* [PATCH V9 01/21] csky: Build infrastructure
2018-10-16 2:58 [PATCH V9 00/21] C-SKY(csky) Linux Kernel Port Guo Ren
@ 2018-10-16 2:58 ` Guo Ren
2018-10-23 0:08 ` Guo Ren
2018-10-16 2:58 ` [PATCH V9 02/21] csky: defconfig Guo Ren
` (20 subsequent siblings)
21 siblings, 1 reply; 68+ messages in thread
From: Guo Ren @ 2018-10-16 2:58 UTC (permalink / raw)
To: akpm, arnd, daniel.lezcano, davem, gregkh, hch, marc.zyngier,
mark.rutland, peterz, robh, tglx
Cc: linux-kernel, linux-arch, devicetree, robh+dt,
c-sky_gcc_upstream, Guo Ren
This patch adds Makefile, Kconfig for build infrastructure.
Signed-off-by: Guo Ren <ren_guo@c-sky.com>
Cc: Arnd Bergmann <arnd@arndb.de>
---
Changelog:
- make allmodconfig is OK.
- Use BUILTIN_DTB as string, remove bool define for compile.
- Add xor.h in asm/Kbuild
- Add compat.h in asm/Kbuild.
- Add select DMA_DIRECT_OPS in Kconfig.
- remove kernel/platform.c in Makefile.
---
---
arch/csky/Kconfig | 205 +++++++++++++++++++++++++++++++++
arch/csky/Kconfig.debug | 9 ++
arch/csky/Makefile | 93 +++++++++++++++
arch/csky/abiv1/Makefile | 8 ++
arch/csky/abiv2/Makefile | 10 ++
arch/csky/boot/Makefile | 24 ++++
arch/csky/boot/dts/Makefile | 13 +++
arch/csky/boot/dts/include/dt-bindings | 1 +
arch/csky/include/asm/Kbuild | 72 ++++++++++++
arch/csky/include/uapi/asm/Kbuild | 33 ++++++
arch/csky/kernel/Makefile | 8 ++
arch/csky/lib/Makefile | 1 +
arch/csky/mm/Makefile | 13 +++
13 files changed, 490 insertions(+)
create mode 100644 arch/csky/Kconfig
create mode 100644 arch/csky/Kconfig.debug
create mode 100644 arch/csky/Makefile
create mode 100644 arch/csky/abiv1/Makefile
create mode 100644 arch/csky/abiv2/Makefile
create mode 100644 arch/csky/boot/Makefile
create mode 100644 arch/csky/boot/dts/Makefile
create mode 120000 arch/csky/boot/dts/include/dt-bindings
create mode 100644 arch/csky/include/asm/Kbuild
create mode 100644 arch/csky/include/uapi/asm/Kbuild
create mode 100644 arch/csky/kernel/Makefile
create mode 100644 arch/csky/lib/Makefile
create mode 100644 arch/csky/mm/Makefile
diff --git a/arch/csky/Kconfig b/arch/csky/Kconfig
new file mode 100644
index 0000000..0a05585
--- /dev/null
+++ b/arch/csky/Kconfig
@@ -0,0 +1,205 @@
+config CSKY
+ def_bool y
+ select ARCH_HAS_SYNC_DMA_FOR_CPU
+ select ARCH_HAS_SYNC_DMA_FOR_DEVICE
+ select ARCH_USE_BUILTIN_BSWAP
+ select ARCH_USE_QUEUED_RWLOCKS if NR_CPUS>2
+ select COMMON_CLK
+ select CLKSRC_MMIO
+ select CLKSRC_OF
+ select DMA_DIRECT_OPS
+ select DMA_NONCOHERENT_OPS
+ select IRQ_DOMAIN
+ select HANDLE_DOMAIN_IRQ
+ select DW_APB_TIMER_OF
+ select GENERIC_LIB_ASHLDI3
+ select GENERIC_LIB_ASHRDI3
+ select GENERIC_LIB_LSHRDI3
+ select GENERIC_LIB_MULDI3
+ select GENERIC_LIB_CMPDI2
+ select GENERIC_LIB_UCMPDI2
+ select GENERIC_ALLOCATOR
+ select GENERIC_ATOMIC64
+ select GENERIC_CLOCKEVENTS
+ select GENERIC_CPU_DEVICES
+ select GENERIC_IRQ_CHIP
+ select GENERIC_IRQ_PROBE
+ select GENERIC_IRQ_SHOW
+ select GENERIC_IRQ_MULTI_HANDLER
+ select GENERIC_SCHED_CLOCK
+ select GENERIC_SMP_IDLE_THREAD
+ select HAVE_ARCH_TRACEHOOK
+ select HAVE_GENERIC_DMA_COHERENT
+ select HAVE_KERNEL_GZIP
+ select HAVE_KERNEL_LZO
+ select HAVE_KERNEL_LZMA
+ select HAVE_C_RECORDMCOUNT
+ select HAVE_DMA_API_DEBUG
+ select HAVE_DMA_CONTIGUOUS
+ select HAVE_MEMBLOCK
+ select MAY_HAVE_SPARSE_IRQ
+ select MODULES_USE_ELF_RELA if MODULES
+ select NO_BOOTMEM
+ select OF
+ select OF_EARLY_FLATTREE
+ select OF_RESERVED_MEM
+ select PERF_USE_VMALLOC
+ select RTC_LIB
+ select TIMER_OF
+ select USB_ARCH_HAS_EHCI
+ select USB_ARCH_HAS_OHCI
+
+config CPU_HAS_CACHEV2
+ bool
+
+config CPU_HAS_FPUV2
+ bool
+
+config CPU_HAS_HILO
+ bool
+
+config CPU_HAS_TLBI
+ bool
+
+config CPU_HAS_LDSTEX
+ bool
+ help
+ For SMP, CPU needs "ldex&stex" instrcutions to atomic operations.
+
+config CPU_NEED_TLBSYNC
+ bool
+
+config CPU_NEED_SOFTALIGN
+ bool
+
+config CPU_NO_USER_BKPT
+ bool
+ help
+ For abiv2 we couldn't use "trap 1" as user space bkpt in gdbserver, because
+ abiv2 is 16/32bit instruction set and "trap 1" is 32bit.
+ So we need a 16bit instruction as user space bkpt, and it will cause an illegal
+ instruction exception.
+ In kernel we parse the *regs->pc to determine whether to send SIGTRAP or not.
+
+config GENERIC_CALIBRATE_DELAY
+ def_bool y
+
+config GENERIC_CSUM
+ def_bool y
+
+config GENERIC_HWEIGHT
+ def_bool y
+
+config MMU
+ def_bool y
+
+config RWSEM_GENERIC_SPINLOCK
+ def_bool y
+
+config TIME_LOW_RES
+ def_bool y
+
+config TRACE_IRQFLAGS_SUPPORT
+ def_bool y
+
+config CPU_TLB_SIZE
+ int
+ default "128" if (CPU_CK610 || CPU_CK807 || CPU_CK810)
+ default "1024" if (CPU_CK860)
+
+config CPU_ASID_BITS
+ int
+ default "8" if (CPU_CK610 || CPU_CK807 || CPU_CK810)
+ default "12" if (CPU_CK860)
+
+config L1_CACHE_SHIFT
+ int
+ default "4" if (CPU_CK610)
+ default "5" if (CPU_CK807 || CPU_CK810)
+ default "6" if (CPU_CK860)
+
+menu "Processor type and features"
+
+choice
+ prompt "CPU MODEL"
+ default CPU_CK807
+
+config CPU_CK610
+ bool "CSKY CPU ck610"
+ select CPU_NEED_TLBSYNC
+ select CPU_NEED_SOFTALIGN
+ select CPU_NO_USER_BKPT
+
+config CPU_CK810
+ bool "CSKY CPU ck810"
+ select CPU_HAS_HILO
+ select CPU_NEED_TLBSYNC
+
+config CPU_CK807
+ bool "CSKY CPU ck807"
+ select CPU_HAS_HILO
+
+config CPU_CK860
+ bool "CSKY CPU ck860"
+ select CPU_HAS_TLBI
+ select CPU_HAS_CACHEV2
+ select CPU_HAS_LDSTEX
+ select CPU_HAS_FPUV2
+endchoice
+
+choice
+ prompt "Power Manager Instruction (wait/doze/stop)"
+ default CPU_PM_NONE
+
+config CPU_PM_NONE
+ bool "None"
+
+config CPU_PM_WAIT
+ bool "wait"
+
+config CPU_PM_DOZE
+ bool "doze"
+
+config CPU_PM_STOP
+ bool "stop"
+endchoice
+
+config CPU_HAS_VDSP
+ bool "CPU has VDSP coprocessor"
+ depends on CPU_HAS_FPU && CPU_HAS_FPUV2
+
+config CPU_HAS_FPU
+ bool "CPU has FPU coprocessor"
+ depends on CPU_CK807 || CPU_CK810 || CPU_CK860
+
+config CPU_HAS_TEE
+ bool "CPU has Trusted Execution Environment"
+ depends on CPU_CK810
+
+config SMP
+ bool "Symmetric Multi-Processing (SMP) support for C-SKY"
+ depends on CPU_CK860
+ default n
+
+config NR_CPUS
+ int "Maximum number of CPUs (2-32)"
+ range 2 32
+ depends on SMP
+ default "2"
+
+config HIGHMEM
+ bool "High Memory Support"
+ depends on !CPU_CK610
+ default y
+
+config FORCE_MAX_ZONEORDER
+ int "Maximum zone order"
+ default "11"
+
+config RAM_BASE
+ hex "DRAM start addr (the same with memory-section in dts)"
+ default 0x0
+
+endmenu
+
+source "kernel/Kconfig.hz"
diff --git a/arch/csky/Kconfig.debug b/arch/csky/Kconfig.debug
new file mode 100644
index 0000000..48cf6ff
--- /dev/null
+++ b/arch/csky/Kconfig.debug
@@ -0,0 +1,9 @@
+menu "C-SKY Debug Options"
+config CSKY_BUILTIN_DTB
+ string "Use kernel builtin dtb"
+ help
+ User could define the dtb instead of the one which is passed from
+ bootloader.
+ Sometimes for debug, we want to use a built-in dtb and then we needn't
+ modify bootloader at all.
+endmenu
diff --git a/arch/csky/Makefile b/arch/csky/Makefile
new file mode 100644
index 0000000..67a4ae1
--- /dev/null
+++ b/arch/csky/Makefile
@@ -0,0 +1,93 @@
+OBJCOPYFLAGS :=-O binary
+GZFLAGS :=-9
+KBUILD_DEFCONFIG := defconfig
+
+ifdef CONFIG_CPU_HAS_FPU
+FPUEXT = f
+endif
+
+ifdef CONFIG_CPU_HAS_VDSP
+VDSPEXT = v
+endif
+
+ifdef CONFIG_CPU_HAS_TEE
+TEEEXT = t
+endif
+
+ifdef CONFIG_CPU_CK610
+CPUTYPE = ck610
+CSKYABI = abiv1
+endif
+
+ifdef CONFIG_CPU_CK810
+CPUTYPE = ck810
+CSKYABI = abiv2
+endif
+
+ifdef CONFIG_CPU_CK807
+CPUTYPE = ck807
+CSKYABI = abiv2
+endif
+
+ifdef CONFIG_CPU_CK860
+CPUTYPE = ck860
+CSKYABI = abiv2
+endif
+
+ifneq ($(CSKYABI),)
+MCPU_STR = $(CPUTYPE)$(FPUEXT)$(VDSPEXT)$(TEEEXT)
+KBUILD_CFLAGS += -mcpu=$(MCPU_STR)
+KBUILD_CFLAGS += -DCSKYCPU_DEF_NAME=\"$(MCPU_STR)\"
+KBUILD_CFLAGS += -msoft-float -mdiv
+KBUILD_CFLAGS += -fno-tree-vectorize
+endif
+
+KBUILD_CFLAGS += -pipe
+ifeq ($(CSKYABI),abiv2)
+KBUILD_CFLAGS += -mno-stack-size
+endif
+
+abidirs := $(patsubst %,arch/csky/%/,$(CSKYABI))
+KBUILD_CFLAGS += $(patsubst %,-I$(srctree)/%inc,$(abidirs))
+
+KBUILD_CPPFLAGS += -mlittle-endian
+LDFLAGS += -EL
+
+KBUILD_AFLAGS += $(KBUILD_CFLAGS)
+
+head-y := arch/csky/kernel/head.o
+
+core-y += arch/csky/kernel/
+core-y += arch/csky/mm/
+core-y += arch/csky/$(CSKYABI)/
+
+libs-y += arch/csky/lib/ \
+ $(shell $(CC) $(KBUILD_CFLAGS) $(KCFLAGS) -print-libgcc-file-name)
+
+boot := arch/csky/boot
+ifneq '$(CONFIG_CSKY_BUILTIN_DTB)' '""'
+core-y += $(boot)/dts/
+endif
+
+all: zImage
+
+
+dtbs: scripts
+ $(Q)$(MAKE) $(build)=$(boot)/dts
+
+%.dtb %.dtb.S %.dtb.o: scripts
+ $(Q)$(MAKE) $(build)=$(boot)/dts $(boot)/dts/$@
+
+zImage Image uImage: vmlinux dtbs
+ $(Q)$(MAKE) $(build)=$(boot) $(boot)/$@
+
+archclean:
+ $(Q)$(MAKE) $(clean)=$(boot)
+ $(Q)$(MAKE) $(clean)=$(boot)/dts
+ rm -rf arch/csky/include/generated
+
+define archhelp
+ echo '* zImage - Compressed kernel image (arch/$(ARCH)/boot/zImage)'
+ echo ' Image - Uncompressed kernel image (arch/$(ARCH)/boot/Image)'
+ echo ' uImage - U-Boot wrapped zImage'
+endef
diff --git a/arch/csky/abiv1/Makefile b/arch/csky/abiv1/Makefile
new file mode 100644
index 0000000..7c06276
--- /dev/null
+++ b/arch/csky/abiv1/Makefile
@@ -0,0 +1,8 @@
+obj-$(CONFIG_CPU_NEED_SOFTALIGN) += alignment.o
+obj-y += bswapdi.o
+obj-y += bswapsi.o
+obj-y += cacheflush.o
+obj-y += mmap.o
+obj-y += memcpy.o
+obj-y += memset.o
+obj-y += strksyms.o
diff --git a/arch/csky/abiv2/Makefile b/arch/csky/abiv2/Makefile
new file mode 100644
index 0000000..069ca72
--- /dev/null
+++ b/arch/csky/abiv2/Makefile
@@ -0,0 +1,10 @@
+obj-y += cacheflush.o
+obj-$(CONFIG_CPU_HAS_FPU) += fpu.o
+obj-y += memcmp.o
+obj-y += memcpy.o
+obj-y += memmove.o
+obj-y += memset.o
+obj-y += strcmp.o
+obj-y += strcpy.o
+obj-y += strlen.o
+obj-y += strksyms.o
diff --git a/arch/csky/boot/Makefile b/arch/csky/boot/Makefile
new file mode 100644
index 0000000..47d3d72
--- /dev/null
+++ b/arch/csky/boot/Makefile
@@ -0,0 +1,24 @@
+targets := Image zImage uImage
+targets += $(dtb-y)
+
+$(obj)/Image: vmlinux FORCE
+ $(call if_changed,objcopy)
+ @echo ' Kernel: $@ is ready'
+
+compress-$(CONFIG_KERNEL_GZIP) = gzip
+compress-$(CONFIG_KERNEL_LZO) = lzo
+compress-$(CONFIG_KERNEL_LZMA) = lzma
+compress-$(CONFIG_KERNEL_XZ) = xzkern
+compress-$(CONFIG_KERNEL_LZ4) = lz4
+
+$(obj)/zImage: $(obj)/Image FORCE
+ $(call if_changed,$(compress-y))
+ @echo ' Kernel: $@ is ready'
+
+UIMAGE_ARCH = sandbox
+UIMAGE_COMPRESSION = $(compress-y)
+UIMAGE_LOADADDR = $(shell $(NM) vmlinux | awk '$$NF == "_start" {print $$1}')
+
+$(obj)/uImage: $(obj)/zImage
+ $(call if_changed,uimage)
+ @echo 'Image: $@ is ready'
diff --git a/arch/csky/boot/dts/Makefile b/arch/csky/boot/dts/Makefile
new file mode 100644
index 0000000..305e81a
--- /dev/null
+++ b/arch/csky/boot/dts/Makefile
@@ -0,0 +1,13 @@
+dtstree := $(srctree)/$(src)
+
+ifneq '$(CONFIG_CSKY_BUILTIN_DTB)' '""'
+builtindtb-y := $(patsubst "%",%,$(CONFIG_CSKY_BUILTIN_DTB))
+dtb-y += $(builtindtb-y).dtb
+obj-y += $(builtindtb-y).dtb.o
+.SECONDARY: $(obj)/$(builtindtb-y).dtb.S
+else
+dtb-y := $(patsubst $(dtstree)/%.dts,%.dtb, $(wildcard $(dtstree)/*.dts))
+endif
+
+always += $(dtb-y)
+clean-files += *.dtb *.dtb.S
diff --git a/arch/csky/boot/dts/include/dt-bindings b/arch/csky/boot/dts/include/dt-bindings
new file mode 120000
index 0000000..08c00e4
--- /dev/null
+++ b/arch/csky/boot/dts/include/dt-bindings
@@ -0,0 +1 @@
+../../../../../include/dt-bindings
\ No newline at end of file
diff --git a/arch/csky/include/asm/Kbuild b/arch/csky/include/asm/Kbuild
new file mode 100644
index 0000000..014eb57
--- /dev/null
+++ b/arch/csky/include/asm/Kbuild
@@ -0,0 +1,72 @@
+generic-y += asm-offsets.h
+generic-y += bitsperlong.h
+generic-y += bugs.h
+generic-y += clkdev.h
+generic-y += compat.h
+generic-y += current.h
+generic-y += delay.h
+generic-y += device.h
+generic-y += div64.h
+generic-y += dma.h
+generic-y += dma-contiguous.h
+generic-y += dma-mapping.h
+generic-y += emergency-restart.h
+generic-y += errno.h
+generic-y += exec.h
+generic-y += fb.h
+generic-y += ftrace.h
+generic-y += futex.h
+generic-y += gpio.h
+generic-y += hardirq.h
+generic-y += hw_irq.h
+generic-y += ioctl.h
+generic-y += ioctls.h
+generic-y += irq.h
+generic-y += irq_regs.h
+generic-y += ipcbuf.h
+generic-y += irq_work.h
+generic-y += kdebug.h
+generic-y += kmap_types.h
+generic-y += kprobes.h
+generic-y += kvm_para.h
+generic-y += linkage.h
+generic-y += local.h
+generic-y += local64.h
+generic-y += mman.h
+generic-y += mm-arch-hooks.h
+generic-y += module.h
+generic-y += msgbuf.h
+generic-y += mutex.h
+generic-y += param.h
+generic-y += pci.h
+generic-y += percpu.h
+generic-y += posix_types.h
+generic-y += poll.h
+generic-y += preempt.h
+generic-y += qrwlock.h
+generic-y += resource.h
+generic-y += scatterlist.h
+generic-y += sections.h
+generic-y += sembuf.h
+generic-y += serial.h
+generic-y += setup.h
+generic-y += shmbuf.h
+generic-y += shm.h
+generic-y += siginfo.h
+generic-y += signal.h
+generic-y += socket.h
+generic-y += sockios.h
+generic-y += statfs.h
+generic-y += termbits.h
+generic-y += termios.h
+generic-y += timex.h
+generic-y += topology.h
+generic-y += trace_clock.h
+generic-y += types.h
+generic-y += ucontext.h
+generic-y += unaligned.h
+generic-y += user.h
+generic-y += vga.h
+generic-y += vmlinux.lds.h
+generic-y += word-at-a-time.h
+generic-y += xor.h
diff --git a/arch/csky/include/uapi/asm/Kbuild b/arch/csky/include/uapi/asm/Kbuild
new file mode 100644
index 0000000..b65252b
--- /dev/null
+++ b/arch/csky/include/uapi/asm/Kbuild
@@ -0,0 +1,33 @@
+include include/uapi/asm-generic/Kbuild.asm
+
+header-y += cachectl.h
+header-y += stat.h
+
+generic-y += auxvec.h
+generic-y += param.h
+generic-y += bpf_perf_event.h
+generic-y += errno.h
+generic-y += fcntl.h
+generic-y += ioctl.h
+generic-y += ioctls.h
+generic-y += ipcbuf.h
+generic-y += shmbuf.h
+generic-y += bitsperlong.h
+generic-y += mman.h
+generic-y += msgbuf.h
+generic-y += poll.h
+generic-y += posix_types.h
+generic-y += resource.h
+generic-y += sembuf.h
+generic-y += siginfo.h
+generic-y += signal.h
+generic-y += socket.h
+generic-y += sockios.h
+generic-y += statfs.h
+generic-y += stat.h
+generic-y += setup.h
+generic-y += swab.h
+generic-y += termbits.h
+generic-y += termios.h
+generic-y += types.h
+generic-y += ucontext.h
diff --git a/arch/csky/kernel/Makefile b/arch/csky/kernel/Makefile
new file mode 100644
index 0000000..4422de7
--- /dev/null
+++ b/arch/csky/kernel/Makefile
@@ -0,0 +1,8 @@
+extra-y := head.o vmlinux.lds
+
+obj-y += entry.o atomic.o signal.o traps.o irq.o time.o vdso.o
+obj-y += power.o syscall.o syscall_table.o setup.o
+obj-y += process.o cpu-probe.o ptrace.o dumpstack.o
+
+obj-$(CONFIG_MODULES) += module.o
+obj-$(CONFIG_SMP) += smp.o
diff --git a/arch/csky/lib/Makefile b/arch/csky/lib/Makefile
new file mode 100644
index 0000000..d1f368c
--- /dev/null
+++ b/arch/csky/lib/Makefile
@@ -0,0 +1 @@
+lib-y := usercopy.o delay.o
diff --git a/arch/csky/mm/Makefile b/arch/csky/mm/Makefile
new file mode 100644
index 0000000..c870eb3
--- /dev/null
+++ b/arch/csky/mm/Makefile
@@ -0,0 +1,13 @@
+ifeq ($(CONFIG_CPU_HAS_CACHEV2),y)
+obj-y += cachev2.o
+else
+obj-y += cachev1.o
+endif
+
+obj-y += dma-mapping.o
+obj-y += fault.o
+obj-$(CONFIG_HIGHMEM) += highmem.o
+obj-y += init.o
+obj-y += ioremap.o
+obj-y += syscache.o
+obj-y += tlb.o
--
2.7.4
^ permalink raw reply related [flat|nested] 68+ messages in thread* Re: [PATCH V9 01/21] csky: Build infrastructure
2018-10-16 2:58 ` [PATCH V9 01/21] csky: Build infrastructure Guo Ren
@ 2018-10-23 0:08 ` Guo Ren
2018-10-24 22:53 ` Arnd Bergmann
0 siblings, 1 reply; 68+ messages in thread
From: Guo Ren @ 2018-10-23 0:08 UTC (permalink / raw)
To: arnd; +Cc: linux-kernel, linux-arch
Hi Arnd,
Could you help me review the patch, seems the patch is skipped.
ref: https://lkml.org/lkml/2018/10/16/224
Thx
On Tue, Oct 16, 2018 at 10:58:20AM +0800, Guo Ren wrote:
> This patch adds Makefile, Kconfig for build infrastructure.
>
> Signed-off-by: Guo Ren <ren_guo@c-sky.com>
> Cc: Arnd Bergmann <arnd@arndb.de>
> ---
> Changelog:
> - make allmodconfig is OK.
> - Use BUILTIN_DTB as string, remove bool define for compile.
> - Add xor.h in asm/Kbuild
> - Add compat.h in asm/Kbuild.
> - Add select DMA_DIRECT_OPS in Kconfig.
> - remove kernel/platform.c in Makefile.
> ---
> ---
> arch/csky/Kconfig | 205 +++++++++++++++++++++++++++++++++
> arch/csky/Kconfig.debug | 9 ++
> arch/csky/Makefile | 93 +++++++++++++++
> arch/csky/abiv1/Makefile | 8 ++
> arch/csky/abiv2/Makefile | 10 ++
> arch/csky/boot/Makefile | 24 ++++
> arch/csky/boot/dts/Makefile | 13 +++
> arch/csky/boot/dts/include/dt-bindings | 1 +
> arch/csky/include/asm/Kbuild | 72 ++++++++++++
> arch/csky/include/uapi/asm/Kbuild | 33 ++++++
> arch/csky/kernel/Makefile | 8 ++
> arch/csky/lib/Makefile | 1 +
> arch/csky/mm/Makefile | 13 +++
> 13 files changed, 490 insertions(+)
> create mode 100644 arch/csky/Kconfig
> create mode 100644 arch/csky/Kconfig.debug
> create mode 100644 arch/csky/Makefile
> create mode 100644 arch/csky/abiv1/Makefile
> create mode 100644 arch/csky/abiv2/Makefile
> create mode 100644 arch/csky/boot/Makefile
> create mode 100644 arch/csky/boot/dts/Makefile
> create mode 120000 arch/csky/boot/dts/include/dt-bindings
> create mode 100644 arch/csky/include/asm/Kbuild
> create mode 100644 arch/csky/include/uapi/asm/Kbuild
> create mode 100644 arch/csky/kernel/Makefile
> create mode 100644 arch/csky/lib/Makefile
> create mode 100644 arch/csky/mm/Makefile
>
> diff --git a/arch/csky/Kconfig b/arch/csky/Kconfig
> new file mode 100644
> index 0000000..0a05585
> --- /dev/null
> +++ b/arch/csky/Kconfig
> @@ -0,0 +1,205 @@
> +config CSKY
> + def_bool y
> + select ARCH_HAS_SYNC_DMA_FOR_CPU
> + select ARCH_HAS_SYNC_DMA_FOR_DEVICE
> + select ARCH_USE_BUILTIN_BSWAP
> + select ARCH_USE_QUEUED_RWLOCKS if NR_CPUS>2
> + select COMMON_CLK
> + select CLKSRC_MMIO
> + select CLKSRC_OF
> + select DMA_DIRECT_OPS
> + select DMA_NONCOHERENT_OPS
> + select IRQ_DOMAIN
> + select HANDLE_DOMAIN_IRQ
> + select DW_APB_TIMER_OF
> + select GENERIC_LIB_ASHLDI3
> + select GENERIC_LIB_ASHRDI3
> + select GENERIC_LIB_LSHRDI3
> + select GENERIC_LIB_MULDI3
> + select GENERIC_LIB_CMPDI2
> + select GENERIC_LIB_UCMPDI2
> + select GENERIC_ALLOCATOR
> + select GENERIC_ATOMIC64
> + select GENERIC_CLOCKEVENTS
> + select GENERIC_CPU_DEVICES
> + select GENERIC_IRQ_CHIP
> + select GENERIC_IRQ_PROBE
> + select GENERIC_IRQ_SHOW
> + select GENERIC_IRQ_MULTI_HANDLER
> + select GENERIC_SCHED_CLOCK
> + select GENERIC_SMP_IDLE_THREAD
> + select HAVE_ARCH_TRACEHOOK
> + select HAVE_GENERIC_DMA_COHERENT
> + select HAVE_KERNEL_GZIP
> + select HAVE_KERNEL_LZO
> + select HAVE_KERNEL_LZMA
> + select HAVE_C_RECORDMCOUNT
> + select HAVE_DMA_API_DEBUG
> + select HAVE_DMA_CONTIGUOUS
> + select HAVE_MEMBLOCK
> + select MAY_HAVE_SPARSE_IRQ
> + select MODULES_USE_ELF_RELA if MODULES
> + select NO_BOOTMEM
> + select OF
> + select OF_EARLY_FLATTREE
> + select OF_RESERVED_MEM
> + select PERF_USE_VMALLOC
> + select RTC_LIB
> + select TIMER_OF
> + select USB_ARCH_HAS_EHCI
> + select USB_ARCH_HAS_OHCI
> +
> +config CPU_HAS_CACHEV2
> + bool
> +
> +config CPU_HAS_FPUV2
> + bool
> +
> +config CPU_HAS_HILO
> + bool
> +
> +config CPU_HAS_TLBI
> + bool
> +
> +config CPU_HAS_LDSTEX
> + bool
> + help
> + For SMP, CPU needs "ldex&stex" instrcutions to atomic operations.
> +
> +config CPU_NEED_TLBSYNC
> + bool
> +
> +config CPU_NEED_SOFTALIGN
> + bool
> +
> +config CPU_NO_USER_BKPT
> + bool
> + help
> + For abiv2 we couldn't use "trap 1" as user space bkpt in gdbserver, because
> + abiv2 is 16/32bit instruction set and "trap 1" is 32bit.
> + So we need a 16bit instruction as user space bkpt, and it will cause an illegal
> + instruction exception.
> + In kernel we parse the *regs->pc to determine whether to send SIGTRAP or not.
> +
> +config GENERIC_CALIBRATE_DELAY
> + def_bool y
> +
> +config GENERIC_CSUM
> + def_bool y
> +
> +config GENERIC_HWEIGHT
> + def_bool y
> +
> +config MMU
> + def_bool y
> +
> +config RWSEM_GENERIC_SPINLOCK
> + def_bool y
> +
> +config TIME_LOW_RES
> + def_bool y
> +
> +config TRACE_IRQFLAGS_SUPPORT
> + def_bool y
> +
> +config CPU_TLB_SIZE
> + int
> + default "128" if (CPU_CK610 || CPU_CK807 || CPU_CK810)
> + default "1024" if (CPU_CK860)
> +
> +config CPU_ASID_BITS
> + int
> + default "8" if (CPU_CK610 || CPU_CK807 || CPU_CK810)
> + default "12" if (CPU_CK860)
> +
> +config L1_CACHE_SHIFT
> + int
> + default "4" if (CPU_CK610)
> + default "5" if (CPU_CK807 || CPU_CK810)
> + default "6" if (CPU_CK860)
> +
> +menu "Processor type and features"
> +
> +choice
> + prompt "CPU MODEL"
> + default CPU_CK807
> +
> +config CPU_CK610
> + bool "CSKY CPU ck610"
> + select CPU_NEED_TLBSYNC
> + select CPU_NEED_SOFTALIGN
> + select CPU_NO_USER_BKPT
> +
> +config CPU_CK810
> + bool "CSKY CPU ck810"
> + select CPU_HAS_HILO
> + select CPU_NEED_TLBSYNC
> +
> +config CPU_CK807
> + bool "CSKY CPU ck807"
> + select CPU_HAS_HILO
> +
> +config CPU_CK860
> + bool "CSKY CPU ck860"
> + select CPU_HAS_TLBI
> + select CPU_HAS_CACHEV2
> + select CPU_HAS_LDSTEX
> + select CPU_HAS_FPUV2
> +endchoice
> +
> +choice
> + prompt "Power Manager Instruction (wait/doze/stop)"
> + default CPU_PM_NONE
> +
> +config CPU_PM_NONE
> + bool "None"
> +
> +config CPU_PM_WAIT
> + bool "wait"
> +
> +config CPU_PM_DOZE
> + bool "doze"
> +
> +config CPU_PM_STOP
> + bool "stop"
> +endchoice
> +
> +config CPU_HAS_VDSP
> + bool "CPU has VDSP coprocessor"
> + depends on CPU_HAS_FPU && CPU_HAS_FPUV2
> +
> +config CPU_HAS_FPU
> + bool "CPU has FPU coprocessor"
> + depends on CPU_CK807 || CPU_CK810 || CPU_CK860
> +
> +config CPU_HAS_TEE
> + bool "CPU has Trusted Execution Environment"
> + depends on CPU_CK810
> +
> +config SMP
> + bool "Symmetric Multi-Processing (SMP) support for C-SKY"
> + depends on CPU_CK860
> + default n
> +
> +config NR_CPUS
> + int "Maximum number of CPUs (2-32)"
> + range 2 32
> + depends on SMP
> + default "2"
> +
> +config HIGHMEM
> + bool "High Memory Support"
> + depends on !CPU_CK610
> + default y
> +
> +config FORCE_MAX_ZONEORDER
> + int "Maximum zone order"
> + default "11"
> +
> +config RAM_BASE
> + hex "DRAM start addr (the same with memory-section in dts)"
> + default 0x0
> +
> +endmenu
> +
> +source "kernel/Kconfig.hz"
> diff --git a/arch/csky/Kconfig.debug b/arch/csky/Kconfig.debug
> new file mode 100644
> index 0000000..48cf6ff
> --- /dev/null
> +++ b/arch/csky/Kconfig.debug
> @@ -0,0 +1,9 @@
> +menu "C-SKY Debug Options"
> +config CSKY_BUILTIN_DTB
> + string "Use kernel builtin dtb"
> + help
> + User could define the dtb instead of the one which is passed from
> + bootloader.
> + Sometimes for debug, we want to use a built-in dtb and then we needn't
> + modify bootloader at all.
> +endmenu
> diff --git a/arch/csky/Makefile b/arch/csky/Makefile
> new file mode 100644
> index 0000000..67a4ae1
> --- /dev/null
> +++ b/arch/csky/Makefile
> @@ -0,0 +1,93 @@
> +OBJCOPYFLAGS :=-O binary
> +GZFLAGS :=-9
> +KBUILD_DEFCONFIG := defconfig
> +
> +ifdef CONFIG_CPU_HAS_FPU
> +FPUEXT = f
> +endif
> +
> +ifdef CONFIG_CPU_HAS_VDSP
> +VDSPEXT = v
> +endif
> +
> +ifdef CONFIG_CPU_HAS_TEE
> +TEEEXT = t
> +endif
> +
> +ifdef CONFIG_CPU_CK610
> +CPUTYPE = ck610
> +CSKYABI = abiv1
> +endif
> +
> +ifdef CONFIG_CPU_CK810
> +CPUTYPE = ck810
> +CSKYABI = abiv2
> +endif
> +
> +ifdef CONFIG_CPU_CK807
> +CPUTYPE = ck807
> +CSKYABI = abiv2
> +endif
> +
> +ifdef CONFIG_CPU_CK860
> +CPUTYPE = ck860
> +CSKYABI = abiv2
> +endif
> +
> +ifneq ($(CSKYABI),)
> +MCPU_STR = $(CPUTYPE)$(FPUEXT)$(VDSPEXT)$(TEEEXT)
> +KBUILD_CFLAGS += -mcpu=$(MCPU_STR)
> +KBUILD_CFLAGS += -DCSKYCPU_DEF_NAME=\"$(MCPU_STR)\"
> +KBUILD_CFLAGS += -msoft-float -mdiv
> +KBUILD_CFLAGS += -fno-tree-vectorize
> +endif
> +
> +KBUILD_CFLAGS += -pipe
> +ifeq ($(CSKYABI),abiv2)
> +KBUILD_CFLAGS += -mno-stack-size
> +endif
> +
> +abidirs := $(patsubst %,arch/csky/%/,$(CSKYABI))
> +KBUILD_CFLAGS += $(patsubst %,-I$(srctree)/%inc,$(abidirs))
> +
> +KBUILD_CPPFLAGS += -mlittle-endian
> +LDFLAGS += -EL
> +
> +KBUILD_AFLAGS += $(KBUILD_CFLAGS)
> +
> +head-y := arch/csky/kernel/head.o
> +
> +core-y += arch/csky/kernel/
> +core-y += arch/csky/mm/
> +core-y += arch/csky/$(CSKYABI)/
> +
> +libs-y += arch/csky/lib/ \
> + $(shell $(CC) $(KBUILD_CFLAGS) $(KCFLAGS) -print-libgcc-file-name)
> +
> +boot := arch/csky/boot
> +ifneq '$(CONFIG_CSKY_BUILTIN_DTB)' '""'
> +core-y += $(boot)/dts/
> +endif
> +
> +all: zImage
> +
> +
> +dtbs: scripts
> + $(Q)$(MAKE) $(build)=$(boot)/dts
> +
> +%.dtb %.dtb.S %.dtb.o: scripts
> + $(Q)$(MAKE) $(build)=$(boot)/dts $(boot)/dts/$@
> +
> +zImage Image uImage: vmlinux dtbs
> + $(Q)$(MAKE) $(build)=$(boot) $(boot)/$@
> +
> +archclean:
> + $(Q)$(MAKE) $(clean)=$(boot)
> + $(Q)$(MAKE) $(clean)=$(boot)/dts
> + rm -rf arch/csky/include/generated
> +
> +define archhelp
> + echo '* zImage - Compressed kernel image (arch/$(ARCH)/boot/zImage)'
> + echo ' Image - Uncompressed kernel image (arch/$(ARCH)/boot/Image)'
> + echo ' uImage - U-Boot wrapped zImage'
> +endef
> diff --git a/arch/csky/abiv1/Makefile b/arch/csky/abiv1/Makefile
> new file mode 100644
> index 0000000..7c06276
> --- /dev/null
> +++ b/arch/csky/abiv1/Makefile
> @@ -0,0 +1,8 @@
> +obj-$(CONFIG_CPU_NEED_SOFTALIGN) += alignment.o
> +obj-y += bswapdi.o
> +obj-y += bswapsi.o
> +obj-y += cacheflush.o
> +obj-y += mmap.o
> +obj-y += memcpy.o
> +obj-y += memset.o
> +obj-y += strksyms.o
> diff --git a/arch/csky/abiv2/Makefile b/arch/csky/abiv2/Makefile
> new file mode 100644
> index 0000000..069ca72
> --- /dev/null
> +++ b/arch/csky/abiv2/Makefile
> @@ -0,0 +1,10 @@
> +obj-y += cacheflush.o
> +obj-$(CONFIG_CPU_HAS_FPU) += fpu.o
> +obj-y += memcmp.o
> +obj-y += memcpy.o
> +obj-y += memmove.o
> +obj-y += memset.o
> +obj-y += strcmp.o
> +obj-y += strcpy.o
> +obj-y += strlen.o
> +obj-y += strksyms.o
> diff --git a/arch/csky/boot/Makefile b/arch/csky/boot/Makefile
> new file mode 100644
> index 0000000..47d3d72
> --- /dev/null
> +++ b/arch/csky/boot/Makefile
> @@ -0,0 +1,24 @@
> +targets := Image zImage uImage
> +targets += $(dtb-y)
> +
> +$(obj)/Image: vmlinux FORCE
> + $(call if_changed,objcopy)
> + @echo ' Kernel: $@ is ready'
> +
> +compress-$(CONFIG_KERNEL_GZIP) = gzip
> +compress-$(CONFIG_KERNEL_LZO) = lzo
> +compress-$(CONFIG_KERNEL_LZMA) = lzma
> +compress-$(CONFIG_KERNEL_XZ) = xzkern
> +compress-$(CONFIG_KERNEL_LZ4) = lz4
> +
> +$(obj)/zImage: $(obj)/Image FORCE
> + $(call if_changed,$(compress-y))
> + @echo ' Kernel: $@ is ready'
> +
> +UIMAGE_ARCH = sandbox
> +UIMAGE_COMPRESSION = $(compress-y)
> +UIMAGE_LOADADDR = $(shell $(NM) vmlinux | awk '$$NF == "_start" {print $$1}')
> +
> +$(obj)/uImage: $(obj)/zImage
> + $(call if_changed,uimage)
> + @echo 'Image: $@ is ready'
> diff --git a/arch/csky/boot/dts/Makefile b/arch/csky/boot/dts/Makefile
> new file mode 100644
> index 0000000..305e81a
> --- /dev/null
> +++ b/arch/csky/boot/dts/Makefile
> @@ -0,0 +1,13 @@
> +dtstree := $(srctree)/$(src)
> +
> +ifneq '$(CONFIG_CSKY_BUILTIN_DTB)' '""'
> +builtindtb-y := $(patsubst "%",%,$(CONFIG_CSKY_BUILTIN_DTB))
> +dtb-y += $(builtindtb-y).dtb
> +obj-y += $(builtindtb-y).dtb.o
> +.SECONDARY: $(obj)/$(builtindtb-y).dtb.S
> +else
> +dtb-y := $(patsubst $(dtstree)/%.dts,%.dtb, $(wildcard $(dtstree)/*.dts))
> +endif
> +
> +always += $(dtb-y)
> +clean-files += *.dtb *.dtb.S
> diff --git a/arch/csky/boot/dts/include/dt-bindings b/arch/csky/boot/dts/include/dt-bindings
> new file mode 120000
> index 0000000..08c00e4
> --- /dev/null
> +++ b/arch/csky/boot/dts/include/dt-bindings
> @@ -0,0 +1 @@
> +../../../../../include/dt-bindings
> \ No newline at end of file
> diff --git a/arch/csky/include/asm/Kbuild b/arch/csky/include/asm/Kbuild
> new file mode 100644
> index 0000000..014eb57
> --- /dev/null
> +++ b/arch/csky/include/asm/Kbuild
> @@ -0,0 +1,72 @@
> +generic-y += asm-offsets.h
> +generic-y += bitsperlong.h
> +generic-y += bugs.h
> +generic-y += clkdev.h
> +generic-y += compat.h
> +generic-y += current.h
> +generic-y += delay.h
> +generic-y += device.h
> +generic-y += div64.h
> +generic-y += dma.h
> +generic-y += dma-contiguous.h
> +generic-y += dma-mapping.h
> +generic-y += emergency-restart.h
> +generic-y += errno.h
> +generic-y += exec.h
> +generic-y += fb.h
> +generic-y += ftrace.h
> +generic-y += futex.h
> +generic-y += gpio.h
> +generic-y += hardirq.h
> +generic-y += hw_irq.h
> +generic-y += ioctl.h
> +generic-y += ioctls.h
> +generic-y += irq.h
> +generic-y += irq_regs.h
> +generic-y += ipcbuf.h
> +generic-y += irq_work.h
> +generic-y += kdebug.h
> +generic-y += kmap_types.h
> +generic-y += kprobes.h
> +generic-y += kvm_para.h
> +generic-y += linkage.h
> +generic-y += local.h
> +generic-y += local64.h
> +generic-y += mman.h
> +generic-y += mm-arch-hooks.h
> +generic-y += module.h
> +generic-y += msgbuf.h
> +generic-y += mutex.h
> +generic-y += param.h
> +generic-y += pci.h
> +generic-y += percpu.h
> +generic-y += posix_types.h
> +generic-y += poll.h
> +generic-y += preempt.h
> +generic-y += qrwlock.h
> +generic-y += resource.h
> +generic-y += scatterlist.h
> +generic-y += sections.h
> +generic-y += sembuf.h
> +generic-y += serial.h
> +generic-y += setup.h
> +generic-y += shmbuf.h
> +generic-y += shm.h
> +generic-y += siginfo.h
> +generic-y += signal.h
> +generic-y += socket.h
> +generic-y += sockios.h
> +generic-y += statfs.h
> +generic-y += termbits.h
> +generic-y += termios.h
> +generic-y += timex.h
> +generic-y += topology.h
> +generic-y += trace_clock.h
> +generic-y += types.h
> +generic-y += ucontext.h
> +generic-y += unaligned.h
> +generic-y += user.h
> +generic-y += vga.h
> +generic-y += vmlinux.lds.h
> +generic-y += word-at-a-time.h
> +generic-y += xor.h
> diff --git a/arch/csky/include/uapi/asm/Kbuild b/arch/csky/include/uapi/asm/Kbuild
> new file mode 100644
> index 0000000..b65252b
> --- /dev/null
> +++ b/arch/csky/include/uapi/asm/Kbuild
> @@ -0,0 +1,33 @@
> +include include/uapi/asm-generic/Kbuild.asm
> +
> +header-y += cachectl.h
> +header-y += stat.h
> +
> +generic-y += auxvec.h
> +generic-y += param.h
> +generic-y += bpf_perf_event.h
> +generic-y += errno.h
> +generic-y += fcntl.h
> +generic-y += ioctl.h
> +generic-y += ioctls.h
> +generic-y += ipcbuf.h
> +generic-y += shmbuf.h
> +generic-y += bitsperlong.h
> +generic-y += mman.h
> +generic-y += msgbuf.h
> +generic-y += poll.h
> +generic-y += posix_types.h
> +generic-y += resource.h
> +generic-y += sembuf.h
> +generic-y += siginfo.h
> +generic-y += signal.h
> +generic-y += socket.h
> +generic-y += sockios.h
> +generic-y += statfs.h
> +generic-y += stat.h
> +generic-y += setup.h
> +generic-y += swab.h
> +generic-y += termbits.h
> +generic-y += termios.h
> +generic-y += types.h
> +generic-y += ucontext.h
> diff --git a/arch/csky/kernel/Makefile b/arch/csky/kernel/Makefile
> new file mode 100644
> index 0000000..4422de7
> --- /dev/null
> +++ b/arch/csky/kernel/Makefile
> @@ -0,0 +1,8 @@
> +extra-y := head.o vmlinux.lds
> +
> +obj-y += entry.o atomic.o signal.o traps.o irq.o time.o vdso.o
> +obj-y += power.o syscall.o syscall_table.o setup.o
> +obj-y += process.o cpu-probe.o ptrace.o dumpstack.o
> +
> +obj-$(CONFIG_MODULES) += module.o
> +obj-$(CONFIG_SMP) += smp.o
> diff --git a/arch/csky/lib/Makefile b/arch/csky/lib/Makefile
> new file mode 100644
> index 0000000..d1f368c
> --- /dev/null
> +++ b/arch/csky/lib/Makefile
> @@ -0,0 +1 @@
> +lib-y := usercopy.o delay.o
> diff --git a/arch/csky/mm/Makefile b/arch/csky/mm/Makefile
> new file mode 100644
> index 0000000..c870eb3
> --- /dev/null
> +++ b/arch/csky/mm/Makefile
> @@ -0,0 +1,13 @@
> +ifeq ($(CONFIG_CPU_HAS_CACHEV2),y)
> +obj-y += cachev2.o
> +else
> +obj-y += cachev1.o
> +endif
> +
> +obj-y += dma-mapping.o
> +obj-y += fault.o
> +obj-$(CONFIG_HIGHMEM) += highmem.o
> +obj-y += init.o
> +obj-y += ioremap.o
> +obj-y += syscache.o
> +obj-y += tlb.o
> --
> 2.7.4
^ permalink raw reply [flat|nested] 68+ messages in thread* Re: [PATCH V9 01/21] csky: Build infrastructure
2018-10-23 0:08 ` Guo Ren
@ 2018-10-24 22:53 ` Arnd Bergmann
2018-10-25 17:04 ` Guo Ren
0 siblings, 1 reply; 68+ messages in thread
From: Arnd Bergmann @ 2018-10-24 22:53 UTC (permalink / raw)
To: Guo Ren; +Cc: Linux Kernel Mailing List, linux-arch
On Tue, Oct 23, 2018 at 1:08 AM Guo Ren <ren_guo@c-sky.com> wrote:
>
> Hi Arnd,
>
> Could you help me review the patch, seems the patch is skipped.
> ref: https://lkml.org/lkml/2018/10/16/224
Ah right, I missed that. I think I originally had some thoughts on it but
didn't immediately comment and then forgot about it. It's basically
ok, so
Acked-by: Arnd Bergmann <arnd@arndb.de>
but please have a look below for a follow-up patch:
> > diff --git a/arch/csky/include/asm/Kbuild b/arch/csky/include/asm/Kbuild
> > new file mode 100644
> > index 0000000..014eb57
> > --- /dev/null
> > +++ b/arch/csky/include/asm/Kbuild
> > @@ -0,0 +1,72 @@
> > +generic-y += asm-offsets.h
> > +generic-y += bitsperlong.h
> > +generic-y += bugs.h
> > +generic-y += clkdev.h
> > +generic-y += compat.h
> > +generic-y += current.h
> > +generic-y += delay.h
> > +generic-y += device.h
> > +generic-y += div64.h
...
> > +generic-y += errno.h
...
> > +generic-y += ioctl.h
> > +generic-y += ioctls.h
A number of the header files that are marked as generic-y in
asm/Kbuild are also marked the same way in uapi/asm/Kbuild.
I think this is a mistake, since we only need one of the two
for each header.
When both exist, we just take the one from asm/, but then we install
the uapi/asm/ version in the 'make headers_install' stage, and they
both have the same contents, so I think you only need the uapi/asm/
version here.
> > diff --git a/arch/csky/include/uapi/asm/Kbuild b/arch/csky/include/uapi/asm/Kbuild
> > new file mode 100644
> > index 0000000..b65252b
> > --- /dev/null
> > +++ b/arch/csky/include/uapi/asm/Kbuild
> > @@ -0,0 +1,33 @@
> > +include include/uapi/asm-generic/Kbuild.asm
> > +
> > +header-y += cachectl.h
> > +header-y += stat.h
It also seems that the second statement here (for stat.h) is from
an older version, since that header no longer exists, at least if
I'm reading your patches correctly.
Arnd
^ permalink raw reply [flat|nested] 68+ messages in thread* Re: [PATCH V9 01/21] csky: Build infrastructure
2018-10-24 22:53 ` Arnd Bergmann
@ 2018-10-25 17:04 ` Guo Ren
0 siblings, 0 replies; 68+ messages in thread
From: Guo Ren @ 2018-10-25 17:04 UTC (permalink / raw)
To: Arnd Bergmann; +Cc: Linux Kernel Mailing List, linux-arch
On Wed, Oct 24, 2018 at 11:53:05PM +0100, Arnd Bergmann wrote:
> On Tue, Oct 23, 2018 at 1:08 AM Guo Ren <ren_guo@c-sky.com> wrote:
> >
> > Hi Arnd,
> >
> > Could you help me review the patch, seems the patch is skipped.
> > ref: https://lkml.org/lkml/2018/10/16/224
>
> Ah right, I missed that. I think I originally had some thoughts on it but
> didn't immediately comment and then forgot about it. It's basically
> ok, so
>
> Acked-by: Arnd Bergmann <arnd@arndb.de>
Thx.
>
> but please have a look below for a follow-up patch:
>
> > > diff --git a/arch/csky/include/asm/Kbuild b/arch/csky/include/asm/Kbuild
> > > new file mode 100644
> > > index 0000000..014eb57
> > > --- /dev/null
> > > +++ b/arch/csky/include/asm/Kbuild
> > > @@ -0,0 +1,72 @@
> > > +generic-y += asm-offsets.h
> > > +generic-y += bitsperlong.h
> > > +generic-y += bugs.h
> > > +generic-y += clkdev.h
> > > +generic-y += compat.h
> > > +generic-y += current.h
> > > +generic-y += delay.h
> > > +generic-y += device.h
> > > +generic-y += div64.h
> ...
> > > +generic-y += errno.h
> ...
> > > +generic-y += ioctl.h
> > > +generic-y += ioctls.h
>
> A number of the header files that are marked as generic-y in
> asm/Kbuild are also marked the same way in uapi/asm/Kbuild.
> I think this is a mistake, since we only need one of the two
> for each header.
>
> When both exist, we just take the one from asm/, but then we install
> the uapi/asm/ version in the 'make headers_install' stage, and they
> both have the same contents, so I think you only need the uapi/asm/
> version here.
Ok, I've cleaned them up.
>
> > > diff --git a/arch/csky/include/uapi/asm/Kbuild b/arch/csky/include/uapi/asm/Kbuild
> > > new file mode 100644
> > > index 0000000..b65252b
> > > --- /dev/null
> > > +++ b/arch/csky/include/uapi/asm/Kbuild
> > > @@ -0,0 +1,33 @@
> > > +include include/uapi/asm-generic/Kbuild.asm
> > > +
> > > +header-y += cachectl.h
> > > +header-y += stat.h
>
> It also seems that the second statement here (for stat.h) is from
> an older version, since that header no longer exists, at least if
> I'm reading your patches correctly.
Yes, no "header-y += stat.h" and I've removed it.
Thx for the reviewed.
Best Regards
Guo Ren
^ permalink raw reply [flat|nested] 68+ messages in thread
* [PATCH V9 02/21] csky: defconfig
2018-10-16 2:58 [PATCH V9 00/21] C-SKY(csky) Linux Kernel Port Guo Ren
2018-10-16 2:58 ` [PATCH V9 01/21] csky: Build infrastructure Guo Ren
@ 2018-10-16 2:58 ` Guo Ren
2018-10-17 14:56 ` Arnd Bergmann
2018-10-16 2:58 ` [PATCH V9 03/21] csky: Kernel booting Guo Ren
` (19 subsequent siblings)
21 siblings, 1 reply; 68+ messages in thread
From: Guo Ren @ 2018-10-16 2:58 UTC (permalink / raw)
To: akpm, arnd, daniel.lezcano, davem, gregkh, hch, marc.zyngier,
mark.rutland, peterz, robh, tglx
Cc: linux-kernel, linux-arch, devicetree, robh+dt,
c-sky_gcc_upstream, Guo Ren
This patch adds csky defconfig.
Signed-off-by: Guo Ren <ren_guo@c-sky.com>
Cc: Arnd Bergmann <arnd@arndb.de>
---
arch/csky/configs/defconfig | 61 +++++++++++++++++++++++++++++++++++++++++++++
1 file changed, 61 insertions(+)
create mode 100644 arch/csky/configs/defconfig
diff --git a/arch/csky/configs/defconfig b/arch/csky/configs/defconfig
new file mode 100644
index 0000000..7ef4289
--- /dev/null
+++ b/arch/csky/configs/defconfig
@@ -0,0 +1,61 @@
+# CONFIG_LOCALVERSION_AUTO is not set
+CONFIG_DEFAULT_HOSTNAME="csky"
+# CONFIG_SWAP is not set
+CONFIG_SYSVIPC=y
+CONFIG_POSIX_MQUEUE=y
+CONFIG_AUDIT=y
+CONFIG_NO_HZ_IDLE=y
+CONFIG_HIGH_RES_TIMERS=y
+CONFIG_BSD_PROCESS_ACCT=y
+CONFIG_BSD_PROCESS_ACCT_V3=y
+CONFIG_MODULES=y
+CONFIG_MODULE_UNLOAD=y
+CONFIG_DEFAULT_DEADLINE=y
+CONFIG_CPU_CK807=y
+CONFIG_CPU_HAS_FPU=y
+CONFIG_NET=y
+CONFIG_PACKET=y
+CONFIG_UNIX=y
+CONFIG_INET=y
+CONFIG_DEVTMPFS=y
+CONFIG_DEVTMPFS_MOUNT=y
+CONFIG_BLK_DEV_LOOP=y
+CONFIG_BLK_DEV_RAM=y
+CONFIG_BLK_DEV_RAM_SIZE=65536
+CONFIG_VT_HW_CONSOLE_BINDING=y
+CONFIG_SERIAL_NONSTANDARD=y
+CONFIG_SERIAL_8250=y
+CONFIG_SERIAL_8250_CONSOLE=y
+CONFIG_SERIAL_OF_PLATFORM=y
+CONFIG_TTY_PRINTK=y
+# CONFIG_VGA_CONSOLE is not set
+CONFIG_CSKY_MPTIMER=y
+CONFIG_GX6605S_TIMER=y
+CONFIG_PM_DEVFREQ=y
+CONFIG_DEVFREQ_GOV_SIMPLE_ONDEMAND=y
+CONFIG_DEVFREQ_GOV_PERFORMANCE=y
+CONFIG_DEVFREQ_GOV_POWERSAVE=y
+CONFIG_DEVFREQ_GOV_USERSPACE=y
+CONFIG_GENERIC_PHY=y
+CONFIG_EXT4_FS=y
+CONFIG_FANOTIFY=y
+CONFIG_QUOTA=y
+CONFIG_FSCACHE=m
+CONFIG_FSCACHE_STATS=y
+CONFIG_CACHEFILES=m
+CONFIG_MSDOS_FS=y
+CONFIG_VFAT_FS=y
+CONFIG_FAT_DEFAULT_UTF8=y
+CONFIG_NTFS_FS=y
+CONFIG_PROC_KCORE=y
+CONFIG_PROC_CHILDREN=y
+CONFIG_TMPFS=y
+CONFIG_TMPFS_POSIX_ACL=y
+CONFIG_CONFIGFS_FS=y
+CONFIG_CRAMFS=y
+CONFIG_ROMFS_FS=y
+CONFIG_NFS_FS=y
+CONFIG_PRINTK_TIME=y
+CONFIG_DEBUG_INFO=y
+CONFIG_DEBUG_FS=y
+CONFIG_MAGIC_SYSRQ=y
--
2.7.4
^ permalink raw reply related [flat|nested] 68+ messages in thread* Re: [PATCH V9 02/21] csky: defconfig
2018-10-16 2:58 ` [PATCH V9 02/21] csky: defconfig Guo Ren
@ 2018-10-17 14:56 ` Arnd Bergmann
0 siblings, 0 replies; 68+ messages in thread
From: Arnd Bergmann @ 2018-10-17 14:56 UTC (permalink / raw)
To: Guo Ren
Cc: Andrew Morton, Daniel Lezcano, David Miller, gregkh,
Christoph Hellwig, Marc Zyngier, Mark Rutland, Peter Zijlstra,
Rob Herring, Thomas Gleixner, Linux Kernel Mailing List,
linux-arch, DTML, Rob Herring, c-sky_gcc_upstream
On Tue, Oct 16, 2018 at 4:59 AM Guo Ren <ren_guo@c-sky.com> wrote:
>
> This patch adds csky defconfig.
>
> Signed-off-by: Guo Ren <ren_guo@c-sky.com>
> Cc: Arnd Bergmann <arnd@arndb.de>
Reviewed-by: Arnd Bergmann <arnd@arndb.de>
^ permalink raw reply [flat|nested] 68+ messages in thread
* [PATCH V9 03/21] csky: Kernel booting
2018-10-16 2:58 [PATCH V9 00/21] C-SKY(csky) Linux Kernel Port Guo Ren
2018-10-16 2:58 ` [PATCH V9 01/21] csky: Build infrastructure Guo Ren
2018-10-16 2:58 ` [PATCH V9 02/21] csky: defconfig Guo Ren
@ 2018-10-16 2:58 ` Guo Ren
2018-10-17 14:58 ` Arnd Bergmann
2018-10-16 2:58 ` [PATCH V9 04/21] csky: Exception handling and mm-fault Guo Ren
` (18 subsequent siblings)
21 siblings, 1 reply; 68+ messages in thread
From: Guo Ren @ 2018-10-16 2:58 UTC (permalink / raw)
To: akpm, arnd, daniel.lezcano, davem, gregkh, hch, marc.zyngier,
mark.rutland, peterz, robh, tglx
Cc: linux-kernel, linux-arch, devicetree, robh+dt,
c-sky_gcc_upstream, Guo Ren
This patch add boot code. Thx boot params is all in dtb and it's
the only way to let kernel get bootloader param information.
Signed-off-by: Guo Ren <ren_guo@c-sky.com>
Cc: Arnd Bergmann <arnd@arndb.de>
---
Changelog:
- Use built-in dtb when dtb param is NULL.
- Add dummy console for allmodconfig.
---
---
arch/csky/kernel/head.S | 77 ++++++++++++++++++++
arch/csky/kernel/setup.c | 162 +++++++++++++++++++++++++++++++++++++++++
arch/csky/kernel/vmlinux.lds.S | 66 +++++++++++++++++
3 files changed, 305 insertions(+)
create mode 100644 arch/csky/kernel/head.S
create mode 100644 arch/csky/kernel/setup.c
create mode 100644 arch/csky/kernel/vmlinux.lds.S
diff --git a/arch/csky/kernel/head.S b/arch/csky/kernel/head.S
new file mode 100644
index 0000000..9c4ec47
--- /dev/null
+++ b/arch/csky/kernel/head.S
@@ -0,0 +1,77 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+
+#include <linux/linkage.h>
+#include <linux/init.h>
+#include <asm/page.h>
+#include <abi/entry.h>
+
+__HEAD
+ENTRY(_start)
+ /* set super user mode */
+ lrw a3, DEFAULT_PSR_VALUE
+ mtcr a3, psr
+ psrset ee
+
+ SETUP_MMU a3
+
+ /* set stack point */
+ lrw a3, init_thread_union + THREAD_SIZE
+ mov sp, a3
+
+ jmpi csky_start
+END(_start)
+
+#ifdef CONFIG_SMP
+.align 10
+ENTRY(_start_smp_secondary)
+ /* Invalid I/Dcache BTB BHT */
+ movi a3, 7
+ lsli a3, 16
+ addi a3, (1<<4) | 3
+ mtcr a3, cr17
+
+ tlbi.alls
+
+ /* setup PAGEMASK */
+ movi a3, 0
+ mtcr a3, cr<6, 15>
+
+ /* setup MEL0/MEL1 */
+ grs a0, _start_smp_pc
+_start_smp_pc:
+ bmaski a1, 13
+ andn a0, a1
+ movi a1, 0x00000006
+ movi a2, 0x00001006
+ or a1, a0
+ or a2, a0
+ mtcr a1, cr<2, 15>
+ mtcr a2, cr<3, 15>
+
+ /* setup MEH */
+ mtcr a0, cr<4, 15>
+
+ /* write TLB */
+ bgeni a3, 28
+ mtcr a3, cr<8, 15>
+
+ SETUP_MMU a3
+
+ /* enable MMU */
+ movi a3, 1
+ mtcr a3, cr18
+
+ jmpi _goto_mmu_on
+_goto_mmu_on:
+ lrw a3, DEFAULT_PSR_VALUE
+ mtcr a3, psr
+ psrset ee
+
+ /* set stack point */
+ lrw a3, secondary_stack
+ ld.w a3, (a3, 0)
+ mov sp, a3
+
+ jmpi csky_start_secondary
+END(_start_smp_secondary)
+#endif
diff --git a/arch/csky/kernel/setup.c b/arch/csky/kernel/setup.c
new file mode 100644
index 0000000..a5e3ab1
--- /dev/null
+++ b/arch/csky/kernel/setup.c
@@ -0,0 +1,162 @@
+// SPDX-License-Identifier: GPL-2.0
+// Copyright (C) 2018 Hangzhou C-SKY Microsystems co.,ltd.
+
+#include <linux/console.h>
+#include <linux/memblock.h>
+#include <linux/bootmem.h>
+#include <linux/initrd.h>
+#include <linux/of.h>
+#include <linux/of_fdt.h>
+#include <linux/start_kernel.h>
+#include <linux/dma-contiguous.h>
+#include <linux/screen_info.h>
+#include <asm/sections.h>
+#include <asm/mmu_context.h>
+#include <asm/pgalloc.h>
+
+#ifdef CONFIG_DUMMY_CONSOLE
+struct screen_info screen_info = {
+ .orig_video_lines = 30,
+ .orig_video_cols = 80,
+ .orig_video_mode = 0,
+ .orig_video_ega_bx = 0,
+ .orig_video_isVGA = 1,
+ .orig_video_points = 8
+};
+#endif
+
+phys_addr_t __init_memblock memblock_end_of_REG0(void)
+{
+ return (memblock.memory.regions[0].base +
+ memblock.memory.regions[0].size);
+}
+
+phys_addr_t __init_memblock memblock_start_of_REG1(void)
+{
+ return memblock.memory.regions[1].base;
+}
+
+size_t __init_memblock memblock_size_of_REG1(void)
+{
+ return memblock.memory.regions[1].size;
+}
+
+static void __init csky_memblock_init(void)
+{
+ unsigned long zone_size[MAX_NR_ZONES];
+ unsigned long zhole_size[MAX_NR_ZONES];
+ signed long size;
+
+ memblock_reserve(__pa(_stext), _end - _stext);
+#ifdef CONFIG_BLK_DEV_INITRD
+ memblock_reserve(__pa(initrd_start), initrd_end - initrd_start);
+#endif
+
+ early_init_fdt_reserve_self();
+ early_init_fdt_scan_reserved_mem();
+
+ memblock_dump_all();
+
+ memset(zone_size, 0, sizeof(zone_size));
+ memset(zhole_size, 0, sizeof(zhole_size));
+
+ min_low_pfn = PFN_UP(memblock_start_of_DRAM());
+ max_pfn = PFN_DOWN(memblock_end_of_DRAM());
+
+ max_low_pfn = PFN_UP(memblock_end_of_REG0());
+ if (max_low_pfn == 0)
+ max_low_pfn = max_pfn;
+
+ size = max_pfn - min_low_pfn;
+
+ if (memblock.memory.cnt > 1) {
+ zone_size[ZONE_NORMAL] =
+ PFN_DOWN(memblock_start_of_REG1()) - min_low_pfn;
+ zhole_size[ZONE_NORMAL] =
+ PFN_DOWN(memblock_start_of_REG1()) - max_low_pfn;
+ } else {
+ if (size <= PFN_DOWN(LOWMEM_LIMIT - PHYS_OFFSET_OFFSET))
+ zone_size[ZONE_NORMAL] = max_pfn - min_low_pfn;
+ else {
+ zone_size[ZONE_NORMAL] =
+ PFN_DOWN(LOWMEM_LIMIT - PHYS_OFFSET_OFFSET);
+ max_low_pfn = min_low_pfn + zone_size[ZONE_NORMAL];
+ }
+ }
+
+#ifdef CONFIG_HIGHMEM
+ size = 0;
+ if (memblock.memory.cnt > 1) {
+ size = PFN_DOWN(memblock_size_of_REG1());
+ highstart_pfn = PFN_DOWN(memblock_start_of_REG1());
+ } else {
+ size = max_pfn - min_low_pfn -
+ PFN_DOWN(LOWMEM_LIMIT - PHYS_OFFSET_OFFSET);
+ highstart_pfn = min_low_pfn +
+ PFN_DOWN(LOWMEM_LIMIT - PHYS_OFFSET_OFFSET);
+ }
+
+ if (size > 0)
+ zone_size[ZONE_HIGHMEM] = size;
+
+ highend_pfn = max_pfn;
+#endif
+ memblock_set_current_limit(PFN_PHYS(max_low_pfn));
+
+ dma_contiguous_reserve(0);
+
+ free_area_init_node(0, zone_size, min_low_pfn, zhole_size);
+}
+
+void __init setup_arch(char **cmdline_p)
+{
+ *cmdline_p = boot_command_line;
+
+ console_verbose();
+
+ pr_info("Phys. mem: %ldMB\n",
+ (unsigned long) memblock_phys_mem_size()/1024/1024);
+
+ init_mm.start_code = (unsigned long) _stext;
+ init_mm.end_code = (unsigned long) _etext;
+ init_mm.end_data = (unsigned long) _edata;
+ init_mm.brk = (unsigned long) _end;
+
+ parse_early_param();
+
+ csky_memblock_init();
+
+ unflatten_and_copy_device_tree();
+
+#ifdef CONFIG_SMP
+ setup_smp();
+#endif
+
+ sparse_init();
+
+#ifdef CONFIG_HIGHMEM
+ kmap_init();
+#endif
+
+#if defined(CONFIG_VT) && defined(CONFIG_DUMMY_CONSOLE)
+ conswitchp = &dummy_con;
+#endif
+}
+
+asmlinkage __visible void __init csky_start(unsigned int unused, void *param)
+{
+ /* Clean up bss section */
+ memset(__bss_start, 0, __bss_stop - __bss_start);
+
+ pre_trap_init();
+ pre_mmu_init();
+
+ if (param == NULL)
+ early_init_dt_scan(__dtb_start);
+ else
+ early_init_dt_scan(param);
+
+ start_kernel();
+
+ asm volatile("br .\n");
+}
diff --git a/arch/csky/kernel/vmlinux.lds.S b/arch/csky/kernel/vmlinux.lds.S
new file mode 100644
index 0000000..ae7961b
--- /dev/null
+++ b/arch/csky/kernel/vmlinux.lds.S
@@ -0,0 +1,66 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+
+#include <asm/vmlinux.lds.h>
+#include <asm/page.h>
+
+OUTPUT_ARCH(csky)
+ENTRY(_start)
+
+#ifndef __cskyBE__
+jiffies = jiffies_64;
+#else
+jiffies = jiffies_64 + 4;
+#endif
+
+#define VBR_BASE \
+ . = ALIGN(1024); \
+ vec_base = .; \
+ . += 512;
+
+SECTIONS
+{
+ . = PAGE_OFFSET + PHYS_OFFSET_OFFSET;
+
+ _stext = .;
+ __init_begin = .;
+ HEAD_TEXT_SECTION
+ INIT_TEXT_SECTION(PAGE_SIZE)
+ INIT_DATA_SECTION(PAGE_SIZE)
+ PERCPU_SECTION(L1_CACHE_BYTES)
+ . = ALIGN(PAGE_SIZE);
+ __init_end = .;
+
+ .text : AT(ADDR(.text) - LOAD_OFFSET) {
+ _text = .;
+ IRQENTRY_TEXT
+ SOFTIRQENTRY_TEXT
+ TEXT_TEXT
+ SCHED_TEXT
+ CPUIDLE_TEXT
+ LOCK_TEXT
+ KPROBES_TEXT
+ *(.fixup)
+ *(.gnu.warning)
+ } = 0
+ _etext = .;
+
+ /* __init_begin __init_end must be page aligned for free_initmem */
+ . = ALIGN(PAGE_SIZE);
+
+
+ _sdata = .;
+ RO_DATA_SECTION(PAGE_SIZE)
+ RW_DATA_SECTION(L1_CACHE_BYTES, PAGE_SIZE, THREAD_SIZE)
+ _edata = .;
+
+ NOTES
+ EXCEPTION_TABLE(L1_CACHE_BYTES)
+ BSS_SECTION(L1_CACHE_BYTES, PAGE_SIZE, L1_CACHE_BYTES)
+ VBR_BASE
+ _end = . ;
+
+ STABS_DEBUG
+ DWARF_DEBUG
+
+ DISCARDS
+}
--
2.7.4
^ permalink raw reply related [flat|nested] 68+ messages in thread* Re: [PATCH V9 03/21] csky: Kernel booting
2018-10-16 2:58 ` [PATCH V9 03/21] csky: Kernel booting Guo Ren
@ 2018-10-17 14:58 ` Arnd Bergmann
0 siblings, 0 replies; 68+ messages in thread
From: Arnd Bergmann @ 2018-10-17 14:58 UTC (permalink / raw)
To: Guo Ren
Cc: Andrew Morton, Daniel Lezcano, David Miller, gregkh,
Christoph Hellwig, Marc Zyngier, Mark Rutland, Peter Zijlstra,
Rob Herring, Thomas Gleixner, Linux Kernel Mailing List,
linux-arch, DTML, Rob Herring, c-sky_gcc_upstream
On Tue, Oct 16, 2018 at 4:59 AM Guo Ren <ren_guo@c-sky.com> wrote:
>
> This patch add boot code. Thx boot params is all in dtb and it's
> the only way to let kernel get bootloader param information.
>
> Signed-off-by: Guo Ren <ren_guo@c-sky.com>
> Cc: Arnd Bergmann <arnd@arndb.de>
> ---
> Changelog:
> - Use built-in dtb when dtb param is NULL.
> - Add dummy console for allmodconfig.
> ---
Reviewed-by: Arnd Bergmann <arnd@arndb.de>
^ permalink raw reply [flat|nested] 68+ messages in thread
* [PATCH V9 04/21] csky: Exception handling and mm-fault
2018-10-16 2:58 [PATCH V9 00/21] C-SKY(csky) Linux Kernel Port Guo Ren
` (2 preceding siblings ...)
2018-10-16 2:58 ` [PATCH V9 03/21] csky: Kernel booting Guo Ren
@ 2018-10-16 2:58 ` Guo Ren
2018-10-17 14:59 ` Arnd Bergmann
2018-10-16 2:58 ` [PATCH V9 05/21] csky: System Call Guo Ren
` (17 subsequent siblings)
21 siblings, 1 reply; 68+ messages in thread
From: Guo Ren @ 2018-10-16 2:58 UTC (permalink / raw)
To: akpm, arnd, daniel.lezcano, davem, gregkh, hch, marc.zyngier,
mark.rutland, peterz, robh, tglx
Cc: linux-kernel, linux-arch, devicetree, robh+dt,
c-sky_gcc_upstream, Guo Ren
This patch adds exception handling code, cpuinfo and mm-fault code.
Signed-off-by: Guo Ren <ren_guo@c-sky.com>
Cc: Arnd Bergmann <arnd@arndb.de>
---
Changelog:
- Fixup r15 may used by unalignment access and remove r1.
---
---
arch/csky/abiv1/alignment.c | 326 +++++++++++++++++++++++++++++++++
arch/csky/abiv1/inc/abi/entry.h | 160 ++++++++++++++++
arch/csky/abiv2/inc/abi/entry.h | 156 ++++++++++++++++
arch/csky/include/asm/traps.h | 44 +++++
arch/csky/include/asm/unistd.h | 4 +
arch/csky/kernel/cpu-probe.c | 79 ++++++++
arch/csky/kernel/entry.S | 396 ++++++++++++++++++++++++++++++++++++++++
arch/csky/kernel/traps.c | 172 +++++++++++++++++
arch/csky/mm/fault.c | 220 ++++++++++++++++++++++
9 files changed, 1557 insertions(+)
create mode 100644 arch/csky/abiv1/alignment.c
create mode 100644 arch/csky/abiv1/inc/abi/entry.h
create mode 100644 arch/csky/abiv2/inc/abi/entry.h
create mode 100644 arch/csky/include/asm/traps.h
create mode 100644 arch/csky/include/asm/unistd.h
create mode 100644 arch/csky/kernel/cpu-probe.c
create mode 100644 arch/csky/kernel/entry.S
create mode 100644 arch/csky/kernel/traps.c
create mode 100644 arch/csky/mm/fault.c
diff --git a/arch/csky/abiv1/alignment.c b/arch/csky/abiv1/alignment.c
new file mode 100644
index 0000000..60205e9
--- /dev/null
+++ b/arch/csky/abiv1/alignment.c
@@ -0,0 +1,326 @@
+// SPDX-License-Identifier: GPL-2.0
+// Copyright (C) 2018 Hangzhou C-SKY Microsystems co.,ltd.
+
+#include <linux/kernel.h>
+#include <linux/uaccess.h>
+#include <linux/ptrace.h>
+
+static int align_enable = 1;
+static int align_count;
+
+static inline uint32_t get_ptreg(struct pt_regs *regs, uint32_t rx)
+{
+ return rx == 15 ? regs->lr : *((uint32_t *)&(regs->a0) - 2 + rx);
+}
+
+static inline void put_ptreg(struct pt_regs *regs, uint32_t rx, uint32_t val)
+{
+ if (rx == 15)
+ regs->lr = val;
+ else
+ *((uint32_t *)&(regs->a0) - 2 + rx) = val;
+}
+
+/*
+ * Get byte-value from addr and set it to *valp.
+ *
+ * Success: return 0
+ * Failure: return 1
+ */
+static int ldb_asm(uint32_t addr, uint32_t *valp)
+{
+ uint32_t val;
+ int err;
+
+ if (!access_ok(VERIFY_READ, (void *)addr, 1))
+ return 1;
+
+ asm volatile (
+ "movi %0, 0\n"
+ "1:\n"
+ "ldb %1, (%2)\n"
+ "br 3f\n"
+ "2:\n"
+ "movi %0, 1\n"
+ "br 3f\n"
+ ".section __ex_table,\"a\"\n"
+ ".align 2\n"
+ ".long 1b, 2b\n"
+ ".previous\n"
+ "3:\n"
+ : "=&r"(err), "=r"(val)
+ : "r" (addr)
+ );
+
+ *valp = val;
+
+ return err;
+}
+
+/*
+ * Put byte-value to addr.
+ *
+ * Success: return 0
+ * Failure: return 1
+ */
+static int stb_asm(uint32_t addr, uint32_t val)
+{
+ int err;
+
+ if (!access_ok(VERIFY_WRITE, (void *)addr, 1))
+ return 1;
+
+ asm volatile (
+ "movi %0, 0\n"
+ "1:\n"
+ "stb %1, (%2)\n"
+ "br 3f\n"
+ "2:\n"
+ "movi %0, 1\n"
+ "br 3f\n"
+ ".section __ex_table,\"a\"\n"
+ ".align 2\n"
+ ".long 1b, 2b\n"
+ ".previous\n"
+ "3:\n"
+ : "=&r"(err)
+ : "r"(val), "r" (addr)
+ );
+
+ return err;
+}
+
+/*
+ * Get half-word from [rx + imm]
+ *
+ * Success: return 0
+ * Failure: return 1
+ */
+static int ldh_c(struct pt_regs *regs, uint32_t rz, uint32_t addr)
+{
+ uint32_t byte0, byte1;
+
+ if (ldb_asm(addr, &byte0))
+ return 1;
+ addr += 1;
+ if (ldb_asm(addr, &byte1))
+ return 1;
+
+ byte0 |= byte1 << 8;
+ put_ptreg(regs, rz, byte0);
+
+ return 0;
+}
+
+/*
+ * Store half-word to [rx + imm]
+ *
+ * Success: return 0
+ * Failure: return 1
+ */
+static int sth_c(struct pt_regs *regs, uint32_t rz, uint32_t addr)
+{
+ uint32_t byte0, byte1;
+
+ byte0 = byte1 = get_ptreg(regs, rz);
+
+ byte0 &= 0xff;
+
+ if (stb_asm(addr, byte0))
+ return 1;
+
+ addr += 1;
+ byte1 = (byte1 >> 8) & 0xff;
+ if (stb_asm(addr, byte1))
+ return 1;
+
+ return 0;
+}
+
+/*
+ * Get word from [rx + imm]
+ *
+ * Success: return 0
+ * Failure: return 1
+ */
+static int ldw_c(struct pt_regs *regs, uint32_t rz, uint32_t addr)
+{
+ uint32_t byte0, byte1, byte2, byte3;
+
+ if (ldb_asm(addr, &byte0))
+ return 1;
+
+ addr += 1;
+ if (ldb_asm(addr, &byte1))
+ return 1;
+
+ addr += 1;
+ if (ldb_asm(addr, &byte2))
+ return 1;
+
+ addr += 1;
+ if (ldb_asm(addr, &byte3))
+ return 1;
+
+ byte0 |= byte1 << 8;
+ byte0 |= byte2 << 16;
+ byte0 |= byte3 << 24;
+
+ put_ptreg(regs, rz, byte0);
+
+ return 0;
+}
+
+/*
+ * Store word to [rx + imm]
+ *
+ * Success: return 0
+ * Failure: return 1
+ */
+static int stw_c(struct pt_regs *regs, uint32_t rz, uint32_t addr)
+{
+ uint32_t byte0, byte1, byte2, byte3;
+
+ byte0 = byte1 = byte2 = byte3 = get_ptreg(regs, rz);
+
+ byte0 &= 0xff;
+
+ if (stb_asm(addr, byte0))
+ return 1;
+
+ addr += 1;
+ byte1 = (byte1 >> 8) & 0xff;
+ if (stb_asm(addr, byte1))
+ return 1;
+
+ addr += 1;
+ byte2 = (byte2 >> 16) & 0xff;
+ if (stb_asm(addr, byte2))
+ return 1;
+
+ addr += 1;
+ byte3 = (byte3 >> 24) & 0xff;
+ if (stb_asm(addr, byte3))
+ return 1;
+
+ align_count++;
+
+ return 0;
+}
+
+extern int fixup_exception(struct pt_regs *regs);
+
+#define OP_LDH 0xc000
+#define OP_STH 0xd000
+#define OP_LDW 0x8000
+#define OP_STW 0x9000
+
+void csky_alignment(struct pt_regs *regs)
+{
+ int ret;
+ uint16_t tmp;
+ uint32_t opcode = 0;
+ uint32_t rx = 0;
+ uint32_t rz = 0;
+ uint32_t imm = 0;
+ uint32_t addr = 0;
+
+ if (!user_mode(regs))
+ goto bad_area;
+
+ ret = get_user(tmp, (uint16_t *)instruction_pointer(regs));
+ if (ret) {
+ pr_err("%s get_user failed.\n", __func__);
+ goto bad_area;
+ }
+
+ opcode = (uint32_t)tmp;
+
+ rx = opcode & 0xf;
+ imm = (opcode >> 4) & 0xf;
+ rz = (opcode >> 8) & 0xf;
+ opcode &= 0xf000;
+
+ if (rx == 0 || rx == 1 || rz == 0 || rz == 1)
+ goto bad_area;
+
+ switch (opcode) {
+ case OP_LDH:
+ addr = get_ptreg(regs, rx) + (imm << 1);
+ ret = ldh_c(regs, rz, addr);
+ break;
+ case OP_LDW:
+ addr = get_ptreg(regs, rx) + (imm << 2);
+ ret = ldw_c(regs, rz, addr);
+ break;
+ case OP_STH:
+ addr = get_ptreg(regs, rx) + (imm << 1);
+ ret = sth_c(regs, rz, addr);
+ break;
+ case OP_STW:
+ addr = get_ptreg(regs, rx) + (imm << 2);
+ ret = stw_c(regs, rz, addr);
+ break;
+ }
+
+ if (ret)
+ goto bad_area;
+
+ regs->pc += 2;
+
+ return;
+
+bad_area:
+ if (!user_mode(regs)) {
+ if (fixup_exception(regs))
+ return;
+
+ bust_spinlocks(1);
+ pr_alert("%s opcode: %x, rz: %d, rx: %d, imm: %d, addr: %x.\n",
+ __func__, opcode, rz, rx, imm, addr);
+ show_regs(regs);
+ bust_spinlocks(0);
+ do_exit(SIGKILL);
+ }
+
+ force_sig_fault(SIGBUS, BUS_ADRALN, (void __user *)addr, current);
+}
+
+static struct ctl_table alignment_tbl[4] = {
+ {
+ .procname = "enable",
+ .data = &align_enable,
+ .maxlen = sizeof(align_enable),
+ .mode = 0666,
+ .proc_handler = &proc_dointvec
+ },
+ {
+ .procname = "count",
+ .data = &align_count,
+ .maxlen = sizeof(align_count),
+ .mode = 0666,
+ .proc_handler = &proc_dointvec
+ },
+ {}
+};
+
+static struct ctl_table sysctl_table[2] = {
+ {
+ .procname = "csky_alignment",
+ .mode = 0555,
+ .child = alignment_tbl},
+ {}
+};
+
+static struct ctl_path sysctl_path[2] = {
+ {.procname = "csky"},
+ {}
+};
+
+static int __init csky_alignment_init(void)
+{
+ register_sysctl_paths(sysctl_path, sysctl_table);
+ return 0;
+}
+
+arch_initcall(csky_alignment_init);
diff --git a/arch/csky/abiv1/inc/abi/entry.h b/arch/csky/abiv1/inc/abi/entry.h
new file mode 100644
index 0000000..3f3faab
--- /dev/null
+++ b/arch/csky/abiv1/inc/abi/entry.h
@@ -0,0 +1,160 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+// Copyright (C) 2018 Hangzhou C-SKY Microsystems co.,ltd.
+
+#ifndef __ASM_CSKY_ENTRY_H
+#define __ASM_CSKY_ENTRY_H
+
+#include <asm/setup.h>
+#include <abi/regdef.h>
+
+#define LSAVE_PC 8
+#define LSAVE_PSR 12
+#define LSAVE_A0 24
+#define LSAVE_A1 28
+#define LSAVE_A2 32
+#define LSAVE_A3 36
+#define LSAVE_A4 40
+#define LSAVE_A5 44
+
+#define EPC_INCREASE 2
+#define EPC_KEEP 0
+
+.macro USPTOKSP
+ mtcr sp, ss1
+ mfcr sp, ss0
+.endm
+
+.macro KSPTOUSP
+ mtcr sp, ss0
+ mfcr sp, ss1
+.endm
+
+.macro INCTRAP rx
+ addi \rx, EPC_INCREASE
+.endm
+
+.macro SAVE_ALL epc_inc
+ mtcr r13, ss2
+ mfcr r13, epsr
+ btsti r13, 31
+ bt 1f
+ USPTOKSP
+1:
+ subi sp, 32
+ subi sp, 32
+ subi sp, 16
+ stw r13, (sp, 12)
+
+ stw lr, (sp, 4)
+
+ mfcr lr, epc
+ movi r13, \epc_inc
+ add lr, r13
+ stw lr, (sp, 8)
+
+ mfcr lr, ss1
+ stw lr, (sp, 16)
+
+ stw a0, (sp, 20)
+ stw a0, (sp, 24)
+ stw a1, (sp, 28)
+ stw a2, (sp, 32)
+ stw a3, (sp, 36)
+
+ addi sp, 32
+ addi sp, 8
+ mfcr r13, ss2
+ stw r6, (sp)
+ stw r7, (sp, 4)
+ stw r8, (sp, 8)
+ stw r9, (sp, 12)
+ stw r10, (sp, 16)
+ stw r11, (sp, 20)
+ stw r12, (sp, 24)
+ stw r13, (sp, 28)
+ stw r14, (sp, 32)
+ stw r1, (sp, 36)
+ subi sp, 32
+ subi sp, 8
+.endm
+
+.macro RESTORE_ALL
+ psrclr ie
+ ldw lr, (sp, 4)
+ ldw a0, (sp, 8)
+ mtcr a0, epc
+ ldw a0, (sp, 12)
+ mtcr a0, epsr
+ btsti a0, 31
+ ldw a0, (sp, 16)
+ mtcr a0, ss1
+
+ ldw a0, (sp, 24)
+ ldw a1, (sp, 28)
+ ldw a2, (sp, 32)
+ ldw a3, (sp, 36)
+
+ addi sp, 32
+ addi sp, 8
+ ldw r6, (sp)
+ ldw r7, (sp, 4)
+ ldw r8, (sp, 8)
+ ldw r9, (sp, 12)
+ ldw r10, (sp, 16)
+ ldw r11, (sp, 20)
+ ldw r12, (sp, 24)
+ ldw r13, (sp, 28)
+ ldw r14, (sp, 32)
+ ldw r1, (sp, 36)
+ addi sp, 32
+ addi sp, 8
+
+ bt 1f
+ KSPTOUSP
+1:
+ rte
+.endm
+
+.macro SAVE_SWITCH_STACK
+ subi sp, 32
+ stm r8-r15, (sp)
+.endm
+
+.macro RESTORE_SWITCH_STACK
+ ldm r8-r15, (sp)
+ addi sp, 32
+.endm
+
+/* MMU registers operators. */
+.macro RD_MIR rx
+ cprcr \rx, cpcr0
+.endm
+
+.macro RD_MEH rx
+ cprcr \rx, cpcr4
+.endm
+
+.macro RD_MCIR rx
+ cprcr \rx, cpcr8
+.endm
+
+.macro RD_PGDR rx
+ cprcr \rx, cpcr29
+.endm
+
+.macro WR_MEH rx
+ cpwcr \rx, cpcr4
+.endm
+
+.macro WR_MCIR rx
+ cpwcr \rx, cpcr8
+.endm
+
+.macro SETUP_MMU rx
+ lrw \rx, PHYS_OFFSET | 0xe
+ cpwcr \rx, cpcr30
+ lrw \rx, (PHYS_OFFSET + 0x20000000) | 0xe
+ cpwcr \rx, cpcr31
+.endm
+
+#endif /* __ASM_CSKY_ENTRY_H */
diff --git a/arch/csky/abiv2/inc/abi/entry.h b/arch/csky/abiv2/inc/abi/entry.h
new file mode 100644
index 0000000..acd0521
--- /dev/null
+++ b/arch/csky/abiv2/inc/abi/entry.h
@@ -0,0 +1,156 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+// Copyright (C) 2018 Hangzhou C-SKY Microsystems co.,ltd.
+
+#ifndef __ASM_CSKY_ENTRY_H
+#define __ASM_CSKY_ENTRY_H
+
+#include <asm/setup.h>
+#include <abi/regdef.h>
+
+#define LSAVE_PC 8
+#define LSAVE_PSR 12
+#define LSAVE_A0 24
+#define LSAVE_A1 28
+#define LSAVE_A2 32
+#define LSAVE_A3 36
+
+#define EPC_INCREASE 4
+#define EPC_KEEP 0
+
+#define KSPTOUSP
+#define USPTOKSP
+
+#define usp cr<14, 1>
+
+.macro INCTRAP rx
+ addi \rx, EPC_INCREASE
+.endm
+
+.macro SAVE_ALL epc_inc
+ subi sp, 152
+ stw tls, (sp, 0)
+ stw lr, (sp, 4)
+
+ mfcr lr, epc
+ movi tls, \epc_inc
+ add lr, tls
+ stw lr, (sp, 8)
+
+ mfcr lr, epsr
+ stw lr, (sp, 12)
+ mfcr lr, usp
+ stw lr, (sp, 16)
+
+ stw a0, (sp, 20)
+ stw a0, (sp, 24)
+ stw a1, (sp, 28)
+ stw a2, (sp, 32)
+ stw a3, (sp, 36)
+
+ addi sp, 40
+ stm r4-r13, (sp)
+
+ addi sp, 40
+ stm r16-r30, (sp)
+#ifdef CONFIG_CPU_HAS_HILO
+ mfhi lr
+ stw lr, (sp, 60)
+ mflo lr
+ stw lr, (sp, 64)
+#endif
+ subi sp, 80
+.endm
+
+.macro RESTORE_ALL
+ psrclr ie
+ ldw tls, (sp, 0)
+ ldw lr, (sp, 4)
+ ldw a0, (sp, 8)
+ mtcr a0, epc
+ ldw a0, (sp, 12)
+ mtcr a0, epsr
+ ldw a0, (sp, 16)
+ mtcr a0, usp
+
+#ifdef CONFIG_CPU_HAS_HILO
+ ldw a0, (sp, 140)
+ mthi a0
+ ldw a0, (sp, 144)
+ mtlo a0
+#endif
+
+ ldw a0, (sp, 24)
+ ldw a1, (sp, 28)
+ ldw a2, (sp, 32)
+ ldw a3, (sp, 36)
+
+ addi sp, 40
+ ldm r4-r13, (sp)
+ addi sp, 40
+ ldm r16-r30, (sp)
+ addi sp, 72
+ rte
+.endm
+
+.macro SAVE_SWITCH_STACK
+ subi sp, 64
+ stm r4-r11, (sp)
+ stw r15, (sp, 32)
+ stw r16, (sp, 36)
+ stw r17, (sp, 40)
+ stw r26, (sp, 44)
+ stw r27, (sp, 48)
+ stw r28, (sp, 52)
+ stw r29, (sp, 56)
+ stw r30, (sp, 60)
+.endm
+
+.macro RESTORE_SWITCH_STACK
+ ldm r4-r11, (sp)
+ ldw r15, (sp, 32)
+ ldw r16, (sp, 36)
+ ldw r17, (sp, 40)
+ ldw r26, (sp, 44)
+ ldw r27, (sp, 48)
+ ldw r28, (sp, 52)
+ ldw r29, (sp, 56)
+ ldw r30, (sp, 60)
+ addi sp, 64
+.endm
+
+/* MMU registers operators. */
+.macro RD_MIR rx
+ mfcr \rx, cr<0, 15>
+.endm
+
+.macro RD_MEH rx
+ mfcr \rx, cr<4, 15>
+.endm
+
+.macro RD_MCIR rx
+ mfcr \rx, cr<8, 15>
+.endm
+
+.macro RD_PGDR rx
+ mfcr \rx, cr<29, 15>
+.endm
+
+.macro RD_PGDR_K rx
+ mfcr \rx, cr<28, 15>
+.endm
+
+.macro WR_MEH rx
+ mtcr \rx, cr<4, 15>
+.endm
+
+.macro WR_MCIR rx
+ mtcr \rx, cr<8, 15>
+.endm
+
+.macro SETUP_MMU rx
+ lrw \rx, PHYS_OFFSET | 0xe
+ mtcr \rx, cr<30, 15>
+ lrw \rx, (PHYS_OFFSET + 0x20000000) | 0xe
+ mtcr \rx, cr<31, 15>
+.endm
+#endif /* __ASM_CSKY_ENTRY_H */
diff --git a/arch/csky/include/asm/traps.h b/arch/csky/include/asm/traps.h
new file mode 100644
index 0000000..1c08180
--- /dev/null
+++ b/arch/csky/include/asm/traps.h
@@ -0,0 +1,44 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+// Copyright (C) 2018 Hangzhou C-SKY Microsystems co.,ltd.
+
+#ifndef __ASM_CSKY_TRAPS_H
+#define __ASM_CSKY_TRAPS_H
+
+#define VEC_RESET 0
+#define VEC_ALIGN 1
+#define VEC_ACCESS 2
+#define VEC_ZERODIV 3
+#define VEC_ILLEGAL 4
+#define VEC_PRIV 5
+#define VEC_TRACE 6
+#define VEC_BREAKPOINT 7
+#define VEC_UNRECOVER 8
+#define VEC_SOFTRESET 9
+#define VEC_AUTOVEC 10
+#define VEC_FAUTOVEC 11
+#define VEC_HWACCEL 12
+
+#define VEC_TLBMISS 14
+#define VEC_TLBMODIFIED 15
+
+#define VEC_TRAP0 16
+#define VEC_TRAP1 17
+#define VEC_TRAP2 18
+#define VEC_TRAP3 19
+
+#define VEC_TLBINVALIDL 20
+#define VEC_TLBINVALIDS 21
+
+#define VEC_PRFL 29
+#define VEC_FPE 30
+
+extern void *vec_base[];
+
+#define VEC_INIT(i, func) \
+do { \
+ vec_base[i] = (void *)func; \
+} while (0)
+
+void csky_alignment(struct pt_regs *regs);
+
+#endif /* __ASM_CSKY_TRAPS_H */
diff --git a/arch/csky/include/asm/unistd.h b/arch/csky/include/asm/unistd.h
new file mode 100644
index 0000000..2844874
--- /dev/null
+++ b/arch/csky/include/asm/unistd.h
@@ -0,0 +1,4 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+// Copyright (C) 2018 Hangzhou C-SKY Microsystems co.,ltd.
+
+#include <uapi/asm/unistd.h>
diff --git a/arch/csky/kernel/cpu-probe.c b/arch/csky/kernel/cpu-probe.c
new file mode 100644
index 0000000..5f15ca3
--- /dev/null
+++ b/arch/csky/kernel/cpu-probe.c
@@ -0,0 +1,79 @@
+// SPDX-License-Identifier: GPL-2.0
+// Copyright (C) 2018 Hangzhou C-SKY Microsystems co.,ltd.
+
+#include <linux/of.h>
+#include <linux/init.h>
+#include <linux/seq_file.h>
+#include <linux/memblock.h>
+
+#include <abi/reg_ops.h>
+
+static void percpu_print(void *arg)
+{
+ struct seq_file *m = (struct seq_file *)arg;
+ unsigned int cur, next, i;
+
+ seq_printf(m, "processor : %d\n", smp_processor_id());
+ seq_printf(m, "C-SKY CPU model : %s\n", CSKYCPU_DEF_NAME);
+
+ /* read processor id, max is 100 */
+ cur = mfcr("cr13");
+ for (i = 0; i < 100; i++) {
+ seq_printf(m, "product info[%d] : 0x%08x\n", i, cur);
+
+ next = mfcr("cr13");
+
+ /* some CPU only has one id reg */
+ if (cur == next)
+ break;
+
+ cur = next;
+
+ /* cpid index is 31-28, reset */
+ if (!(next >> 28)) {
+ while ((mfcr("cr13") >> 28) != i);
+ break;
+ }
+ }
+
+ /* CPU feature regs, setup by bootloader or gdbinit */
+ seq_printf(m, "hint (CPU funcs): 0x%08x\n", mfcr_hint());
+ seq_printf(m, "ccr (L1C & MMU): 0x%08x\n", mfcr("cr18"));
+ seq_printf(m, "ccr2 (L2C) : 0x%08x\n", mfcr_ccr2());
+ seq_printf(m, "\n");
+}
+
+static int c_show(struct seq_file *m, void *v)
+{
+ int cpu;
+
+ for_each_online_cpu(cpu)
+ smp_call_function_single(cpu, percpu_print, m, true);
+
+#ifdef CSKY_ARCH_VERSION
+ seq_printf(m, "arch-version : %s\n", CSKY_ARCH_VERSION);
+ seq_printf(m, "\n");
+#endif
+
+ return 0;
+}
+
+static void *c_start(struct seq_file *m, loff_t *pos)
+{
+ return *pos < 1 ? (void *)1 : NULL;
+}
+
+static void *c_next(struct seq_file *m, void *v, loff_t *pos)
+{
+ ++*pos;
+ return NULL;
+}
+
+static void c_stop(struct seq_file *m, void *v) {}
+
+const struct seq_operations cpuinfo_op = {
+ .start = c_start,
+ .next = c_next,
+ .stop = c_stop,
+ .show = c_show,
+};
diff --git a/arch/csky/kernel/entry.S b/arch/csky/kernel/entry.S
new file mode 100644
index 0000000..79f92b8
--- /dev/null
+++ b/arch/csky/kernel/entry.S
@@ -0,0 +1,396 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+// Copyright (C) 2018 Hangzhou C-SKY Microsystems co.,ltd.
+
+#include <linux/linkage.h>
+#include <abi/entry.h>
+#include <abi/pgtable-bits.h>
+#include <asm/errno.h>
+#include <asm/setup.h>
+#include <asm/unistd.h>
+#include <asm/asm-offsets.h>
+#include <linux/threads.h>
+#include <asm/setup.h>
+#include <asm/page.h>
+#include <asm/thread_info.h>
+
+#define PTE_INDX_MSK 0xffc
+#define PTE_INDX_SHIFT 10
+#define _PGDIR_SHIFT 22
+
+.macro tlbop_begin name, val0, val1, val2
+ENTRY(csky_\name)
+ mtcr a3, ss2
+ mtcr r6, ss3
+ mtcr a2, ss4
+
+ RD_PGDR r6
+ RD_MEH a3
+#ifdef CONFIG_CPU_HAS_TLBI
+ tlbi.vaas a3
+ sync.is
+
+ btsti a3, 31
+ bf 1f
+ RD_PGDR_K r6
+1:
+#else
+ bgeni a2, 31
+ WR_MCIR a2
+ bgeni a2, 25
+ WR_MCIR a2
+#endif
+ bclri r6, 0
+ lrw a2, PHYS_OFFSET
+ subu r6, a2
+ bseti r6, 31
+
+ mov a2, a3
+ lsri a2, _PGDIR_SHIFT
+ lsli a2, 2
+ addu r6, a2
+ ldw r6, (r6)
+
+ lrw a2, PHYS_OFFSET
+ subu r6, a2
+ bseti r6, 31
+
+ lsri a3, PTE_INDX_SHIFT
+ lrw a2, PTE_INDX_MSK
+ and a3, a2
+ addu r6, a3
+ ldw a3, (r6)
+
+ movi a2, (_PAGE_PRESENT | \val0)
+ and a3, a2
+ cmpne a3, a2
+ bt \name
+
+ /* First read/write the page, just update the flags */
+ ldw a3, (r6)
+ bgeni a2, PAGE_VALID_BIT
+ bseti a2, PAGE_ACCESSED_BIT
+ bseti a2, \val1
+ bseti a2, \val2
+ or a3, a2
+ stw a3, (r6)
+
+ /* Some cpu tlb-hardrefill bypass the cache */
+#ifdef CONFIG_CPU_NEED_TLBSYNC
+ movi a2, 0x22
+ bseti a2, 6
+ mtcr r6, cr22
+ mtcr a2, cr17
+ sync
+#endif
+
+ mfcr a3, ss2
+ mfcr r6, ss3
+ mfcr a2, ss4
+ rte
+\name:
+ mfcr a3, ss2
+ mfcr r6, ss3
+ mfcr a2, ss4
+ SAVE_ALL EPC_KEEP
+.endm
+.macro tlbop_end is_write
+ RD_MEH a2
+ psrset ee, ie
+ mov a0, sp
+ movi a1, \is_write
+ jbsr do_page_fault
+ movi r11_sig, 0 /* r11 = 0, Not a syscall. */
+ jmpi ret_from_exception
+.endm
+
+.text
+
+tlbop_begin tlbinvalidl, _PAGE_READ, PAGE_VALID_BIT, PAGE_ACCESSED_BIT
+tlbop_end 0
+
+tlbop_begin tlbinvalids, _PAGE_WRITE, PAGE_DIRTY_BIT, PAGE_MODIFIED_BIT
+tlbop_end 1
+
+tlbop_begin tlbmodified, _PAGE_WRITE, PAGE_DIRTY_BIT, PAGE_MODIFIED_BIT
+#ifndef CONFIG_CPU_HAS_LDSTEX
+jbsr csky_cmpxchg_fixup
+#endif
+tlbop_end 1
+
+ENTRY(csky_systemcall)
+ SAVE_ALL EPC_INCREASE
+
+ psrset ee, ie
+
+ /* Stack frame for syscall, origin call set_esp0 */
+ mov r12, sp
+
+ bmaski r11, 13
+ andn r12, r11
+ bgeni r11, 9
+ addi r11, 32
+ addu r12, r11
+ st sp, (r12, 0)
+
+ lrw r11, __NR_syscalls
+ cmphs syscallid, r11 /* Check nr of syscall */
+ bt ret_from_exception
+
+ lrw r13, sys_call_table
+ ixw r13, syscallid
+ ldw r11, (r13)
+ cmpnei r11, 0
+ bf ret_from_exception
+
+ mov r9, sp
+ bmaski r10, THREAD_SHIFT
+ andn r9, r10
+ ldw r8, (r9, TINFO_FLAGS)
+ btsti r8, TIF_SYSCALL_TRACE
+ bt 1f
+#if defined(__CSKYABIV2__)
+ subi sp, 8
+ stw r5, (sp, 0x4)
+ stw r4, (sp, 0x0)
+ jsr r11 /* Do system call */
+ addi sp, 8
+#else
+ jsr r11
+#endif
+ stw a0, (sp, LSAVE_A0) /* Save return value */
+ jmpi ret_from_exception
+
+1:
+ movi a0, 0 /* enter system call */
+ mov a1, sp /* sp = pt_regs pointer */
+ jbsr syscall_trace
+ /* Prepare args before do system call */
+ ldw a0, (sp, LSAVE_A0)
+ ldw a1, (sp, LSAVE_A1)
+ ldw a2, (sp, LSAVE_A2)
+ ldw a3, (sp, LSAVE_A3)
+#if defined(__CSKYABIV2__)
+ subi sp, 8
+ stw r5, (sp, 0x4)
+ stw r4, (sp, 0x0)
+#else
+ ldw r6, (sp, LSAVE_A4)
+ ldw r7, (sp, LSAVE_A5)
+#endif
+ jsr r11 /* Do system call */
+#if defined(__CSKYABIV2__)
+ addi sp, 8
+#endif
+ stw a0, (sp, LSAVE_A0) /* Save return value */
+
+ movi a0, 1 /* leave system call */
+ mov a1, sp /* sp = pt_regs pointer */
+ jbsr syscall_trace
+
+syscall_exit_work:
+ ld syscallid, (sp, LSAVE_PSR)
+ btsti syscallid, 31
+ bt 2f
+
+ jmpi resume_userspace
+
+2: RESTORE_ALL
+
+ENTRY(ret_from_kernel_thread)
+ jbsr schedule_tail
+ mov a0, r8
+ jsr r9
+ jbsr ret_from_exception
+
+ENTRY(ret_from_fork)
+ jbsr schedule_tail
+ mov r9, sp
+ bmaski r10, THREAD_SHIFT
+ andn r9, r10
+ ldw r8, (r9, TINFO_FLAGS)
+ movi r11_sig, 1
+ btsti r8, TIF_SYSCALL_TRACE
+ bf 3f
+ movi a0, 1
+ mov a1, sp /* sp = pt_regs pointer */
+ jbsr syscall_trace
+3:
+ jbsr ret_from_exception
+
+ret_from_exception:
+ ld syscallid, (sp, LSAVE_PSR)
+ btsti syscallid, 31
+ bt 1f
+
+ /*
+ * Load address of current->thread_info, Then get address of task_struct
+ * Get task_needreshed in task_struct
+ */
+ mov r9, sp
+ bmaski r10, THREAD_SHIFT
+ andn r9, r10
+
+resume_userspace:
+ ldw r8, (r9, TINFO_FLAGS)
+ andi r8, (_TIF_SIGPENDING | _TIF_NOTIFY_RESUME | _TIF_NEED_RESCHED)
+ cmpnei r8, 0
+ bt exit_work
+1: RESTORE_ALL
+
+exit_work:
+ mov a0, sp /* Stack address is arg[0] */
+ jbsr set_esp0 /* Call C level */
+ btsti r8, TIF_NEED_RESCHED
+ bt work_resched
+ /* If thread_info->flag is empty, RESTORE_ALL */
+ cmpnei r8, 0
+ bf 1b
+ mov a1, sp
+ mov a0, r8
+ mov a2, r11_sig /* syscall? */
+ btsti r8, TIF_SIGPENDING /* delivering a signal? */
+ /* prevent further restarts(set r11 = 0) */
+ clrt r11_sig
+ jbsr do_notify_resume /* do signals */
+ br resume_userspace
+
+work_resched:
+ lrw syscallid, ret_from_exception
+ mov r15, syscallid /* Return address in link */
+ jmpi schedule
+
+ENTRY(sys_rt_sigreturn)
+ movi r11_sig, 0
+ jmpi do_rt_sigreturn
+
+ENTRY(csky_trap)
+ SAVE_ALL EPC_KEEP
+ psrset ee
+ movi r11_sig, 0 /* r11 = 0, Not a syscall. */
+ mov a0, sp /* Push Stack pointer arg */
+ jbsr trap_c /* Call C-level trap handler */
+ jmpi ret_from_exception
+
+/*
+ * Prototype from libc for abiv1:
+ * register unsigned int __result asm("a0");
+ * asm( "trap 3" :"=r"(__result)::);
+ */
+ENTRY(csky_get_tls)
+ USPTOKSP
+
+ /* increase epc for continue */
+ mfcr a0, epc
+ INCTRAP a0
+ mtcr a0, epc
+
+ /* get current task thread_info with kernel 8K stack */
+ bmaski a0, THREAD_SHIFT
+ not a0
+ subi sp, 1
+ and a0, sp
+ addi sp, 1
+
+ /* get tls */
+ ldw a0, (a0, TINFO_TP_VALUE)
+
+ KSPTOUSP
+ rte
+
+ENTRY(csky_irq)
+ SAVE_ALL EPC_KEEP
+ psrset ee
+ movi r11_sig, 0 /* r11 = 0, Not a syscall. */
+
+#ifdef CONFIG_PREEMPT
+ mov r9, sp /* Get current stack pointer */
+ bmaski r10, THREAD_SHIFT
+ andn r9, r10 /* Get thread_info */
+
+ /*
+ * Get task_struct->stack.preempt_count for current,
+ * and increase 1.
+ */
+ ldw r8, (r9, TINFO_PREEMPT)
+ addi r8, 1
+ stw r8, (r9, TINFO_PREEMPT)
+#endif
+
+ mov a0, sp
+ jbsr csky_do_IRQ
+
+#ifdef CONFIG_PREEMPT
+ subi r8, 1
+ stw r8, (r9, TINFO_PREEMPT)
+ cmpnei r8, 0
+ bt 2f
+ ldw r8, (r9, TINFO_FLAGS)
+ btsti r8, TIF_NEED_RESCHED
+ bf 2f
+1:
+ jbsr preempt_schedule_irq /* irq en/disable is done inside */
+ ldw r7, (r9, TINFO_FLAGS) /* get new tasks TI_FLAGS */
+ btsti r7, TIF_NEED_RESCHED
+ bt 1b /* go again */
+#endif
+2:
+ jmpi ret_from_exception
+
+/*
+ * a0 = prev task_struct *
+ * a1 = next task_struct *
+ * a0 = return next
+ */
+ENTRY(__switch_to)
+ lrw a3, TASK_THREAD
+ addu a3, a0
+
+ mfcr a2, psr /* Save PSR value */
+ stw a2, (a3, THREAD_SR) /* Save PSR in task struct */
+ bclri a2, 6 /* Disable interrupts */
+ mtcr a2, psr
+
+ SAVE_SWITCH_STACK
+
+ stw sp, (a3, THREAD_KSP)
+
+#ifdef CONFIG_CPU_HAS_HILO
+ lrw r10, THREAD_DSPHI
+ add r10, a3
+ mfhi r6
+ mflo r7
+ stw r6, (r10, 0) /* THREAD_DSPHI */
+ stw r7, (r10, 4) /* THREAD_DSPLO */
+ mfcr r6, cr14
+ stw r6, (r10, 8) /* THREAD_DSPCSR */
+#endif
+
+ /* Set up next process to run */
+ lrw a3, TASK_THREAD
+ addu a3, a1
+
+ ldw sp, (a3, THREAD_KSP) /* Set next kernel sp */
+
+#ifdef CONFIG_CPU_HAS_HILO
+ lrw r10, THREAD_DSPHI
+ add r10, a3
+ ldw r6, (r10, 8) /* THREAD_DSPCSR */
+ mtcr r6, cr14
+ ldw r6, (r10, 0) /* THREAD_DSPHI */
+ ldw r7, (r10, 4) /* THREAD_DSPLO */
+ mthi r6
+ mtlo r7
+#endif
+
+ ldw a2, (a3, THREAD_SR) /* Set next PSR */
+ mtcr a2, psr
+
+#if defined(__CSKYABIV2__)
+ addi r7, a1, TASK_THREAD_INFO
+ ldw tls, (r7, TINFO_TP_VALUE)
+#endif
+
+ RESTORE_SWITCH_STACK
+
+ rts
+ENDPROC(__switch_to)
diff --git a/arch/csky/kernel/traps.c b/arch/csky/kernel/traps.c
new file mode 100644
index 0000000..4d43730
--- /dev/null
+++ b/arch/csky/kernel/traps.c
@@ -0,0 +1,172 @@
+// SPDX-License-Identifier: GPL-2.0
+// Copyright (C) 2018 Hangzhou C-SKY Microsystems co.,ltd.
+
+#include <linux/sched.h>
+#include <linux/signal.h>
+#include <linux/kernel.h>
+#include <linux/mm.h>
+#include <linux/module.h>
+#include <linux/user.h>
+#include <linux/string.h>
+#include <linux/linkage.h>
+#include <linux/init.h>
+#include <linux/ptrace.h>
+#include <linux/kallsyms.h>
+#include <linux/rtc.h>
+#include <linux/uaccess.h>
+
+#include <asm/setup.h>
+#include <asm/traps.h>
+#include <asm/pgalloc.h>
+#include <asm/siginfo.h>
+
+#include <asm/mmu_context.h>
+
+#ifdef CONFIG_CPU_HAS_FPU
+#include <abi/fpu.h>
+#endif
+
+/* Defined in entry.S */
+asmlinkage void csky_trap(void);
+
+asmlinkage void csky_systemcall(void);
+asmlinkage void csky_cmpxchg(void);
+asmlinkage void csky_get_tls(void);
+asmlinkage void csky_irq(void);
+
+asmlinkage void csky_tlbinvalidl(void);
+asmlinkage void csky_tlbinvalids(void);
+asmlinkage void csky_tlbmodified(void);
+
+/* Defined in head.S */
+asmlinkage void _start_smp_secondary(void);
+
+void __init pre_trap_init(void)
+{
+ int i;
+
+ mtcr("vbr", vec_base);
+
+ for (i = 1; i < 128; i++)
+ VEC_INIT(i, csky_trap);
+}
+
+void __init trap_init(void)
+{
+ VEC_INIT(VEC_AUTOVEC, csky_irq);
+
+ /* setup trap0 trap2 trap3 */
+ VEC_INIT(VEC_TRAP0, csky_systemcall);
+ VEC_INIT(VEC_TRAP2, csky_cmpxchg);
+ VEC_INIT(VEC_TRAP3, csky_get_tls);
+
+ /* setup MMU TLB exception */
+ VEC_INIT(VEC_TLBINVALIDL, csky_tlbinvalidl);
+ VEC_INIT(VEC_TLBINVALIDS, csky_tlbinvalids);
+ VEC_INIT(VEC_TLBMODIFIED, csky_tlbmodified);
+
+#ifdef CONFIG_CPU_HAS_FPU
+ init_fpu();
+#endif
+
+#ifdef CONFIG_SMP
+ mtcr("cr<28, 0>", virt_to_phys(vec_base));
+
+ VEC_INIT(VEC_RESET, (void *)virt_to_phys(_start_smp_secondary));
+#endif
+}
+
+void die_if_kernel(char *str, struct pt_regs *regs, int nr)
+{
+ if (user_mode(regs))
+ return;
+
+ console_verbose();
+ pr_err("%s: %08x\n", str, nr);
+ show_regs(regs);
+ add_taint(TAINT_DIE, LOCKDEP_NOW_UNRELIABLE);
+ do_exit(SIGSEGV);
+}
+
+void buserr(struct pt_regs *regs)
+{
+ siginfo_t info;
+#ifdef CONFIG_CPU_CK810
+ static unsigned long prev_pc;
+
+ if ((regs->pc == prev_pc) && prev_pc != 0) {
+ prev_pc = 0;
+ } else {
+ prev_pc = regs->pc;
+ return;
+ }
+#endif
+
+ die_if_kernel("Kernel mode BUS error", regs, 0);
+
+ pr_err("User mode Bus Error\n");
+ show_regs(regs);
+
+ current->thread.esp0 = (unsigned long) regs;
+ info.si_signo = SIGSEGV;
+ info.si_errno = 0;
+ force_sig_info(SIGSEGV, &info, current);
+}
+
+#define USR_BKPT 0x1464
+asmlinkage void trap_c(struct pt_regs *regs)
+{
+ int sig;
+ unsigned long vector;
+ siginfo_t info;
+
+ vector = (mfcr("psr") >> 16) & 0xff;
+
+ switch (vector) {
+ case VEC_ZERODIV:
+ sig = SIGFPE;
+ break;
+ /* ptrace */
+ case VEC_TRACE:
+ info.si_code = TRAP_TRACE;
+ sig = SIGTRAP;
+ break;
+ case VEC_ILLEGAL:
+#ifndef CONFIG_CPU_NO_USER_BKPT
+ if (*(uint16_t *)instruction_pointer(regs) != USR_BKPT)
+#endif
+ {
+ sig = SIGILL;
+ break;
+ }
+ /* gdbserver breakpoint */
+ case VEC_TRAP1:
+ /* jtagserver breakpoint */
+ case VEC_BREAKPOINT:
+ info.si_code = TRAP_BRKPT;
+ sig = SIGTRAP;
+ break;
+ case VEC_ACCESS:
+ return buserr(regs);
+#ifdef CONFIG_CPU_NEED_SOFTALIGN
+ case VEC_ALIGN:
+ return csky_alignment(regs);
+#endif
+#ifdef CONFIG_CPU_HAS_FPU
+ case VEC_FPE:
+ return fpu_fpe(regs);
+ case VEC_PRIV:
+ if (fpu_libc_helper(regs))
+ return;
+#endif
+ default:
+ sig = SIGSEGV;
+ break;
+ }
+ send_sig(sig, current, 0);
+}
+
+asmlinkage void set_esp0(unsigned long ssp)
+{
+ current->thread.esp0 = ssp;
+}
diff --git a/arch/csky/mm/fault.c b/arch/csky/mm/fault.c
new file mode 100644
index 0000000..0c337b1
--- /dev/null
+++ b/arch/csky/mm/fault.c
@@ -0,0 +1,220 @@
+// SPDX-License-Identifier: GPL-2.0
+// Copyright (C) 2018 Hangzhou C-SKY Microsystems co.,ltd.
+
+#include <linux/signal.h>
+#include <linux/module.h>
+#include <linux/sched.h>
+#include <linux/interrupt.h>
+#include <linux/kernel.h>
+#include <linux/errno.h>
+#include <linux/string.h>
+#include <linux/types.h>
+#include <linux/ptrace.h>
+#include <linux/mman.h>
+#include <linux/mm.h>
+#include <linux/smp.h>
+#include <linux/version.h>
+#include <linux/vt_kern.h>
+#include <linux/kernel.h>
+#include <linux/extable.h>
+#include <linux/uaccess.h>
+
+#include <asm/hardirq.h>
+#include <asm/mmu_context.h>
+#include <asm/traps.h>
+#include <asm/page.h>
+
+int fixup_exception(struct pt_regs *regs)
+{
+ const struct exception_table_entry *fixup;
+
+ fixup = search_exception_tables(instruction_pointer(regs));
+ if (fixup) {
+ regs->pc = fixup->nextinsn;
+
+ return 1;
+ }
+
+ return 0;
+}
+
+/*
+ * This routine handles page faults. It determines the address,
+ * and the problem, and then passes it off to one of the appropriate
+ * routines.
+ */
+asmlinkage void do_page_fault(struct pt_regs *regs, unsigned long write,
+ unsigned long mmu_meh)
+{
+ struct vm_area_struct *vma = NULL;
+ struct task_struct *tsk = current;
+ struct mm_struct *mm = tsk->mm;
+ siginfo_t info;
+ int fault;
+ unsigned long address = mmu_meh & PAGE_MASK;
+
+ info.si_code = SEGV_MAPERR;
+
+#ifndef CONFIG_CPU_HAS_TLBI
+ /*
+ * We fault-in kernel-space virtual memory on-demand. The
+ * 'reference' page table is init_mm.pgd.
+ *
+ * NOTE! We MUST NOT take any locks for this case. We may
+ * be in an interrupt or a critical region, and should
+ * only copy the information from the master page table,
+ * nothing more.
+ */
+ if (unlikely(address >= VMALLOC_START) &&
+ unlikely(address <= VMALLOC_END)) {
+ /*
+ * Synchronize this task's top level page-table
+ * with the 'reference' page table.
+ *
+ * Do _not_ use "tsk" here. We might be inside
+ * an interrupt in the middle of a task switch..
+ */
+ int offset = __pgd_offset(address);
+ pgd_t *pgd, *pgd_k;
+ pud_t *pud, *pud_k;
+ pmd_t *pmd, *pmd_k;
+ pte_t *pte_k;
+
+ unsigned long pgd_base;
+
+ pgd_base = tlb_get_pgd();
+ pgd = (pgd_t *)pgd_base + offset;
+ pgd_k = init_mm.pgd + offset;
+
+ if (!pgd_present(*pgd_k))
+ goto no_context;
+ set_pgd(pgd, *pgd_k);
+
+ pud = (pud_t *)pgd;
+ pud_k = (pud_t *)pgd_k;
+ if (!pud_present(*pud_k))
+ goto no_context;
+
+ pmd = pmd_offset(pud, address);
+ pmd_k = pmd_offset(pud_k, address);
+ if (!pmd_present(*pmd_k))
+ goto no_context;
+ set_pmd(pmd, *pmd_k);
+
+ pte_k = pte_offset_kernel(pmd_k, address);
+ if (!pte_present(*pte_k))
+ goto no_context;
+ return;
+ }
+#endif
+ /*
+ * If we're in an interrupt or have no user
+ * context, we must not take the fault..
+ */
+ if (in_atomic() || !mm)
+ goto bad_area_nosemaphore;
+
+ down_read(&mm->mmap_sem);
+ vma = find_vma(mm, address);
+ if (!vma)
+ goto bad_area;
+ if (vma->vm_start <= address)
+ goto good_area;
+ if (!(vma->vm_flags & VM_GROWSDOWN))
+ goto bad_area;
+ if (expand_stack(vma, address))
+ goto bad_area;
+ /*
+ * Ok, we have a good vm_area for this memory access, so
+ * we can handle it..
+ */
+good_area:
+ info.si_code = SEGV_ACCERR;
+
+ if (write) {
+ if (!(vma->vm_flags & VM_WRITE))
+ goto bad_area;
+ } else {
+ if (!(vma->vm_flags & (VM_READ | VM_WRITE | VM_EXEC)))
+ goto bad_area;
+ }
+
+ /*
+ * If for any reason at all we couldn't handle the fault,
+ * make sure we exit gracefully rather than endlessly redo
+ * the fault.
+ */
+ fault = handle_mm_fault(vma, address, write ? FAULT_FLAG_WRITE : 0);
+ if (unlikely(fault & VM_FAULT_ERROR)) {
+ if (fault & VM_FAULT_OOM)
+ goto out_of_memory;
+ else if (fault & VM_FAULT_SIGBUS)
+ goto do_sigbus;
+ else if (fault & VM_FAULT_SIGSEGV)
+ goto bad_area;
+ BUG();
+ }
+ if (fault & VM_FAULT_MAJOR)
+ tsk->maj_flt++;
+ else
+ tsk->min_flt++;
+
+ up_read(&mm->mmap_sem);
+ return;
+
+ /*
+ * Something tried to access memory that isn't in our memory map..
+ * Fix it, but check if it's kernel or user first..
+ */
+bad_area:
+ up_read(&mm->mmap_sem);
+
+bad_area_nosemaphore:
+ /* User mode accesses just cause a SIGSEGV */
+ if (user_mode(regs)) {
+ tsk->thread.address = address;
+ tsk->thread.error_code = write;
+ info.si_signo = SIGSEGV;
+ info.si_errno = 0;
+ /* info.si_code has been set above */
+ info.si_addr = (void __user *) address;
+ force_sig_info(SIGSEGV, &info, tsk);
+ return;
+ }
+
+no_context:
+ /* Are we prepared to handle this kernel fault? */
+ if (fixup_exception(regs))
+ return;
+
+ /*
+ * Oops. The kernel tried to access some bad page. We'll have to
+ * terminate things with extreme prejudice.
+ */
+ bust_spinlocks(1);
+ pr_alert("Unable to %s at vaddr: %08lx, epc: %08lx\n",
+ __func__, address, regs->pc);
+ die_if_kernel("Oops", regs, write);
+
+out_of_memory:
+ /*
+ * We ran out of memory, call the OOM killer, and return the userspace
+ * (which will retry the fault, or kill us if we got oom-killed).
+ */
+ pagefault_out_of_memory();
+ return;
+
+do_sigbus:
+ up_read(&mm->mmap_sem);
+
+ /* Kernel mode? Handle exceptions or die */
+ if (!user_mode(regs))
+ goto no_context;
+
+ tsk->thread.address = address;
+ info.si_signo = SIGBUS;
+ info.si_errno = 0;
+ info.si_code = BUS_ADRERR;
+ info.si_addr = (void __user *) address;
+ force_sig_info(SIGBUS, &info, tsk);
+}
--
2.7.4
^ permalink raw reply related [flat|nested] 68+ messages in thread* Re: [PATCH V9 04/21] csky: Exception handling and mm-fault
2018-10-16 2:58 ` [PATCH V9 04/21] csky: Exception handling and mm-fault Guo Ren
@ 2018-10-17 14:59 ` Arnd Bergmann
0 siblings, 0 replies; 68+ messages in thread
From: Arnd Bergmann @ 2018-10-17 14:59 UTC (permalink / raw)
To: Guo Ren
Cc: Andrew Morton, Daniel Lezcano, David Miller, gregkh,
Christoph Hellwig, Marc Zyngier, Mark Rutland, Peter Zijlstra,
Rob Herring, Thomas Gleixner, Linux Kernel Mailing List,
linux-arch, DTML, Rob Herring, c-sky_gcc_upstream
On Tue, Oct 16, 2018 at 5:01 AM Guo Ren <ren_guo@c-sky.com> wrote:
>
> This patch adds exception handling code, cpuinfo and mm-fault code.
>
> Signed-off-by: Guo Ren <ren_guo@c-sky.com>
> Cc: Arnd Bergmann <arnd@arndb.de>
Reviewed-by: Arnd Bergmann <arnd@arndb.de>
^ permalink raw reply [flat|nested] 68+ messages in thread
* [PATCH V9 05/21] csky: System Call
2018-10-16 2:58 [PATCH V9 00/21] C-SKY(csky) Linux Kernel Port Guo Ren
` (3 preceding siblings ...)
2018-10-16 2:58 ` [PATCH V9 04/21] csky: Exception handling and mm-fault Guo Ren
@ 2018-10-16 2:58 ` Guo Ren
2018-10-17 15:02 ` Arnd Bergmann
2018-10-16 2:58 ` [PATCH V9 06/21] csky: Cache and TLB routines Guo Ren
` (16 subsequent siblings)
21 siblings, 1 reply; 68+ messages in thread
From: Guo Ren @ 2018-10-16 2:58 UTC (permalink / raw)
To: akpm, arnd, daniel.lezcano, davem, gregkh, hch, marc.zyngier,
mark.rutland, peterz, robh, tglx
Cc: linux-kernel, linux-arch, devicetree, robh+dt,
c-sky_gcc_upstream, Guo Ren
This patch adds files related to syscall.
Signed-off-by: Guo Ren <ren_guo@c-sky.com>
Cc: Arnd Bergmann <arnd@arndb.de>
---
arch/csky/include/asm/syscall.h | 71 +++++++++++++++++++++++++++++++++++++
arch/csky/include/asm/syscalls.h | 15 ++++++++
arch/csky/include/uapi/asm/unistd.h | 10 ++++++
arch/csky/kernel/syscall.c | 43 ++++++++++++++++++++++
arch/csky/kernel/syscall_table.c | 14 ++++++++
5 files changed, 153 insertions(+)
create mode 100644 arch/csky/include/asm/syscall.h
create mode 100644 arch/csky/include/asm/syscalls.h
create mode 100644 arch/csky/include/uapi/asm/unistd.h
create mode 100644 arch/csky/kernel/syscall.c
create mode 100644 arch/csky/kernel/syscall_table.c
diff --git a/arch/csky/include/asm/syscall.h b/arch/csky/include/asm/syscall.h
new file mode 100644
index 0000000..926a64a
--- /dev/null
+++ b/arch/csky/include/asm/syscall.h
@@ -0,0 +1,71 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+
+#ifndef __ASM_SYSCALL_H
+#define __ASM_SYSCALL_H
+
+#include <linux/sched.h>
+#include <linux/err.h>
+#include <abi/regdef.h>
+
+static inline int
+syscall_get_nr(struct task_struct *task, struct pt_regs *regs)
+{
+ return regs_syscallid(regs);
+}
+
+static inline void
+syscall_rollback(struct task_struct *task, struct pt_regs *regs)
+{
+ regs->a0 = regs->orig_a0;
+}
+
+static inline long
+syscall_get_error(struct task_struct *task, struct pt_regs *regs)
+{
+ unsigned long error = regs->a0;
+
+ return IS_ERR_VALUE(error) ? error : 0;
+}
+
+static inline long
+syscall_get_return_value(struct task_struct *task, struct pt_regs *regs)
+{
+ return regs->a0;
+}
+
+static inline void
+syscall_set_return_value(struct task_struct *task, struct pt_regs *regs,
+ int error, long val)
+{
+ regs->a0 = (long) error ?: val;
+}
+
+static inline void
+syscall_get_arguments(struct task_struct *task, struct pt_regs *regs,
+ unsigned int i, unsigned int n, unsigned long *args)
+{
+ BUG_ON(i + n > 6);
+ if (i == 0) {
+ args[0] = regs->orig_a0;
+ args++;
+ i++;
+ n--;
+ }
+ memcpy(args, ®s->a1 + i * sizeof(regs->a1), n * sizeof(args[0]));
+}
+
+static inline void
+syscall_set_arguments(struct task_struct *task, struct pt_regs *regs,
+ unsigned int i, unsigned int n, const unsigned long *args)
+{
+ BUG_ON(i + n > 6);
+ if (i == 0) {
+ regs->orig_a0 = args[0];
+ args++;
+ i++;
+ n--;
+ }
+ memcpy(®s->a1 + i * sizeof(regs->a1), args, n * sizeof(regs->a0));
+}
+
+#endif /* __ASM_SYSCALL_H */
diff --git a/arch/csky/include/asm/syscalls.h b/arch/csky/include/asm/syscalls.h
new file mode 100644
index 0000000..5d48e5e
--- /dev/null
+++ b/arch/csky/include/asm/syscalls.h
@@ -0,0 +1,15 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+// Copyright (C) 2018 Hangzhou C-SKY Microsystems co.,ltd.
+
+#ifndef __ASM_CSKY_SYSCALLS_H
+#define __ASM_CSKY_SYSCALLS_H
+
+#include <asm-generic/syscalls.h>
+
+long sys_cacheflush(void __user *, unsigned long, int);
+
+long sys_set_thread_area(unsigned long addr);
+
+long sys_csky_fadvise64_64(int fd, int advice, loff_t offset, loff_t len);
+
+#endif /* __ASM_CSKY_SYSCALLS_H */
diff --git a/arch/csky/include/uapi/asm/unistd.h b/arch/csky/include/uapi/asm/unistd.h
new file mode 100644
index 0000000..16ed71e
--- /dev/null
+++ b/arch/csky/include/uapi/asm/unistd.h
@@ -0,0 +1,10 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+// Copyright (C) 2018 Hangzhou C-SKY Microsystems co.,ltd.
+
+#define __ARCH_WANT_SYS_CLONE
+#include <asm-generic/unistd.h>
+
+#define __NR_set_thread_area (__NR_arch_specific_syscall + 0)
+__SYSCALL(__NR_set_thread_area, sys_set_thread_area)
+#define __NR_cacheflush (__NR_arch_specific_syscall + 4)
+__SYSCALL(__NR_cacheflush, sys_cacheflush)
diff --git a/arch/csky/kernel/syscall.c b/arch/csky/kernel/syscall.c
new file mode 100644
index 0000000..3d30e58
--- /dev/null
+++ b/arch/csky/kernel/syscall.c
@@ -0,0 +1,43 @@
+// SPDX-License-Identifier: GPL-2.0
+// Copyright (C) 2018 Hangzhou C-SKY Microsystems co.,ltd.
+
+#include <linux/syscalls.h>
+
+SYSCALL_DEFINE1(set_thread_area, unsigned long, addr)
+{
+ struct thread_info *ti = task_thread_info(current);
+ struct pt_regs *reg = current_pt_regs();
+
+ reg->tls = addr;
+ ti->tp_value = addr;
+
+ return 0;
+}
+
+SYSCALL_DEFINE6(mmap2,
+ unsigned long, addr,
+ unsigned long, len,
+ unsigned long, prot,
+ unsigned long, flags,
+ unsigned long, fd,
+ off_t, offset)
+{
+ if (unlikely(offset & (~PAGE_MASK >> 12)))
+ return -EINVAL;
+
+ return ksys_mmap_pgoff(addr, len, prot, flags, fd,
+ offset >> (PAGE_SHIFT - 12));
+}
+
+/*
+ * for abiv1 the 64bits args should be even th, So we need mov the advice
+ * forward.
+ */
+SYSCALL_DEFINE4(csky_fadvise64_64,
+ int, fd,
+ int, advice,
+ loff_t, offset,
+ loff_t, len)
+{
+ return ksys_fadvise64_64(fd, offset, len, advice);
+}
diff --git a/arch/csky/kernel/syscall_table.c b/arch/csky/kernel/syscall_table.c
new file mode 100644
index 0000000..a0c238c
--- /dev/null
+++ b/arch/csky/kernel/syscall_table.c
@@ -0,0 +1,14 @@
+// SPDX-License-Identifier: GPL-2.0
+// Copyright (C) 2018 Hangzhou C-SKY Microsystems co.,ltd.
+
+#include <linux/syscalls.h>
+#include <asm/syscalls.h>
+
+#undef __SYSCALL
+#define __SYSCALL(nr, call)[nr] = (call),
+
+#define sys_fadvise64_64 sys_csky_fadvise64_64
+void * const sys_call_table[__NR_syscalls] __page_aligned_data = {
+ [0 ... __NR_syscalls - 1] = sys_ni_syscall,
+#include <asm/unistd.h>
+};
--
2.7.4
^ permalink raw reply related [flat|nested] 68+ messages in thread* Re: [PATCH V9 05/21] csky: System Call
2018-10-16 2:58 ` [PATCH V9 05/21] csky: System Call Guo Ren
@ 2018-10-17 15:02 ` Arnd Bergmann
2018-10-18 2:02 ` Guo Ren
0 siblings, 1 reply; 68+ messages in thread
From: Arnd Bergmann @ 2018-10-17 15:02 UTC (permalink / raw)
To: Guo Ren
Cc: Andrew Morton, Daniel Lezcano, David Miller, gregkh,
Christoph Hellwig, Marc Zyngier, Mark Rutland, Peter Zijlstra,
Rob Herring, Thomas Gleixner, Linux Kernel Mailing List,
linux-arch, DTML, Rob Herring, c-sky_gcc_upstream
On Tue, Oct 16, 2018 at 5:01 AM Guo Ren <ren_guo@c-sky.com> wrote:
>
> This patch adds files related to syscall.
>
> Signed-off-by: Guo Ren <ren_guo@c-sky.com>
> Cc: Arnd Bergmann <arnd@arndb.de>
Reviewed-by: Arnd Bergmann <arnd@arndb.de>
I had a detailed look at this one and provided lots of comments, however
one tiny thing escaped me:
> +#define __NR_set_thread_area (__NR_arch_specific_syscall + 0)
> +__SYSCALL(__NR_set_thread_area, sys_set_thread_area)
> +#define __NR_cacheflush (__NR_arch_specific_syscall + 4)
> +__SYSCALL(__NR_cacheflush, sys_cacheflush)
It's odd that the second one uses __NR_arch_specific_syscall + 4 rather
than +1. If you can still change that changing too much other stuff,
that might avoid some head-scratching later. If it causes problems,
just leave it and maybe add a comment saying that the numbers
inbetween are still available.
Arnd
^ permalink raw reply [flat|nested] 68+ messages in thread* Re: [PATCH V9 05/21] csky: System Call
2018-10-17 15:02 ` Arnd Bergmann
@ 2018-10-18 2:02 ` Guo Ren
0 siblings, 0 replies; 68+ messages in thread
From: Guo Ren @ 2018-10-18 2:02 UTC (permalink / raw)
To: Arnd Bergmann
Cc: Andrew Morton, Daniel Lezcano, David Miller, gregkh,
Christoph Hellwig, Marc Zyngier, Mark Rutland, Peter Zijlstra,
Rob Herring, Thomas Gleixner, Linux Kernel Mailing List,
linux-arch, DTML, Rob Herring, c-sky_gcc_upstream
On Wed, Oct 17, 2018 at 05:02:33PM +0200, Arnd Bergmann wrote:
> On Tue, Oct 16, 2018 at 5:01 AM Guo Ren <ren_guo@c-sky.com> wrote:
> >
> > This patch adds files related to syscall.
> >
> > Signed-off-by: Guo Ren <ren_guo@c-sky.com>
> > Cc: Arnd Bergmann <arnd@arndb.de>
>
> Reviewed-by: Arnd Bergmann <arnd@arndb.de>
>
> I had a detailed look at this one and provided lots of comments, however
> one tiny thing escaped me:
>
> > +#define __NR_set_thread_area (__NR_arch_specific_syscall + 0)
> > +__SYSCALL(__NR_set_thread_area, sys_set_thread_area)
> > +#define __NR_cacheflush (__NR_arch_specific_syscall + 4)
> > +__SYSCALL(__NR_cacheflush, sys_cacheflush)
>
> It's odd that the second one uses __NR_arch_specific_syscall + 4 rather
> than +1. If you can still change that changing too much other stuff,
> that might avoid some head-scratching later. If it causes problems,
> just leave it and maybe add a comment saying that the numbers
> inbetween are still available.
Change to (__NR_arch_specific_syscall + 1) is ok for now.
Best Regards
Guo Ren
^ permalink raw reply [flat|nested] 68+ messages in thread
* [PATCH V9 06/21] csky: Cache and TLB routines
2018-10-16 2:58 [PATCH V9 00/21] C-SKY(csky) Linux Kernel Port Guo Ren
` (4 preceding siblings ...)
2018-10-16 2:58 ` [PATCH V9 05/21] csky: System Call Guo Ren
@ 2018-10-16 2:58 ` Guo Ren
2018-10-17 15:08 ` Arnd Bergmann
2018-10-16 2:58 ` [PATCH V9 07/21] csky: MMU and page table management Guo Ren
` (15 subsequent siblings)
21 siblings, 1 reply; 68+ messages in thread
From: Guo Ren @ 2018-10-16 2:58 UTC (permalink / raw)
To: akpm, arnd, daniel.lezcano, davem, gregkh, hch, marc.zyngier,
mark.rutland, peterz, robh, tglx
Cc: linux-kernel, linux-arch, devicetree, robh+dt,
c-sky_gcc_upstream, Guo Ren
This patch adds cache and tlb sync codes for abiv1 & abiv2.
Signed-off-by: Guo Ren <ren_guo@c-sky.com>
Cc: Arnd Bergmann <arnd@arndb.de>
---
Changelog:
- fixup module compile error for EXPORT_SYMBOL(cache_wbinv_range).
- tlb.h & cacheflush.h:
fix flush_cache_range and tlb_start_vma of abiv1 in
flush_cache_range(vma, ...). cache_wbinv_range() couldn't deal with
vma->mm's asid for cache_flush_line and we use cache_wbinv_all() first.
Wwe'll improve with cache_flush(vma, start, end) in future.
For tlb_start_vma, we make it the same as other arch.
- remove the abi/tlb.h
---
---
arch/csky/abiv1/cacheflush.c | 52 ++++++++
arch/csky/abiv1/inc/abi/cacheflush.h | 49 ++++++++
arch/csky/abiv2/cacheflush.c | 60 ++++++++++
arch/csky/abiv2/inc/abi/cacheflush.h | 46 +++++++
arch/csky/include/asm/barrier.h | 49 ++++++++
arch/csky/include/asm/cache.h | 30 +++++
arch/csky/include/asm/cacheflush.h | 9 ++
arch/csky/include/asm/io.h | 24 ++++
arch/csky/include/asm/tlb.h | 25 ++++
arch/csky/include/asm/tlbflush.h | 25 ++++
arch/csky/include/uapi/asm/cachectl.h | 13 ++
arch/csky/mm/cachev1.c | 126 +++++++++++++++++++
arch/csky/mm/cachev2.c | 79 ++++++++++++
arch/csky/mm/syscache.c | 32 +++++
arch/csky/mm/tlb.c | 219 ++++++++++++++++++++++++++++++++++
15 files changed, 838 insertions(+)
create mode 100644 arch/csky/abiv1/cacheflush.c
create mode 100644 arch/csky/abiv1/inc/abi/cacheflush.h
create mode 100644 arch/csky/abiv2/cacheflush.c
create mode 100644 arch/csky/abiv2/inc/abi/cacheflush.h
create mode 100644 arch/csky/include/asm/barrier.h
create mode 100644 arch/csky/include/asm/cache.h
create mode 100644 arch/csky/include/asm/cacheflush.h
create mode 100644 arch/csky/include/asm/io.h
create mode 100644 arch/csky/include/asm/tlb.h
create mode 100644 arch/csky/include/asm/tlbflush.h
create mode 100644 arch/csky/include/uapi/asm/cachectl.h
create mode 100644 arch/csky/mm/cachev1.c
create mode 100644 arch/csky/mm/cachev2.c
create mode 100644 arch/csky/mm/syscache.c
create mode 100644 arch/csky/mm/tlb.c
diff --git a/arch/csky/abiv1/cacheflush.c b/arch/csky/abiv1/cacheflush.c
new file mode 100644
index 0000000..10af8b6
--- /dev/null
+++ b/arch/csky/abiv1/cacheflush.c
@@ -0,0 +1,52 @@
+// SPDX-License-Identifier: GPL-2.0
+// Copyright (C) 2018 Hangzhou C-SKY Microsystems co.,ltd.
+
+#include <linux/kernel.h>
+#include <linux/mm.h>
+#include <linux/fs.h>
+#include <linux/syscalls.h>
+#include <linux/spinlock.h>
+#include <asm/page.h>
+#include <asm/cache.h>
+#include <asm/cacheflush.h>
+#include <asm/cachectl.h>
+
+void flush_dcache_page(struct page *page)
+{
+ struct address_space *mapping = page_mapping(page);
+ unsigned long addr;
+
+ if (mapping && !mapping_mapped(mapping)) {
+ set_bit(PG_arch_1, &(page)->flags);
+ return;
+ }
+
+ /*
+ * We could delay the flush for the !page_mapping case too. But that
+ * case is for exec env/arg pages and those are %99 certainly going to
+ * get faulted into the tlb (and thus flushed) anyways.
+ */
+ addr = (unsigned long) page_address(page);
+ dcache_wb_range(addr, addr + PAGE_SIZE);
+}
+
+void update_mmu_cache(struct vm_area_struct *vma, unsigned long address,
+ pte_t *pte)
+{
+ unsigned long addr;
+ struct page *page;
+ unsigned long pfn;
+
+ pfn = pte_pfn(*pte);
+ if (unlikely(!pfn_valid(pfn)))
+ return;
+
+ page = pfn_to_page(pfn);
+ addr = (unsigned long) page_address(page);
+
+ if (vma->vm_flags & VM_EXEC ||
+ pages_do_alias(addr, address & PAGE_MASK))
+ cache_wbinv_all();
+
+ clear_bit(PG_arch_1, &(page)->flags);
+}
diff --git a/arch/csky/abiv1/inc/abi/cacheflush.h b/arch/csky/abiv1/inc/abi/cacheflush.h
new file mode 100644
index 0000000..5f663ae
--- /dev/null
+++ b/arch/csky/abiv1/inc/abi/cacheflush.h
@@ -0,0 +1,49 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+// Copyright (C) 2018 Hangzhou C-SKY Microsystems co.,ltd.
+
+#ifndef __ABI_CSKY_CACHEFLUSH_H
+#define __ABI_CSKY_CACHEFLUSH_H
+
+#include <linux/compiler.h>
+#include <asm/string.h>
+#include <asm/cache.h>
+
+#define ARCH_IMPLEMENTS_FLUSH_DCACHE_PAGE 1
+extern void flush_dcache_page(struct page *);
+
+#define flush_cache_mm(mm) cache_wbinv_all()
+#define flush_cache_page(vma, page, pfn) cache_wbinv_all()
+#define flush_cache_dup_mm(mm) cache_wbinv_all()
+
+/*
+ * if (current_mm != vma->mm) cache_wbinv_range(start, end) will be broken.
+ * Use cache_wbinv_all() here and need to be improved in future.
+ */
+#define flush_cache_range(vma, start, end) cache_wbinv_all()
+#define flush_cache_vmap(start, end) cache_wbinv_range(start, end)
+#define flush_cache_vunmap(start, end) cache_wbinv_range(start, end)
+
+#define flush_icache_page(vma, page) cache_wbinv_all()
+#define flush_icache_range(start, end) cache_wbinv_range(start, end)
+
+#define flush_icache_user_range(vma, pg, adr, len) \
+ cache_wbinv_range(adr, adr + len)
+
+#define copy_from_user_page(vma, page, vaddr, dst, src, len) \
+do { \
+ cache_wbinv_all(); \
+ memcpy(dst, src, len); \
+ cache_wbinv_all(); \
+} while (0)
+
+#define copy_to_user_page(vma, page, vaddr, dst, src, len) \
+do { \
+ cache_wbinv_all(); \
+ memcpy(dst, src, len); \
+ cache_wbinv_all(); \
+} while (0)
+
+#define flush_dcache_mmap_lock(mapping) do {} while (0)
+#define flush_dcache_mmap_unlock(mapping) do {} while (0)
+
+#endif /* __ABI_CSKY_CACHEFLUSH_H */
diff --git a/arch/csky/abiv2/cacheflush.c b/arch/csky/abiv2/cacheflush.c
new file mode 100644
index 0000000..d22c95f
--- /dev/null
+++ b/arch/csky/abiv2/cacheflush.c
@@ -0,0 +1,60 @@
+// SPDX-License-Identifier: GPL-2.0
+// Copyright (C) 2018 Hangzhou C-SKY Microsystems co.,ltd.
+
+#include <linux/cache.h>
+#include <linux/highmem.h>
+#include <linux/mm.h>
+#include <asm/cache.h>
+
+void flush_icache_page(struct vm_area_struct *vma, struct page *page)
+{
+ unsigned long start;
+
+ start = (unsigned long) kmap_atomic(page);
+
+ cache_wbinv_range(start, start + PAGE_SIZE);
+
+ kunmap_atomic((void *)start);
+}
+
+void flush_icache_user_range(struct vm_area_struct *vma, struct page *page,
+ unsigned long vaddr, int len)
+{
+ unsigned long kaddr;
+
+ kaddr = (unsigned long) kmap_atomic(page) + (vaddr & ~PAGE_MASK);
+
+ cache_wbinv_range(kaddr, kaddr + len);
+
+ kunmap_atomic((void *)kaddr);
+}
+
+void update_mmu_cache(struct vm_area_struct *vma, unsigned long address,
+ pte_t *pte)
+{
+ unsigned long addr, pfn;
+ struct page *page;
+ void *va;
+
+ if (!(vma->vm_flags & VM_EXEC))
+ return;
+
+ pfn = pte_pfn(*pte);
+ if (unlikely(!pfn_valid(pfn)))
+ return;
+
+ page = pfn_to_page(pfn);
+ if (page == ZERO_PAGE(0))
+ return;
+
+ va = page_address(page);
+ addr = (unsigned long) va;
+
+ if (va == NULL && PageHighMem(page))
+ addr = (unsigned long) kmap_atomic(page);
+
+ cache_wbinv_range(addr, addr + PAGE_SIZE);
+
+ if (va == NULL && PageHighMem(page))
+ kunmap_atomic((void *) addr);
+}
diff --git a/arch/csky/abiv2/inc/abi/cacheflush.h b/arch/csky/abiv2/inc/abi/cacheflush.h
new file mode 100644
index 0000000..b8db5e0
--- /dev/null
+++ b/arch/csky/abiv2/inc/abi/cacheflush.h
@@ -0,0 +1,46 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+
+#ifndef __ABI_CSKY_CACHEFLUSH_H
+#define __ABI_CSKY_CACHEFLUSH_H
+
+/* Keep includes the same across arches. */
+#include <linux/mm.h>
+
+/*
+ * The cache doesn't need to be flushed when TLB entries change when
+ * the cache is mapped to physical memory, not virtual memory
+ */
+#define flush_cache_all() do { } while (0)
+#define flush_cache_mm(mm) do { } while (0)
+#define flush_cache_dup_mm(mm) do { } while (0)
+
+#define flush_cache_range(vma, start, end) \
+ do { \
+ if (vma->vm_flags & VM_EXEC) \
+ icache_inv_all(); \
+ } while (0)
+
+#define flush_cache_page(vma, vmaddr, pfn) do { } while (0)
+#define ARCH_IMPLEMENTS_FLUSH_DCACHE_PAGE 0
+#define flush_dcache_page(page) do { } while (0)
+#define flush_dcache_mmap_lock(mapping) do { } while (0)
+#define flush_dcache_mmap_unlock(mapping) do { } while (0)
+
+#define flush_icache_range(start, end) cache_wbinv_range(start, end)
+
+void flush_icache_page(struct vm_area_struct *vma, struct page *page);
+void flush_icache_user_range(struct vm_area_struct *vma, struct page *page,
+ unsigned long vaddr, int len);
+
+#define flush_cache_vmap(start, end) do { } while (0)
+#define flush_cache_vunmap(start, end) do { } while (0)
+
+#define copy_to_user_page(vma, page, vaddr, dst, src, len) \
+do { \
+ memcpy(dst, src, len); \
+ cache_wbinv_range((unsigned long)dst, (unsigned long)dst + len); \
+} while (0)
+#define copy_from_user_page(vma, page, vaddr, dst, src, len) \
+ memcpy(dst, src, len)
+
+#endif /* __ABI_CSKY_CACHEFLUSH_H */
diff --git a/arch/csky/include/asm/barrier.h b/arch/csky/include/asm/barrier.h
new file mode 100644
index 0000000..476eb78
--- /dev/null
+++ b/arch/csky/include/asm/barrier.h
@@ -0,0 +1,49 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+// Copyright (C) 2018 Hangzhou C-SKY Microsystems co.,ltd.
+
+#ifndef __ASM_CSKY_BARRIER_H
+#define __ASM_CSKY_BARRIER_H
+
+#ifndef __ASSEMBLY__
+
+#define nop() asm volatile ("nop\n":::"memory")
+
+/*
+ * sync: completion barrier
+ * sync.s: completion barrier and shareable to other cores
+ * sync.i: completion barrier with flush cpu pipeline
+ * sync.is: completion barrier with flush cpu pipeline and shareable to
+ * other cores
+ *
+ * bar.brwarw: ordering barrier for all load/store instructions before it
+ * bar.brwarws: ordering barrier for all load/store instructions before it
+ * and shareable to other cores
+ * bar.brar: ordering barrier for all load instructions before it
+ * bar.brars: ordering barrier for all load instructions before it
+ * and shareable to other cores
+ * bar.bwaw: ordering barrier for all store instructions before it
+ * bar.bwaws: ordering barrier for all store instructions before it
+ * and shareable to other cores
+ */
+
+#ifdef CONFIG_CPU_HAS_CACHEV2
+#define mb() asm volatile ("bar.brwarw\n":::"memory")
+#define rmb() asm volatile ("bar.brar\n":::"memory")
+#define wmb() asm volatile ("bar.bwaw\n":::"memory")
+
+#ifdef CONFIG_SMP
+#define __smp_mb() asm volatile ("bar.brwarws\n":::"memory")
+#define __smp_rmb() asm volatile ("bar.brars\n":::"memory")
+#define __smp_wmb() asm volatile ("bar.bwaws\n":::"memory")
+#endif /* CONFIG_SMP */
+
+#define sync_is() asm volatile ("sync.is\n":::"memory")
+
+#else /* !CONFIG_CPU_HAS_CACHEV2 */
+#define mb() asm volatile ("sync\n":::"memory")
+#endif
+
+#include <asm-generic/barrier.h>
+
+#endif /* __ASSEMBLY__ */
+#endif /* __ASM_CSKY_BARRIER_H */
diff --git a/arch/csky/include/asm/cache.h b/arch/csky/include/asm/cache.h
new file mode 100644
index 0000000..d683734
--- /dev/null
+++ b/arch/csky/include/asm/cache.h
@@ -0,0 +1,30 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+
+#ifndef __ASM_CSKY_CACHE_H
+#define __ASM_CSKY_CACHE_H
+
+/* bytes per L1 cache line */
+#define L1_CACHE_SHIFT CONFIG_L1_CACHE_SHIFT
+
+#define L1_CACHE_BYTES (1 << L1_CACHE_SHIFT)
+
+#define ARCH_DMA_MINALIGN L1_CACHE_BYTES
+
+#ifndef __ASSEMBLY__
+
+void dcache_wb_line(unsigned long start);
+
+void icache_inv_range(unsigned long start, unsigned long end);
+void icache_inv_all(void);
+
+void dcache_wb_range(unsigned long start, unsigned long end);
+void dcache_wbinv_all(void);
+
+void cache_wbinv_range(unsigned long start, unsigned long end);
+void cache_wbinv_all(void);
+
+void dma_wbinv_range(unsigned long start, unsigned long end);
+void dma_wb_range(unsigned long start, unsigned long end);
+
+#endif
+#endif /* __ASM_CSKY_CACHE_H */
diff --git a/arch/csky/include/asm/cacheflush.h b/arch/csky/include/asm/cacheflush.h
new file mode 100644
index 0000000..a96da67
--- /dev/null
+++ b/arch/csky/include/asm/cacheflush.h
@@ -0,0 +1,9 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+// Copyright (C) 2018 Hangzhou C-SKY Microsystems co.,ltd.
+
+#ifndef __ASM_CSKY_CACHEFLUSH_H
+#define __ASM_CSKY_CACHEFLUSH_H
+
+#include <abi/cacheflush.h>
+
+#endif /* __ASM_CSKY_CACHEFLUSH_H */
diff --git a/arch/csky/include/asm/io.h b/arch/csky/include/asm/io.h
new file mode 100644
index 0000000..ecae6b3
--- /dev/null
+++ b/arch/csky/include/asm/io.h
@@ -0,0 +1,24 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+// Copyright (C) 2018 Hangzhou C-SKY Microsystems co.,ltd.
+
+#ifndef __ASM_CSKY_IO_H
+#define __ASM_CSKY_IO_H
+
+#include <abi/pgtable-bits.h>
+#include <linux/types.h>
+#include <linux/version.h>
+
+extern void __iomem *ioremap(phys_addr_t offset, size_t size);
+
+extern void iounmap(void *addr);
+
+extern int remap_area_pages(unsigned long address, phys_addr_t phys_addr,
+ size_t size, unsigned long flags);
+
+#define ioremap_nocache(phy, sz) ioremap(phy, sz)
+#define ioremap_wc ioremap_nocache
+#define ioremap_wt ioremap_nocache
+
+#include <asm-generic/io.h>
+
+#endif /* __ASM_CSKY_IO_H */
diff --git a/arch/csky/include/asm/tlb.h b/arch/csky/include/asm/tlb.h
new file mode 100644
index 0000000..8c7cc09
--- /dev/null
+++ b/arch/csky/include/asm/tlb.h
@@ -0,0 +1,25 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+// Copyright (C) 2018 Hangzhou C-SKY Microsystems co.,ltd.
+
+#ifndef __ASM_CSKY_TLB_H
+#define __ASM_CSKY_TLB_H
+
+#include <asm/cacheflush.h>
+
+#define tlb_start_vma(tlb, vma) \
+ do { \
+ if (!tlb->fullmm) \
+ flush_cache_range(vma, vma->vm_start, vma->vm_end); \
+ } while (0)
+
+#define tlb_end_vma(tlb, vma) \
+ do { \
+ if (!tlb->fullmm) \
+ flush_tlb_range(vma, vma->vm_start, vma->vm_end); \
+ } while (0)
+
+#define tlb_flush(tlb) flush_tlb_mm((tlb)->mm)
+
+#include <asm-generic/tlb.h>
+
+#endif /* __ASM_CSKY_TLB_H */
diff --git a/arch/csky/include/asm/tlbflush.h b/arch/csky/include/asm/tlbflush.h
new file mode 100644
index 0000000..6845b06
--- /dev/null
+++ b/arch/csky/include/asm/tlbflush.h
@@ -0,0 +1,25 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+// Copyright (C) 2018 Hangzhou C-SKY Microsystems co.,ltd.
+
+#ifndef __ASM_TLBFLUSH_H
+#define __ASM_TLBFLUSH_H
+
+/*
+ * TLB flushing:
+ *
+ * - flush_tlb_all() flushes all processes TLB entries
+ * - flush_tlb_mm(mm) flushes the specified mm context TLB entries
+ * - flush_tlb_page(vma, vmaddr) flushes one page
+ * - flush_tlb_range(vma, start, end) flushes a range of pages
+ * - flush_tlb_kernel_range(start, end) flushes a range of kernel pages
+ */
+extern void flush_tlb_all(void);
+extern void flush_tlb_mm(struct mm_struct *mm);
+extern void flush_tlb_page(struct vm_area_struct *vma, unsigned long page);
+extern void flush_tlb_range(struct vm_area_struct *vma, unsigned long start,
+ unsigned long end);
+extern void flush_tlb_kernel_range(unsigned long start, unsigned long end);
+
+extern void flush_tlb_one(unsigned long vaddr);
+
+#endif
diff --git a/arch/csky/include/uapi/asm/cachectl.h b/arch/csky/include/uapi/asm/cachectl.h
new file mode 100644
index 0000000..ddf2f39
--- /dev/null
+++ b/arch/csky/include/uapi/asm/cachectl.h
@@ -0,0 +1,13 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+
+#ifndef __ASM_CSKY_CACHECTL_H
+#define __ASM_CSKY_CACHECTL_H
+
+/*
+ * See "man cacheflush"
+ */
+#define ICACHE (1<<0)
+#define DCACHE (1<<1)
+#define BCACHE (ICACHE|DCACHE)
+
+#endif /* __ASM_CSKY_CACHECTL_H */
diff --git a/arch/csky/mm/cachev1.c b/arch/csky/mm/cachev1.c
new file mode 100644
index 0000000..b8a75cc
--- /dev/null
+++ b/arch/csky/mm/cachev1.c
@@ -0,0 +1,126 @@
+// SPDX-License-Identifier: GPL-2.0
+// Copyright (C) 2018 Hangzhou C-SKY Microsystems co.,ltd.
+
+#include <linux/spinlock.h>
+#include <asm/cache.h>
+#include <abi/reg_ops.h>
+
+/* for L1-cache */
+#define INS_CACHE (1 << 0)
+#define DATA_CACHE (1 << 1)
+#define CACHE_INV (1 << 4)
+#define CACHE_CLR (1 << 5)
+#define CACHE_OMS (1 << 6)
+#define CACHE_ITS (1 << 7)
+#define CACHE_LICF (1 << 31)
+
+/* for L2-cache */
+#define CR22_LEVEL_SHIFT (1)
+#define CR22_SET_SHIFT (7)
+#define CR22_WAY_SHIFT (30)
+#define CR22_WAY_SHIFT_L2 (29)
+
+static DEFINE_SPINLOCK(cache_lock);
+
+static inline void cache_op_line(unsigned long i, unsigned int val)
+{
+ mtcr("cr22", i);
+ mtcr("cr17", val);
+}
+
+#define CCR2_L2E (1 << 3)
+static void cache_op_all(unsigned int value, unsigned int l2)
+{
+ mtcr("cr17", value | CACHE_CLR);
+ mb();
+
+ if (l2 && (mfcr_ccr2() & CCR2_L2E)) {
+ mtcr("cr24", value | CACHE_CLR);
+ mb();
+ }
+}
+
+static void cache_op_range(
+ unsigned int start,
+ unsigned int end,
+ unsigned int value,
+ unsigned int l2)
+{
+ unsigned long i, flags;
+ unsigned int val = value | CACHE_CLR | CACHE_OMS;
+ bool l2_sync;
+
+ if (unlikely((end - start) >= PAGE_SIZE) ||
+ unlikely(start < PAGE_OFFSET) ||
+ unlikely(start >= PAGE_OFFSET + LOWMEM_LIMIT)) {
+ cache_op_all(value, l2);
+ return;
+ }
+
+ if ((mfcr_ccr2() & CCR2_L2E) && l2)
+ l2_sync = 1;
+ else
+ l2_sync = 0;
+
+ spin_lock_irqsave(&cache_lock, flags);
+
+ i = start & ~(L1_CACHE_BYTES - 1);
+ for (; i < end; i += L1_CACHE_BYTES) {
+ cache_op_line(i, val);
+ if (l2_sync) {
+ mb();
+ mtcr("cr24", val);
+ }
+ }
+ spin_unlock_irqrestore(&cache_lock, flags);
+
+ mb();
+}
+
+void dcache_wb_line(unsigned long start)
+{
+ asm volatile("idly4\n":::"memory");
+ cache_op_line(start, DATA_CACHE|CACHE_CLR);
+ mb();
+}
+
+void icache_inv_range(unsigned long start, unsigned long end)
+{
+ cache_op_range(start, end, INS_CACHE|CACHE_INV, 0);
+}
+
+void icache_inv_all(void)
+{
+ cache_op_all(INS_CACHE|CACHE_INV, 0);
+}
+
+void dcache_wb_range(unsigned long start, unsigned long end)
+{
+ cache_op_range(start, end, DATA_CACHE|CACHE_CLR, 0);
+}
+
+void dcache_wbinv_all(void)
+{
+ cache_op_all(DATA_CACHE|CACHE_CLR|CACHE_INV, 0);
+}
+
+void cache_wbinv_range(unsigned long start, unsigned long end)
+{
+ cache_op_range(start, end, INS_CACHE|DATA_CACHE|CACHE_CLR|CACHE_INV, 0);
+}
+EXPORT_SYMBOL(cache_wbinv_range);
+
+void cache_wbinv_all(void)
+{
+ cache_op_all(INS_CACHE|DATA_CACHE|CACHE_CLR|CACHE_INV, 0);
+}
+
+void dma_wbinv_range(unsigned long start, unsigned long end)
+{
+ cache_op_range(start, end, DATA_CACHE|CACHE_CLR|CACHE_INV, 1);
+}
+
+void dma_wb_range(unsigned long start, unsigned long end)
+{
+ cache_op_range(start, end, DATA_CACHE|CACHE_INV, 1);
+}
diff --git a/arch/csky/mm/cachev2.c b/arch/csky/mm/cachev2.c
new file mode 100644
index 0000000..baaf05d
--- /dev/null
+++ b/arch/csky/mm/cachev2.c
@@ -0,0 +1,79 @@
+// SPDX-License-Identifier: GPL-2.0
+// Copyright (C) 2018 Hangzhou C-SKY Microsystems co.,ltd.
+
+#include <linux/spinlock.h>
+#include <linux/smp.h>
+#include <asm/cache.h>
+#include <asm/barrier.h>
+
+inline void dcache_wb_line(unsigned long start)
+{
+ asm volatile("dcache.cval1 %0\n"::"r"(start):"memory");
+ sync_is();
+}
+
+void icache_inv_range(unsigned long start, unsigned long end)
+{
+ unsigned long i = start & ~(L1_CACHE_BYTES - 1);
+
+ for (; i < end; i += L1_CACHE_BYTES)
+ asm volatile("icache.iva %0\n"::"r"(i):"memory");
+ sync_is();
+}
+
+void icache_inv_all(void)
+{
+ asm volatile("icache.ialls\n":::"memory");
+ sync_is();
+}
+
+void dcache_wb_range(unsigned long start, unsigned long end)
+{
+ unsigned long i = start & ~(L1_CACHE_BYTES - 1);
+
+ for (; i < end; i += L1_CACHE_BYTES)
+ asm volatile("dcache.cval1 %0\n"::"r"(i):"memory");
+ sync_is();
+}
+
+void dcache_inv_range(unsigned long start, unsigned long end)
+{
+ unsigned long i = start & ~(L1_CACHE_BYTES - 1);
+
+ for (; i < end; i += L1_CACHE_BYTES)
+ asm volatile("dcache.civa %0\n"::"r"(i):"memory");
+ sync_is();
+}
+
+void cache_wbinv_range(unsigned long start, unsigned long end)
+{
+ unsigned long i = start & ~(L1_CACHE_BYTES - 1);
+
+ for (; i < end; i += L1_CACHE_BYTES)
+ asm volatile("dcache.cval1 %0\n"::"r"(i):"memory");
+ sync_is();
+
+ i = start & ~(L1_CACHE_BYTES - 1);
+ for (; i < end; i += L1_CACHE_BYTES)
+ asm volatile("icache.iva %0\n"::"r"(i):"memory");
+ sync_is();
+}
+EXPORT_SYMBOL(cache_wbinv_range);
+
+void dma_wbinv_range(unsigned long start, unsigned long end)
+{
+ unsigned long i = start & ~(L1_CACHE_BYTES - 1);
+
+ for (; i < end; i += L1_CACHE_BYTES)
+ asm volatile("dcache.civa %0\n"::"r"(i):"memory");
+ sync_is();
+}
+
+void dma_wb_range(unsigned long start, unsigned long end)
+{
+ unsigned long i = start & ~(L1_CACHE_BYTES - 1);
+
+ for (; i < end; i += L1_CACHE_BYTES)
+ asm volatile("dcache.civa %0\n"::"r"(i):"memory");
+ sync_is();
+}
diff --git a/arch/csky/mm/syscache.c b/arch/csky/mm/syscache.c
new file mode 100644
index 0000000..c4645e4
--- /dev/null
+++ b/arch/csky/mm/syscache.c
@@ -0,0 +1,32 @@
+// SPDX-License-Identifier: GPL-2.0
+// Copyright (C) 2018 Hangzhou C-SKY Microsystems co.,ltd.
+
+#include <linux/syscalls.h>
+#include <asm/page.h>
+#include <asm/cache.h>
+#include <asm/cachectl.h>
+
+SYSCALL_DEFINE3(cacheflush,
+ void __user *, addr,
+ unsigned long, bytes,
+ int, cache)
+{
+ switch (cache) {
+ case ICACHE:
+ icache_inv_range((unsigned long)addr,
+ (unsigned long)addr + bytes);
+ break;
+ case DCACHE:
+ dcache_wb_range((unsigned long)addr,
+ (unsigned long)addr + bytes);
+ break;
+ case BCACHE:
+ cache_wbinv_range((unsigned long)addr,
+ (unsigned long)addr + bytes);
+ break;
+ default:
+ return -EINVAL;
+ }
+
+ return 0;
+}
diff --git a/arch/csky/mm/tlb.c b/arch/csky/mm/tlb.c
new file mode 100644
index 0000000..08b8394
--- /dev/null
+++ b/arch/csky/mm/tlb.c
@@ -0,0 +1,219 @@
+// SPDX-License-Identifier: GPL-2.0
+// Copyright (C) 2018 Hangzhou C-SKY Microsystems co.,ltd.
+
+#include <linux/init.h>
+#include <linux/mm.h>
+#include <linux/module.h>
+#include <linux/sched.h>
+
+#include <asm/mmu_context.h>
+#include <asm/pgtable.h>
+#include <asm/setup.h>
+
+#define CSKY_TLB_SIZE CONFIG_CPU_TLB_SIZE
+
+void flush_tlb_all(void)
+{
+ tlb_invalid_all();
+}
+
+void flush_tlb_mm(struct mm_struct *mm)
+{
+ int cpu = smp_processor_id();
+
+ if (cpu_context(cpu, mm) != 0)
+ drop_mmu_context(mm, cpu);
+
+ tlb_invalid_all();
+}
+
+#define restore_asid_inv_utlb(oldpid, newpid) \
+do { \
+ if ((oldpid & ASID_MASK) == newpid) \
+ write_mmu_entryhi(oldpid + 1); \
+ write_mmu_entryhi(oldpid); \
+} while (0)
+
+void flush_tlb_range(struct vm_area_struct *vma, unsigned long start,
+ unsigned long end)
+{
+ struct mm_struct *mm = vma->vm_mm;
+ int cpu = smp_processor_id();
+
+ if (cpu_context(cpu, mm) != 0) {
+ unsigned long size, flags;
+ int newpid = cpu_asid(cpu, mm);
+
+ local_irq_save(flags);
+ size = (end - start + (PAGE_SIZE - 1)) >> PAGE_SHIFT;
+ size = (size + 1) >> 1;
+ if (size <= CSKY_TLB_SIZE/2) {
+ start &= (PAGE_MASK << 1);
+ end += ((PAGE_SIZE << 1) - 1);
+ end &= (PAGE_MASK << 1);
+#ifdef CONFIG_CPU_HAS_TLBI
+ while (start < end) {
+ asm volatile("tlbi.vaas %0"
+ ::"r"(start | newpid));
+ start += (PAGE_SIZE << 1);
+ }
+ sync_is();
+#else
+ {
+ int oldpid = read_mmu_entryhi();
+
+ while (start < end) {
+ int idx;
+
+ write_mmu_entryhi(start | newpid);
+ start += (PAGE_SIZE << 1);
+ tlb_probe();
+ idx = read_mmu_index();
+ if (idx >= 0)
+ tlb_invalid_indexed();
+ }
+ restore_asid_inv_utlb(oldpid, newpid);
+ }
+#endif
+ } else {
+ drop_mmu_context(mm, cpu);
+ }
+ local_irq_restore(flags);
+ }
+}
+
+void flush_tlb_kernel_range(unsigned long start, unsigned long end)
+{
+ unsigned long size, flags;
+
+ local_irq_save(flags);
+ size = (end - start + (PAGE_SIZE - 1)) >> PAGE_SHIFT;
+ if (size <= CSKY_TLB_SIZE) {
+ start &= (PAGE_MASK << 1);
+ end += ((PAGE_SIZE << 1) - 1);
+ end &= (PAGE_MASK << 1);
+#ifdef CONFIG_CPU_HAS_TLBI
+ while (start < end) {
+ asm volatile("tlbi.vaas %0"::"r"(start));
+ start += (PAGE_SIZE << 1);
+ }
+ sync_is();
+#else
+ {
+ int oldpid = read_mmu_entryhi();
+
+ while (start < end) {
+ int idx;
+
+ write_mmu_entryhi(start);
+ start += (PAGE_SIZE << 1);
+ tlb_probe();
+ idx = read_mmu_index();
+ if (idx >= 0)
+ tlb_invalid_indexed();
+ }
+ restore_asid_inv_utlb(oldpid, 0);
+ }
+#endif
+ } else {
+ flush_tlb_all();
+ }
+
+ local_irq_restore(flags);
+}
+
+void flush_tlb_page(struct vm_area_struct *vma, unsigned long page)
+{
+ int cpu = smp_processor_id();
+ int newpid = cpu_asid(cpu, vma->vm_mm);
+
+ if (!vma || cpu_context(cpu, vma->vm_mm) != 0) {
+ page &= (PAGE_MASK << 1);
+
+#ifdef CONFIG_CPU_HAS_TLBI
+ asm volatile("tlbi.vaas %0"::"r"(page | newpid));
+ sync_is();
+#else
+ {
+ int oldpid, idx;
+ unsigned long flags;
+
+ local_irq_save(flags);
+ oldpid = read_mmu_entryhi();
+ write_mmu_entryhi(page | newpid);
+ tlb_probe();
+ idx = read_mmu_index();
+ if (idx >= 0)
+ tlb_invalid_indexed();
+
+ restore_asid_inv_utlb(oldpid, newpid);
+ local_irq_restore(flags);
+ }
+#endif
+ }
+}
+
+/*
+ * Remove one kernel space TLB entry. This entry is assumed to be marked
+ * global so we don't do the ASID thing.
+ */
+void flush_tlb_one(unsigned long page)
+{
+ int oldpid;
+
+ oldpid = read_mmu_entryhi();
+ page &= (PAGE_MASK << 1);
+
+#ifdef CONFIG_CPU_HAS_TLBI
+ page = page | (oldpid & 0xfff);
+ asm volatile("tlbi.vaas %0"::"r"(page));
+ sync_is();
+#else
+ {
+ int idx;
+ unsigned long flags;
+
+ page = page | (oldpid & 0xff);
+
+ local_irq_save(flags);
+ write_mmu_entryhi(page);
+ tlb_probe();
+ idx = read_mmu_index();
+ if (idx >= 0)
+ tlb_invalid_indexed();
+ restore_asid_inv_utlb(oldpid, oldpid);
+ local_irq_restore(flags);
+ }
+#endif
+}
+EXPORT_SYMBOL(flush_tlb_one);
+
+/* show current 32 jtlbs */
+void show_jtlb_table(void)
+{
+ unsigned long flags;
+ int entryhi, entrylo0, entrylo1;
+ int entry;
+ int oldpid;
+
+ local_irq_save(flags);
+ entry = 0;
+ pr_info("\n\n\n");
+
+ oldpid = read_mmu_entryhi();
+ while (entry < CSKY_TLB_SIZE) {
+ write_mmu_index(entry);
+ tlb_read();
+ entryhi = read_mmu_entryhi();
+ entrylo0 = read_mmu_entrylo0();
+ entrylo0 = entrylo0;
+ entrylo1 = read_mmu_entrylo1();
+ entrylo1 = entrylo1;
+ pr_info("jtlb[%d]: entryhi - 0x%x; entrylo0 - 0x%x;"
+ " entrylo1 - 0x%x\n",
+ entry, entryhi, entrylo0, entrylo1);
+ entry++;
+ }
+ write_mmu_entryhi(oldpid);
+ local_irq_restore(flags);
+}
--
2.7.4
^ permalink raw reply related [flat|nested] 68+ messages in thread* Re: [PATCH V9 06/21] csky: Cache and TLB routines
2018-10-16 2:58 ` [PATCH V9 06/21] csky: Cache and TLB routines Guo Ren
@ 2018-10-17 15:08 ` Arnd Bergmann
0 siblings, 0 replies; 68+ messages in thread
From: Arnd Bergmann @ 2018-10-17 15:08 UTC (permalink / raw)
To: Guo Ren
Cc: Andrew Morton, Daniel Lezcano, David Miller, gregkh,
Christoph Hellwig, Marc Zyngier, Mark Rutland, Peter Zijlstra,
Rob Herring, Thomas Gleixner, Linux Kernel Mailing List,
linux-arch, DTML, Rob Herring, c-sky_gcc_upstream
On Tue, Oct 16, 2018 at 5:01 AM Guo Ren <ren_guo@c-sky.com> wrote:
>
> This patch adds cache and tlb sync codes for abiv1 & abiv2.
>
> Signed-off-by: Guo Ren <ren_guo@c-sky.com>
> Cc: Arnd Bergmann <arnd@arndb.de>
Reviewed-by: Arnd Bergmann <arnd@arndb.de>
^ permalink raw reply [flat|nested] 68+ messages in thread
* [PATCH V9 07/21] csky: MMU and page table management
2018-10-16 2:58 [PATCH V9 00/21] C-SKY(csky) Linux Kernel Port Guo Ren
` (5 preceding siblings ...)
2018-10-16 2:58 ` [PATCH V9 06/21] csky: Cache and TLB routines Guo Ren
@ 2018-10-16 2:58 ` Guo Ren
2018-10-17 15:06 ` Arnd Bergmann
2018-10-16 2:58 ` [PATCH V9 08/21] csky: Process management and Signal Guo Ren
` (14 subsequent siblings)
21 siblings, 1 reply; 68+ messages in thread
From: Guo Ren @ 2018-10-16 2:58 UTC (permalink / raw)
To: akpm, arnd, daniel.lezcano, davem, gregkh, hch, marc.zyngier,
mark.rutland, peterz, robh, tglx
Cc: linux-kernel, linux-arch, devicetree, robh+dt,
c-sky_gcc_upstream, Guo Ren
This patch adds files related to memory management and here is our
memory-layout:
Fixmap : 0xffc02000 – 0xfffff000 (4 MB - 12KB)
Pkmap : 0xff800000 – 0xffc00000 (4 MB)
Vmalloc : 0xf0200000 – 0xff000000 (238 MB)
Lowmem : 0x80000000 – 0xc0000000 (1GB)
abiv1 CPU (CK610) is VIPT cache and it doesn't support highmem.
abiv2 CPUs are all PIPT cache and they could support highmem.
Lowmem is directly mapped by msa0 & msa1 reg, and we needn't setup
memory page table for it.
Link:https://lore.kernel.org/lkml/20180518215548.GH17671@n2100.armlinux.org.uk/
Signed-off-by: Guo Ren <ren_guo@c-sky.com>
Cc: Christoph Hellwig <hch@infradead.org>
Cc: Arnd Bergmann <arnd@arndb.de>
---
Changelog:
- fixup pfn_valid(pfn) compile error for modules.
- fixup compile warning for CONFIG_DEBUG_HIGHMEM.
- fixup module compile error for EXPORT_SYMBOL(empty_zero_page).
- dma-mapping: fix up dma_mapping error
The arch_sync_dma_for_cpu()/arch_sync_dma_for_device() implementation is
broken for some combinations that end up in a BUG() instead of performing
the necessary flushes.
- Fixup arch_sync_dma() broken when size is larger than PAGE_SIZE.
Ignore the DMA_ATTR_NON_CONSISTENT, remove the BUG() and use return NULL
instead.
The implementation of arch should follow the following rules:
map for_cpu for_device unmap
TO_DEV writeback none writeback none
TO_CPU invalidate invalidate* invalidate invalidate*
BIDIR writeback invalidate writeback invalidate
---
---
arch/csky/abiv1/inc/abi/ckmmu.h | 75 ++++++++
arch/csky/abiv1/inc/abi/page.h | 27 +++
arch/csky/abiv1/inc/abi/pgtable-bits.h | 37 ++++
arch/csky/abiv1/mmap.c | 66 +++++++
arch/csky/abiv2/inc/abi/ckmmu.h | 87 ++++++++++
arch/csky/abiv2/inc/abi/page.h | 14 ++
arch/csky/abiv2/inc/abi/pgtable-bits.h | 37 ++++
arch/csky/include/asm/addrspace.h | 10 ++
arch/csky/include/asm/fixmap.h | 27 +++
arch/csky/include/asm/highmem.h | 51 ++++++
arch/csky/include/asm/mmu.h | 12 ++
arch/csky/include/asm/page.h | 104 +++++++++++
arch/csky/include/asm/pgalloc.h | 115 +++++++++++++
arch/csky/include/asm/pgtable.h | 306 +++++++++++++++++++++++++++++++++
arch/csky/include/asm/segment.h | 19 ++
arch/csky/include/asm/shmparam.h | 11 ++
arch/csky/mm/dma-mapping.c | 254 +++++++++++++++++++++++++++
arch/csky/mm/highmem.c | 198 +++++++++++++++++++++
arch/csky/mm/init.c | 122 +++++++++++++
arch/csky/mm/ioremap.c | 48 ++++++
20 files changed, 1620 insertions(+)
create mode 100644 arch/csky/abiv1/inc/abi/ckmmu.h
create mode 100644 arch/csky/abiv1/inc/abi/page.h
create mode 100644 arch/csky/abiv1/inc/abi/pgtable-bits.h
create mode 100644 arch/csky/abiv1/mmap.c
create mode 100644 arch/csky/abiv2/inc/abi/ckmmu.h
create mode 100644 arch/csky/abiv2/inc/abi/page.h
create mode 100644 arch/csky/abiv2/inc/abi/pgtable-bits.h
create mode 100644 arch/csky/include/asm/addrspace.h
create mode 100644 arch/csky/include/asm/fixmap.h
create mode 100644 arch/csky/include/asm/highmem.h
create mode 100644 arch/csky/include/asm/mmu.h
create mode 100644 arch/csky/include/asm/page.h
create mode 100644 arch/csky/include/asm/pgalloc.h
create mode 100644 arch/csky/include/asm/pgtable.h
create mode 100644 arch/csky/include/asm/segment.h
create mode 100644 arch/csky/include/asm/shmparam.h
create mode 100644 arch/csky/mm/dma-mapping.c
create mode 100644 arch/csky/mm/highmem.c
create mode 100644 arch/csky/mm/init.c
create mode 100644 arch/csky/mm/ioremap.c
diff --git a/arch/csky/abiv1/inc/abi/ckmmu.h b/arch/csky/abiv1/inc/abi/ckmmu.h
new file mode 100644
index 0000000..3a00201
--- /dev/null
+++ b/arch/csky/abiv1/inc/abi/ckmmu.h
@@ -0,0 +1,75 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+// Copyright (C) 2018 Hangzhou C-SKY Microsystems co.,ltd.
+
+#ifndef __ASM_CSKY_CKMMUV1_H
+#define __ASM_CSKY_CKMMUV1_H
+#include <abi/reg_ops.h>
+
+static inline int read_mmu_index(void)
+{
+ return cprcr("cpcr0");
+}
+
+static inline void write_mmu_index(int value)
+{
+ cpwcr("cpcr0", value);
+}
+
+static inline int read_mmu_entrylo0(void)
+{
+ return cprcr("cpcr2") << 6;
+}
+
+static inline int read_mmu_entrylo1(void)
+{
+ return cprcr("cpcr3") << 6;
+}
+
+static inline void write_mmu_pagemask(int value)
+{
+ cpwcr("cpcr6", value);
+}
+
+static inline int read_mmu_entryhi(void)
+{
+ return cprcr("cpcr4");
+}
+
+static inline void write_mmu_entryhi(int value)
+{
+ cpwcr("cpcr4", value);
+}
+
+/*
+ * TLB operations.
+ */
+static inline void tlb_probe(void)
+{
+ cpwcr("cpcr8", 0x80000000);
+}
+
+static inline void tlb_read(void)
+{
+ cpwcr("cpcr8", 0x40000000);
+}
+
+static inline void tlb_invalid_all(void)
+{
+ cpwcr("cpcr8", 0x04000000);
+}
+
+static inline void tlb_invalid_indexed(void)
+{
+ cpwcr("cpcr8", 0x02000000);
+}
+
+static inline void setup_pgd(unsigned long pgd, bool kernel)
+{
+ cpwcr("cpcr29", pgd);
+}
+
+static inline unsigned long get_pgd(void)
+{
+ return cprcr("cpcr29");
+}
+#endif /* __ASM_CSKY_CKMMUV1_H */
diff --git a/arch/csky/abiv1/inc/abi/page.h b/arch/csky/abiv1/inc/abi/page.h
new file mode 100644
index 0000000..6336e92
--- /dev/null
+++ b/arch/csky/abiv1/inc/abi/page.h
@@ -0,0 +1,27 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+// Copyright (C) 2018 Hangzhou C-SKY Microsystems co.,ltd.
+
+extern unsigned long shm_align_mask;
+extern void flush_dcache_page(struct page *page);
+
+static inline unsigned long pages_do_alias(unsigned long addr1,
+ unsigned long addr2)
+{
+ return (addr1 ^ addr2) & shm_align_mask;
+}
+
+static inline void clear_user_page(void *addr, unsigned long vaddr,
+ struct page *page)
+{
+ clear_page(addr);
+ if (pages_do_alias((unsigned long) addr, vaddr & PAGE_MASK))
+ flush_dcache_page(page);
+}
+
+static inline void copy_user_page(void *to, void *from, unsigned long vaddr,
+ struct page *page)
+{
+ copy_page(to, from);
+ if (pages_do_alias((unsigned long) to, vaddr & PAGE_MASK))
+ flush_dcache_page(page);
+}
diff --git a/arch/csky/abiv1/inc/abi/pgtable-bits.h b/arch/csky/abiv1/inc/abi/pgtable-bits.h
new file mode 100644
index 0000000..455075b
--- /dev/null
+++ b/arch/csky/abiv1/inc/abi/pgtable-bits.h
@@ -0,0 +1,37 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+// Copyright (C) 2018 Hangzhou C-SKY Microsystems co.,ltd.
+
+#ifndef __ASM_CSKY_PGTABLE_BITS_H
+#define __ASM_CSKY_PGTABLE_BITS_H
+
+/* implemented in software */
+#define _PAGE_ACCESSED (1<<3)
+#define PAGE_ACCESSED_BIT (3)
+
+#define _PAGE_READ (1<<1)
+#define _PAGE_WRITE (1<<2)
+#define _PAGE_PRESENT (1<<0)
+
+#define _PAGE_MODIFIED (1<<4)
+#define PAGE_MODIFIED_BIT (4)
+
+/* implemented in hardware */
+#define _PAGE_GLOBAL (1<<6)
+
+#define _PAGE_VALID (1<<7)
+#define PAGE_VALID_BIT (7)
+
+#define _PAGE_DIRTY (1<<8)
+#define PAGE_DIRTY_BIT (8)
+
+#define _PAGE_CACHE (3<<9)
+#define _PAGE_UNCACHE (2<<9)
+
+#define _CACHE_MASK (7<<9)
+
+#define _CACHE_CACHED (_PAGE_VALID | _PAGE_CACHE)
+#define _CACHE_UNCACHED (_PAGE_VALID | _PAGE_UNCACHE)
+
+#define HAVE_ARCH_UNMAPPED_AREA
+
+#endif /* __ASM_CSKY_PGTABLE_BITS_H */
diff --git a/arch/csky/abiv1/mmap.c b/arch/csky/abiv1/mmap.c
new file mode 100644
index 0000000..b462fd5
--- /dev/null
+++ b/arch/csky/abiv1/mmap.c
@@ -0,0 +1,66 @@
+// SPDX-License-Identifier: GPL-2.0
+// Copyright (C) 2018 Hangzhou C-SKY Microsystems co.,ltd.
+
+#include <linux/fs.h>
+#include <linux/mm.h>
+#include <linux/mman.h>
+#include <linux/shm.h>
+#include <linux/sched.h>
+#include <linux/random.h>
+#include <linux/io.h>
+
+unsigned long shm_align_mask = (0x4000 >> 1) - 1; /* Sane caches */
+
+#define COLOUR_ALIGN(addr, pgoff) \
+ ((((addr) + shm_align_mask) & ~shm_align_mask) + \
+ (((pgoff) << PAGE_SHIFT) & shm_align_mask))
+
+unsigned long arch_get_unmapped_area(struct file *filp, unsigned long addr,
+ unsigned long len, unsigned long pgoff, unsigned long flags)
+{
+ struct vm_area_struct *vmm;
+ int do_color_align;
+
+ if (flags & MAP_FIXED) {
+ /*
+ * We do not accept a shared mapping if it would violate
+ * cache aliasing constraints.
+ */
+ if ((flags & MAP_SHARED) &&
+ ((addr - (pgoff << PAGE_SHIFT)) & shm_align_mask))
+ return -EINVAL;
+ return addr;
+ }
+
+ if (len > TASK_SIZE)
+ return -ENOMEM;
+ do_color_align = 0;
+ if (filp || (flags & MAP_SHARED))
+ do_color_align = 1;
+ if (addr) {
+ if (do_color_align)
+ addr = COLOUR_ALIGN(addr, pgoff);
+ else
+ addr = PAGE_ALIGN(addr);
+ vmm = find_vma(current->mm, addr);
+ if (TASK_SIZE - len >= addr &&
+ (!vmm || addr + len <= vmm->vm_start))
+ return addr;
+ }
+ addr = TASK_UNMAPPED_BASE;
+ if (do_color_align)
+ addr = COLOUR_ALIGN(addr, pgoff);
+ else
+ addr = PAGE_ALIGN(addr);
+
+ for (vmm = find_vma(current->mm, addr); ; vmm = vmm->vm_next) {
+ /* At this point: (!vmm || addr < vmm->vm_end). */
+ if (TASK_SIZE - len < addr)
+ return -ENOMEM;
+ if (!vmm || addr + len <= vmm->vm_start)
+ return addr;
+ addr = vmm->vm_end;
+ if (do_color_align)
+ addr = COLOUR_ALIGN(addr, pgoff);
+ }
+}
diff --git a/arch/csky/abiv2/inc/abi/ckmmu.h b/arch/csky/abiv2/inc/abi/ckmmu.h
new file mode 100644
index 0000000..97230ad
--- /dev/null
+++ b/arch/csky/abiv2/inc/abi/ckmmu.h
@@ -0,0 +1,87 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+// Copyright (C) 2018 Hangzhou C-SKY Microsystems co.,ltd.
+
+#ifndef __ASM_CSKY_CKMMUV2_H
+#define __ASM_CSKY_CKMMUV2_H
+
+#include <abi/reg_ops.h>
+#include <asm/barrier.h>
+
+static inline int read_mmu_index(void)
+{
+ return mfcr("cr<0, 15>");
+}
+
+static inline void write_mmu_index(int value)
+{
+ mtcr("cr<0, 15>", value);
+}
+
+static inline int read_mmu_entrylo0(void)
+{
+ return mfcr("cr<2, 15>");
+}
+
+static inline int read_mmu_entrylo1(void)
+{
+ return mfcr("cr<3, 15>");
+}
+
+static inline void write_mmu_pagemask(int value)
+{
+ mtcr("cr<6, 15>", value);
+}
+
+static inline int read_mmu_entryhi(void)
+{
+ return mfcr("cr<4, 15>");
+}
+
+static inline void write_mmu_entryhi(int value)
+{
+ mtcr("cr<4, 15>", value);
+}
+
+/*
+ * TLB operations.
+ */
+static inline void tlb_probe(void)
+{
+ mtcr("cr<8, 15>", 0x80000000);
+}
+
+static inline void tlb_read(void)
+{
+ mtcr("cr<8, 15>", 0x40000000);
+}
+
+static inline void tlb_invalid_all(void)
+{
+#ifdef CONFIG_CPU_HAS_TLBI
+ asm volatile("tlbi.alls\n":::"memory");
+ sync_is();
+#else
+ mtcr("cr<8, 15>", 0x04000000);
+#endif
+}
+
+static inline void tlb_invalid_indexed(void)
+{
+ mtcr("cr<8, 15>", 0x02000000);
+}
+
+/* setup hardrefil pgd */
+static inline unsigned long get_pgd(void)
+{
+ return mfcr("cr<29, 15>");
+}
+
+static inline void setup_pgd(unsigned long pgd, bool kernel)
+{
+ if (kernel)
+ mtcr("cr<28, 15>", pgd);
+ else
+ mtcr("cr<29, 15>", pgd);
+}
+
+#endif /* __ASM_CSKY_CKMMUV2_H */
diff --git a/arch/csky/abiv2/inc/abi/page.h b/arch/csky/abiv2/inc/abi/page.h
new file mode 100644
index 0000000..0a70cb5
--- /dev/null
+++ b/arch/csky/abiv2/inc/abi/page.h
@@ -0,0 +1,14 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+// Copyright (C) 2018 Hangzhou C-SKY Microsystems co.,ltd.
+
+static inline void clear_user_page(void *addr, unsigned long vaddr,
+ struct page *page)
+{
+ clear_page(addr);
+}
+
+static inline void copy_user_page(void *to, void *from, unsigned long vaddr,
+ struct page *page)
+{
+ copy_page(to, from);
+}
diff --git a/arch/csky/abiv2/inc/abi/pgtable-bits.h b/arch/csky/abiv2/inc/abi/pgtable-bits.h
new file mode 100644
index 0000000..b20ae19
--- /dev/null
+++ b/arch/csky/abiv2/inc/abi/pgtable-bits.h
@@ -0,0 +1,37 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+// Copyright (C) 2018 Hangzhou C-SKY Microsystems co.,ltd.
+
+#ifndef __ASM_CSKY_PGTABLE_BITS_H
+#define __ASM_CSKY_PGTABLE_BITS_H
+
+/* implemented in software */
+#define _PAGE_ACCESSED (1<<7)
+#define PAGE_ACCESSED_BIT (7)
+
+#define _PAGE_READ (1<<8)
+#define _PAGE_WRITE (1<<9)
+#define _PAGE_PRESENT (1<<10)
+
+#define _PAGE_MODIFIED (1<<11)
+#define PAGE_MODIFIED_BIT (11)
+
+/* implemented in hardware */
+#define _PAGE_GLOBAL (1<<0)
+
+#define _PAGE_VALID (1<<1)
+#define PAGE_VALID_BIT (1)
+
+#define _PAGE_DIRTY (1<<2)
+#define PAGE_DIRTY_BIT (2)
+
+#define _PAGE_SO (1<<5)
+#define _PAGE_BUF (1<<6)
+
+#define _PAGE_CACHE (1<<3)
+
+#define _CACHE_MASK _PAGE_CACHE
+
+#define _CACHE_CACHED (_PAGE_VALID | _PAGE_CACHE | _PAGE_BUF)
+#define _CACHE_UNCACHED (_PAGE_VALID | _PAGE_SO)
+
+#endif /* __ASM_CSKY_PGTABLE_BITS_H */
diff --git a/arch/csky/include/asm/addrspace.h b/arch/csky/include/asm/addrspace.h
new file mode 100644
index 0000000..d1c2ede
--- /dev/null
+++ b/arch/csky/include/asm/addrspace.h
@@ -0,0 +1,10 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+// Copyright (C) 2018 Hangzhou C-SKY Microsystems co.,ltd.
+
+#ifndef __ASM_CSKY_ADDRSPACE_H
+#define __ASM_CSKY_ADDRSPACE_H
+
+#define KSEG0 0x80000000ul
+#define KSEG0ADDR(a) (((unsigned long)a & 0x1fffffff) | KSEG0)
+
+#endif /* __ASM_CSKY_ADDRSPACE_H */
diff --git a/arch/csky/include/asm/fixmap.h b/arch/csky/include/asm/fixmap.h
new file mode 100644
index 0000000..380ff0a
--- /dev/null
+++ b/arch/csky/include/asm/fixmap.h
@@ -0,0 +1,27 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+// Copyright (C) 2018 Hangzhou C-SKY Microsystems co.,ltd.
+
+#ifndef __ASM_CSKY_FIXMAP_H
+#define __ASM_CSKY_FIXMAP_H
+
+#include <asm/page.h>
+#ifdef CONFIG_HIGHMEM
+#include <linux/threads.h>
+#include <asm/kmap_types.h>
+#endif
+
+enum fixed_addresses {
+#ifdef CONFIG_HIGHMEM
+ FIX_KMAP_BEGIN,
+ FIX_KMAP_END = FIX_KMAP_BEGIN + (KM_TYPE_NR * NR_CPUS) - 1,
+#endif
+ __end_of_fixed_addresses
+};
+
+#define FIXADDR_TOP 0xffffc000
+#define FIXADDR_SIZE (__end_of_fixed_addresses << PAGE_SHIFT)
+#define FIXADDR_START (FIXADDR_TOP - FIXADDR_SIZE)
+
+#include <asm-generic/fixmap.h>
+
+#endif /* __ASM_CSKY_FIXMAP_H */
diff --git a/arch/csky/include/asm/highmem.h b/arch/csky/include/asm/highmem.h
new file mode 100644
index 0000000..a345a2f
--- /dev/null
+++ b/arch/csky/include/asm/highmem.h
@@ -0,0 +1,51 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+// Copyright (C) 2018 Hangzhou C-SKY Microsystems co.,ltd.
+
+#ifndef __ASM_CSKY_HIGHMEM_H
+#define __ASM_CSKY_HIGHMEM_H
+
+#ifdef __KERNEL__
+
+#include <linux/init.h>
+#include <linux/interrupt.h>
+#include <linux/uaccess.h>
+#include <asm/kmap_types.h>
+#include <asm/cache.h>
+
+/* undef for production */
+#define HIGHMEM_DEBUG 1
+
+/* declarations for highmem.c */
+extern unsigned long highstart_pfn, highend_pfn;
+
+extern pte_t *pkmap_page_table;
+
+/*
+ * Right now we initialize only a single pte table. It can be extended
+ * easily, subsequent pte tables have to be allocated in one physical
+ * chunk of RAM.
+ */
+#define LAST_PKMAP 1024
+#define LAST_PKMAP_MASK (LAST_PKMAP-1)
+#define PKMAP_NR(virt) ((virt-PKMAP_BASE) >> PAGE_SHIFT)
+#define PKMAP_ADDR(nr) (PKMAP_BASE + ((nr) << PAGE_SHIFT))
+
+extern void *kmap_high(struct page *page);
+extern void kunmap_high(struct page *page);
+
+extern void *kmap(struct page *page);
+extern void kunmap(struct page *page);
+extern void *kmap_atomic(struct page *page);
+extern void __kunmap_atomic(void *kvaddr);
+extern void *kmap_atomic_pfn(unsigned long pfn);
+extern struct page *kmap_atomic_to_page(void *ptr);
+
+#define flush_cache_kmaps() do {} while (0)
+
+extern void kmap_init(void);
+
+#define kmap_prot PAGE_KERNEL
+
+#endif /* __KERNEL__ */
+
+#endif /* __ASM_CSKY_HIGHMEM_H */
diff --git a/arch/csky/include/asm/mmu.h b/arch/csky/include/asm/mmu.h
new file mode 100644
index 0000000..cb34467
--- /dev/null
+++ b/arch/csky/include/asm/mmu.h
@@ -0,0 +1,12 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+// Copyright (C) 2018 Hangzhou C-SKY Microsystems co.,ltd.
+
+#ifndef __ASM_CSKY_MMU_H
+#define __ASM_CSKY_MMU_H
+
+typedef struct {
+ unsigned long asid[NR_CPUS];
+ void *vdso;
+} mm_context_t;
+
+#endif /* __ASM_CSKY_MMU_H */
diff --git a/arch/csky/include/asm/page.h b/arch/csky/include/asm/page.h
new file mode 100644
index 0000000..73cf2bd
--- /dev/null
+++ b/arch/csky/include/asm/page.h
@@ -0,0 +1,104 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+
+#ifndef __ASM_CSKY_PAGE_H
+#define __ASM_CSKY_PAGE_H
+
+#include <asm/setup.h>
+#include <asm/cache.h>
+#include <linux/const.h>
+
+/*
+ * PAGE_SHIFT determines the page size
+ */
+#define PAGE_SHIFT 12
+#define PAGE_SIZE (_AC(1, UL) << PAGE_SHIFT)
+#define PAGE_MASK (~(PAGE_SIZE - 1))
+#define THREAD_SIZE (PAGE_SIZE * 2)
+#define THREAD_MASK (~(THREAD_SIZE - 1))
+#define THREAD_SHIFT (PAGE_SHIFT + 1)
+
+/*
+ * NOTE: virtual isn't really correct, actually it should be the offset into the
+ * memory node, but we have no highmem, so that works for now.
+ * TODO: implement (fast) pfn<->pgdat_idx conversion functions, this makes lots
+ * of the shifts unnecessary.
+ */
+
+#ifndef __ASSEMBLY__
+
+#include <linux/pfn.h>
+
+#define virt_to_pfn(kaddr) (__pa(kaddr) >> PAGE_SHIFT)
+#define pfn_to_virt(pfn) __va((pfn) << PAGE_SHIFT)
+
+#define virt_addr_valid(kaddr) ((void *)(kaddr) >= (void *)PAGE_OFFSET && \
+ (void *)(kaddr) < high_memory)
+#define pfn_valid(pfn) ((pfn) >= ARCH_PFN_OFFSET && ((pfn) - ARCH_PFN_OFFSET) < max_mapnr)
+
+extern void *memset(void *dest, int c, size_t l);
+extern void *memcpy(void *to, const void *from, size_t l);
+
+#define clear_page(page) memset((page), 0, PAGE_SIZE)
+#define copy_page(to, from) memcpy((to), (from), PAGE_SIZE)
+
+#define page_to_phys(page) (page_to_pfn(page) << PAGE_SHIFT)
+#define phys_to_page(paddr) (pfn_to_page(PFN_DOWN(paddr)))
+
+struct page;
+
+#include <abi/page.h>
+
+struct vm_area_struct;
+
+/*
+ * These are used to make use of C type-checking..
+ */
+typedef struct { unsigned long pte_low; } pte_t;
+#define pte_val(x) ((x).pte_low)
+
+typedef struct { unsigned long pgd; } pgd_t;
+typedef struct { unsigned long pgprot; } pgprot_t;
+typedef struct page *pgtable_t;
+
+#define pgd_val(x) ((x).pgd)
+#define pgprot_val(x) ((x).pgprot)
+
+#define ptep_buddy(x) ((pte_t *)((unsigned long)(x) ^ sizeof(pte_t)))
+
+#define __pte(x) ((pte_t) { (x) })
+#define __pgd(x) ((pgd_t) { (x) })
+#define __pgprot(x) ((pgprot_t) { (x) })
+
+#endif /* !__ASSEMBLY__ */
+
+#define PHYS_OFFSET (CONFIG_RAM_BASE & ~(LOWMEM_LIMIT - 1))
+#define PHYS_OFFSET_OFFSET (CONFIG_RAM_BASE & (LOWMEM_LIMIT - 1))
+#define ARCH_PFN_OFFSET PFN_DOWN(CONFIG_RAM_BASE)
+
+#define PAGE_OFFSET 0x80000000
+#define LOWMEM_LIMIT 0x40000000
+
+#define __pa(x) ((unsigned long)(x) - PAGE_OFFSET + PHYS_OFFSET)
+#define __va(x) ((void *)((unsigned long)(x) + PAGE_OFFSET - \
+ PHYS_OFFSET))
+#define __pa_symbol(x) __pa(RELOC_HIDE((unsigned long)(x), 0))
+
+#define MAP_NR(x) PFN_DOWN((unsigned long)(x) - PAGE_OFFSET - \
+ PHYS_OFFSET_OFFSET)
+#define virt_to_page(x) (mem_map + MAP_NR(x))
+
+#define VM_DATA_DEFAULT_FLAGS (VM_READ | VM_WRITE | VM_EXEC | \
+ VM_MAYREAD | VM_MAYWRITE | VM_MAYEXEC)
+
+/*
+ * main RAM and kernel working space are coincident at 0x80000000, but to make
+ * life more interesting, there's also an uncached virtual shadow at 0xb0000000
+ * - these mappings are fixed in the MMU
+ */
+
+#define pfn_to_kaddr(x) __va(PFN_PHYS(x))
+
+#include <asm-generic/memory_model.h>
+#include <asm-generic/getorder.h>
+
+#endif /* __ASM_CSKY_PAGE_H */
diff --git a/arch/csky/include/asm/pgalloc.h b/arch/csky/include/asm/pgalloc.h
new file mode 100644
index 0000000..bf4f4a0
--- /dev/null
+++ b/arch/csky/include/asm/pgalloc.h
@@ -0,0 +1,115 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+// Copyright (C) 2018 Hangzhou C-SKY Microsystems co.,ltd.
+
+#ifndef __ASM_CSKY_PGALLOC_H
+#define __ASM_CSKY_PGALLOC_H
+
+#include <linux/highmem.h>
+#include <linux/mm.h>
+#include <linux/sched.h>
+
+static inline void pmd_populate_kernel(struct mm_struct *mm, pmd_t *pmd,
+ pte_t *pte)
+{
+ set_pmd(pmd, __pmd(__pa(pte)));
+}
+
+static inline void pmd_populate(struct mm_struct *mm, pmd_t *pmd,
+ pgtable_t pte)
+{
+ set_pmd(pmd, __pmd(__pa(page_address(pte))));
+}
+
+#define pmd_pgtable(pmd) pmd_page(pmd)
+
+extern void pgd_init(unsigned long *p);
+
+static inline pte_t *pte_alloc_one_kernel(struct mm_struct *mm,
+ unsigned long address)
+{
+ pte_t *pte;
+ unsigned long *kaddr, i;
+
+ pte = (pte_t *) __get_free_pages(GFP_KERNEL | __GFP_RETRY_MAYFAIL,
+ PTE_ORDER);
+ kaddr = (unsigned long *)pte;
+ if (address & 0x80000000)
+ for (i = 0; i < (PAGE_SIZE/4); i++)
+ *(kaddr + i) = 0x1;
+ else
+ clear_page(kaddr);
+
+ return pte;
+}
+
+static inline struct page *pte_alloc_one(struct mm_struct *mm,
+ unsigned long address)
+{
+ struct page *pte;
+ unsigned long *kaddr, i;
+
+ pte = alloc_pages(GFP_KERNEL | __GFP_RETRY_MAYFAIL, PTE_ORDER);
+ if (pte) {
+ kaddr = kmap_atomic(pte);
+ if (address & 0x80000000) {
+ for (i = 0; i < (PAGE_SIZE/4); i++)
+ *(kaddr + i) = 0x1;
+ } else
+ clear_page(kaddr);
+ kunmap_atomic(kaddr);
+ pgtable_page_ctor(pte);
+ }
+ return pte;
+}
+
+static inline void pte_free_kernel(struct mm_struct *mm, pte_t *pte)
+{
+ free_pages((unsigned long)pte, PTE_ORDER);
+}
+
+static inline void pte_free(struct mm_struct *mm, pgtable_t pte)
+{
+ pgtable_page_dtor(pte);
+ __free_pages(pte, PTE_ORDER);
+}
+
+static inline void pgd_free(struct mm_struct *mm, pgd_t *pgd)
+{
+ free_pages((unsigned long)pgd, PGD_ORDER);
+}
+
+static inline pgd_t *pgd_alloc(struct mm_struct *mm)
+{
+ pgd_t *ret;
+ pgd_t *init;
+
+ ret = (pgd_t *) __get_free_pages(GFP_KERNEL, PGD_ORDER);
+ if (ret) {
+ init = pgd_offset(&init_mm, 0UL);
+ pgd_init((unsigned long *)ret);
+ memcpy(ret + USER_PTRS_PER_PGD, init + USER_PTRS_PER_PGD,
+ (PTRS_PER_PGD - USER_PTRS_PER_PGD) * sizeof(pgd_t));
+ /* prevent out of order excute */
+ smp_mb();
+#ifdef CONFIG_CPU_NEED_TLBSYNC
+ dcache_wb_range((unsigned int)ret,
+ (unsigned int)(ret + PTRS_PER_PGD));
+#endif
+ }
+
+ return ret;
+}
+
+#define __pte_free_tlb(tlb, pte, address) \
+do { \
+ pgtable_page_dtor(pte); \
+ tlb_remove_page(tlb, pte); \
+} while (0)
+
+#define check_pgt_cache() do {} while (0)
+
+extern void pagetable_init(void);
+extern void pre_mmu_init(void);
+extern void pre_trap_init(void);
+
+#endif /* __ASM_CSKY_PGALLOC_H */
diff --git a/arch/csky/include/asm/pgtable.h b/arch/csky/include/asm/pgtable.h
new file mode 100644
index 0000000..edfcbb2
--- /dev/null
+++ b/arch/csky/include/asm/pgtable.h
@@ -0,0 +1,306 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+// Copyright (C) 2018 Hangzhou C-SKY Microsystems co.,ltd.
+
+#ifndef __ASM_CSKY_PGTABLE_H
+#define __ASM_CSKY_PGTABLE_H
+
+#include <asm/fixmap.h>
+#include <asm/addrspace.h>
+#include <abi/pgtable-bits.h>
+#include <asm-generic/pgtable-nopmd.h>
+
+#define PGDIR_SHIFT 22
+#define PGDIR_SIZE (1UL << PGDIR_SHIFT)
+#define PGDIR_MASK (~(PGDIR_SIZE-1))
+
+#define USER_PTRS_PER_PGD (0x80000000UL/PGDIR_SIZE)
+#define FIRST_USER_ADDRESS 0UL
+
+#define PKMAP_BASE (0xff800000)
+
+#define VMALLOC_START (0xc0008000)
+#define VMALLOC_END (PKMAP_BASE - 2*PAGE_SIZE)
+
+/*
+ * C-SKY is two-level paging structure:
+ */
+#define PGD_ORDER 0
+#define PTE_ORDER 0
+
+#define PTRS_PER_PGD ((PAGE_SIZE << PGD_ORDER) / sizeof(pgd_t))
+#define PTRS_PER_PMD 1
+#define PTRS_PER_PTE ((PAGE_SIZE << PTE_ORDER) / sizeof(pte_t))
+
+#define pte_ERROR(e) \
+ pr_err("%s:%d: bad pte %08lx.\n", __FILE__, __LINE__, (e).pte_low)
+#define pgd_ERROR(e) \
+ pr_err("%s:%d: bad pgd %08lx.\n", __FILE__, __LINE__, pgd_val(e))
+
+/* Find an entry in the third-level page table.. */
+#define __pte_offset_t(address) \
+ (((address) >> PAGE_SHIFT) & (PTRS_PER_PTE - 1))
+#define pte_offset_kernel(dir, address) \
+ (pmd_page_vaddr(*(dir)) + __pte_offset_t(address))
+#define pte_offset_map(dir, address) \
+ ((pte_t *)page_address(pmd_page(*(dir))) + __pte_offset_t(address))
+#define pmd_page(pmd) (pfn_to_page(pmd_phys(pmd) >> PAGE_SHIFT))
+#define pte_clear(mm, addr, ptep) set_pte((ptep), \
+ (((unsigned int)addr&0x80000000)?__pte(1):__pte(0)))
+#define pte_none(pte) (!(pte_val(pte)&0xfffffffe))
+#define pte_present(pte) (pte_val(pte) & _PAGE_PRESENT)
+#define pte_pfn(x) ((unsigned long)((x).pte_low >> PAGE_SHIFT))
+#define pfn_pte(pfn, prot) __pte(((unsigned long long)(pfn) << PAGE_SHIFT) \
+ | pgprot_val(prot))
+
+#define __READABLE (_PAGE_READ | _PAGE_VALID | _PAGE_ACCESSED)
+#define __WRITEABLE (_PAGE_WRITE | _PAGE_DIRTY | _PAGE_MODIFIED)
+
+#define _PAGE_CHG_MASK (PAGE_MASK | _PAGE_ACCESSED | _PAGE_MODIFIED | \
+ _CACHE_MASK)
+
+#define pte_unmap(pte) ((void)(pte))
+
+#define __swp_type(x) (((x).val >> 4) & 0xff)
+#define __swp_offset(x) ((x).val >> 12)
+#define __swp_entry(type, offset) ((swp_entry_t) {((type) << 4) | \
+ ((offset) << 12) })
+#define __pte_to_swp_entry(pte) ((swp_entry_t) { pte_val(pte) })
+#define __swp_entry_to_pte(x) ((pte_t) { (x).val })
+
+#define pte_page(x) pfn_to_page(pte_pfn(x))
+#define __mk_pte(page_nr, pgprot) __pte(((page_nr) << PAGE_SHIFT) | \
+ pgprot_val(pgprot))
+
+/*
+ * CSKY can't do page protection for execute, and considers that the same like
+ * read. Also, write permissions imply read permissions. This is the closest
+ * we can get by reasonable means..
+ */
+#define PAGE_NONE __pgprot(_PAGE_PRESENT | _CACHE_CACHED)
+#define PAGE_SHARED __pgprot(_PAGE_PRESENT | _PAGE_READ | _PAGE_WRITE | \
+ _CACHE_CACHED)
+#define PAGE_COPY __pgprot(_PAGE_PRESENT | _PAGE_READ | _CACHE_CACHED)
+#define PAGE_READONLY __pgprot(_PAGE_PRESENT | _PAGE_READ | _CACHE_CACHED)
+#define PAGE_KERNEL __pgprot(_PAGE_PRESENT | __READABLE | __WRITEABLE | \
+ _PAGE_GLOBAL | _CACHE_CACHED)
+#define PAGE_USERIO __pgprot(_PAGE_PRESENT | _PAGE_READ | _PAGE_WRITE | \
+ _CACHE_CACHED)
+
+#define __P000 PAGE_NONE
+#define __P001 PAGE_READONLY
+#define __P010 PAGE_COPY
+#define __P011 PAGE_COPY
+#define __P100 PAGE_READONLY
+#define __P101 PAGE_READONLY
+#define __P110 PAGE_COPY
+#define __P111 PAGE_COPY
+
+#define __S000 PAGE_NONE
+#define __S001 PAGE_READONLY
+#define __S010 PAGE_SHARED
+#define __S011 PAGE_SHARED
+#define __S100 PAGE_READONLY
+#define __S101 PAGE_READONLY
+#define __S110 PAGE_SHARED
+#define __S111 PAGE_SHARED
+
+extern unsigned long empty_zero_page[PAGE_SIZE / sizeof(unsigned long)];
+#define ZERO_PAGE(vaddr) (virt_to_page(empty_zero_page))
+
+extern void load_pgd(unsigned long pg_dir);
+extern pte_t invalid_pte_table[PTRS_PER_PTE];
+
+static inline int pte_special(pte_t pte) { return 0; }
+static inline pte_t pte_mkspecial(pte_t pte) { return pte; }
+
+static inline void set_pte(pte_t *p, pte_t pte)
+{
+ *p = pte;
+#if defined(CONFIG_CPU_NEED_TLBSYNC)
+ dcache_wb_line((u32)p);
+#endif
+ /* prevent out of order excution */
+ smp_mb();
+}
+#define set_pte_at(mm, addr, ptep, pteval) set_pte(ptep, pteval)
+
+static inline pte_t *pmd_page_vaddr(pmd_t pmd)
+{
+ unsigned long ptr;
+
+ ptr = pmd_val(pmd);
+
+ return __va(ptr);
+}
+
+#define pmd_phys(pmd) pmd_val(pmd)
+
+static inline void set_pmd(pmd_t *p, pmd_t pmd)
+{
+ *p = pmd;
+#if defined(CONFIG_CPU_NEED_TLBSYNC)
+ dcache_wb_line((u32)p);
+#endif
+ /* prevent specul excute */
+ smp_mb();
+}
+
+
+static inline int pmd_none(pmd_t pmd)
+{
+ return pmd_val(pmd) == __pa(invalid_pte_table);
+}
+
+#define pmd_bad(pmd) (pmd_val(pmd) & ~PAGE_MASK)
+
+static inline int pmd_present(pmd_t pmd)
+{
+ return (pmd_val(pmd) != __pa(invalid_pte_table));
+}
+
+static inline void pmd_clear(pmd_t *p)
+{
+ pmd_val(*p) = (__pa(invalid_pte_table));
+#if defined(CONFIG_CPU_NEED_TLBSYNC)
+ dcache_wb_line((u32)p);
+#endif
+}
+
+/*
+ * The following only work if pte_present() is true.
+ * Undefined behaviour if not..
+ */
+static inline int pte_read(pte_t pte)
+{
+ return pte.pte_low & _PAGE_READ;
+}
+
+static inline int pte_write(pte_t pte)
+{
+ return (pte).pte_low & _PAGE_WRITE;
+}
+
+static inline int pte_dirty(pte_t pte)
+{
+ return (pte).pte_low & _PAGE_MODIFIED;
+}
+
+static inline int pte_young(pte_t pte)
+{
+ return (pte).pte_low & _PAGE_ACCESSED;
+}
+
+static inline pte_t pte_wrprotect(pte_t pte)
+{
+ pte_val(pte) &= ~(_PAGE_WRITE | _PAGE_DIRTY);
+ return pte;
+}
+
+static inline pte_t pte_mkclean(pte_t pte)
+{
+ pte_val(pte) &= ~(_PAGE_MODIFIED|_PAGE_DIRTY);
+ return pte;
+}
+
+static inline pte_t pte_mkold(pte_t pte)
+{
+ pte_val(pte) &= ~(_PAGE_ACCESSED|_PAGE_VALID);
+ return pte;
+}
+
+static inline pte_t pte_mkwrite(pte_t pte)
+{
+ pte_val(pte) |= _PAGE_WRITE;
+ if (pte_val(pte) & _PAGE_MODIFIED)
+ pte_val(pte) |= _PAGE_DIRTY;
+ return pte;
+}
+
+static inline pte_t pte_mkdirty(pte_t pte)
+{
+ pte_val(pte) |= _PAGE_MODIFIED;
+ if (pte_val(pte) & _PAGE_WRITE)
+ pte_val(pte) |= _PAGE_DIRTY;
+ return pte;
+}
+
+static inline pte_t pte_mkyoung(pte_t pte)
+{
+ pte_val(pte) |= _PAGE_ACCESSED;
+ if (pte_val(pte) & _PAGE_READ)
+ pte_val(pte) |= _PAGE_VALID;
+ return pte;
+}
+
+#define __pgd_offset(address) pgd_index(address)
+#define __pud_offset(address) (((address) >> PUD_SHIFT) & (PTRS_PER_PUD-1))
+#define __pmd_offset(address) (((address) >> PMD_SHIFT) & (PTRS_PER_PMD-1))
+
+/* to find an entry in a kernel page-table-directory */
+#define pgd_offset_k(address) pgd_offset(&init_mm, address)
+
+#define pgd_index(address) ((address) >> PGDIR_SHIFT)
+
+/*
+ * Macro to make mark a page protection value as "uncacheable". Note
+ * that "protection" is really a misnomer here as the protection value
+ * contains the memory attribute bits, dirty bits, and various other
+ * bits as well.
+ */
+#define pgprot_noncached pgprot_noncached
+
+static inline pgprot_t pgprot_noncached(pgprot_t _prot)
+{
+ unsigned long prot = pgprot_val(_prot);
+
+ prot = (prot & ~_CACHE_MASK) | _CACHE_UNCACHED;
+
+ return __pgprot(prot);
+}
+
+/*
+ * Conversion functions: convert a page and protection to a page entry,
+ * and a page entry and page directory to the page they refer to.
+ */
+#define mk_pte(page, pgprot) pfn_pte(page_to_pfn(page), (pgprot))
+static inline pte_t pte_modify(pte_t pte, pgprot_t newprot)
+{
+ return __pte((pte_val(pte) & _PAGE_CHG_MASK) |
+ (pgprot_val(newprot)));
+}
+
+/* to find an entry in a page-table-directory */
+static inline pgd_t *pgd_offset(struct mm_struct *mm, unsigned long address)
+{
+ return mm->pgd + pgd_index(address);
+}
+
+/* Find an entry in the third-level page table.. */
+static inline pte_t *pte_offset(pmd_t *dir, unsigned long address)
+{
+ return (pte_t *) (pmd_page_vaddr(*dir)) +
+ ((address >> PAGE_SHIFT) & (PTRS_PER_PTE - 1));
+}
+
+extern pgd_t swapper_pg_dir[PTRS_PER_PGD];
+extern void paging_init(void);
+
+extern void show_jtlb_table(void);
+
+void update_mmu_cache(struct vm_area_struct *vma, unsigned long address,
+ pte_t *pte);
+
+/* Needs to be defined here and not in linux/mm.h, as it is arch dependent */
+#define kern_addr_valid(addr) (1)
+
+/*
+ * No page table caches to initialise
+ */
+#define pgtable_cache_init() do {} while (0)
+
+#define io_remap_pfn_range(vma, vaddr, pfn, size, prot) \
+ remap_pfn_range(vma, vaddr, pfn, size, prot)
+
+#include <asm-generic/pgtable.h>
+
+#endif /* __ASM_CSKY_PGTABLE_H */
diff --git a/arch/csky/include/asm/segment.h b/arch/csky/include/asm/segment.h
new file mode 100644
index 0000000..ffdc4c4
--- /dev/null
+++ b/arch/csky/include/asm/segment.h
@@ -0,0 +1,19 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+// Copyright (C) 2018 Hangzhou C-SKY Microsystems co.,ltd.
+
+#ifndef __ASM_CSKY_SEGMENT_H
+#define __ASM_CSKY_SEGMENT_H
+
+typedef struct {
+ unsigned long seg;
+} mm_segment_t;
+
+#define KERNEL_DS ((mm_segment_t) { 0xFFFFFFFF })
+#define get_ds() KERNEL_DS
+
+#define USER_DS ((mm_segment_t) { 0x80000000UL })
+#define get_fs() (current_thread_info()->addr_limit)
+#define set_fs(x) (current_thread_info()->addr_limit = (x))
+#define segment_eq(a, b) ((a).seg == (b).seg)
+
+#endif /* __ASM_CSKY_SEGMENT_H */
diff --git a/arch/csky/include/asm/shmparam.h b/arch/csky/include/asm/shmparam.h
new file mode 100644
index 0000000..efafe4c
--- /dev/null
+++ b/arch/csky/include/asm/shmparam.h
@@ -0,0 +1,11 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+// Copyright (C) 2018 Hangzhou C-SKY Microsystems co.,ltd.
+
+#ifndef __ASM_CSKY_SHMPARAM_H
+#define __ASM_CSKY_SHMPARAM_H
+
+#define SHMLBA (4 * PAGE_SIZE)
+
+#define __ARCH_FORCE_SHMLBA
+
+#endif /* __ASM_CSKY_SHMPARAM_H */
diff --git a/arch/csky/mm/dma-mapping.c b/arch/csky/mm/dma-mapping.c
new file mode 100644
index 0000000..85437b2
--- /dev/null
+++ b/arch/csky/mm/dma-mapping.c
@@ -0,0 +1,254 @@
+// SPDX-License-Identifier: GPL-2.0
+// Copyright (C) 2018 Hangzhou C-SKY Microsystems co.,ltd.
+
+#include <linux/cache.h>
+#include <linux/dma-mapping.h>
+#include <linux/dma-contiguous.h>
+#include <linux/dma-noncoherent.h>
+#include <linux/genalloc.h>
+#include <linux/highmem.h>
+#include <linux/io.h>
+#include <linux/mm.h>
+#include <linux/scatterlist.h>
+#include <linux/types.h>
+#include <linux/version.h>
+#include <asm/cache.h>
+
+static struct gen_pool *atomic_pool;
+static size_t atomic_pool_size __initdata = SZ_256K;
+
+static int __init early_coherent_pool(char *p)
+{
+ atomic_pool_size = memparse(p, &p);
+ return 0;
+}
+early_param("coherent_pool", early_coherent_pool);
+
+static int __init atomic_pool_init(void)
+{
+ struct page *page;
+ size_t size = atomic_pool_size;
+ void *ptr;
+ int ret;
+
+ atomic_pool = gen_pool_create(PAGE_SHIFT, -1);
+ if (!atomic_pool)
+ BUG();
+
+ page = alloc_pages(GFP_KERNEL | GFP_DMA, get_order(size));
+ if (!page)
+ BUG();
+
+ ptr = dma_common_contiguous_remap(page, size, VM_ALLOC,
+ pgprot_noncached(PAGE_KERNEL),
+ __builtin_return_address(0));
+ if (!ptr)
+ BUG();
+
+ ret = gen_pool_add_virt(atomic_pool, (unsigned long)ptr,
+ page_to_phys(page), atomic_pool_size, -1);
+ if (ret)
+ BUG();
+
+ gen_pool_set_algo(atomic_pool, gen_pool_first_fit_order_align, NULL);
+
+ pr_info("DMA: preallocated %zu KiB pool for atomic coherent pool\n",
+ atomic_pool_size / 1024);
+
+ pr_info("DMA: vaddr: 0x%x phy: 0x%lx,\n", (unsigned int)ptr,
+ page_to_phys(page));
+
+ return 0;
+}
+postcore_initcall(atomic_pool_init);
+
+static void *csky_dma_alloc_atomic(struct device *dev, size_t size,
+ dma_addr_t *dma_handle)
+{
+ unsigned long addr;
+
+ addr = gen_pool_alloc(atomic_pool, size);
+ if (addr)
+ *dma_handle = gen_pool_virt_to_phys(atomic_pool, addr);
+
+ return (void *)addr;
+}
+
+static void csky_dma_free_atomic(struct device *dev, size_t size, void *vaddr,
+ dma_addr_t dma_handle, unsigned long attrs)
+{
+ gen_pool_free(atomic_pool, (unsigned long)vaddr, size);
+}
+
+static void __dma_clear_buffer(struct page *page, size_t size)
+{
+ if (PageHighMem(page)) {
+ unsigned int count = PAGE_ALIGN(size) >> PAGE_SHIFT;
+
+ do {
+ void *ptr = kmap_atomic(page);
+ size_t _size = (size < PAGE_SIZE) ? size : PAGE_SIZE;
+
+ memset(ptr, 0, _size);
+ dma_wbinv_range((unsigned long)ptr,
+ (unsigned long)ptr + _size);
+
+ kunmap_atomic(ptr);
+
+ page++;
+ size -= PAGE_SIZE;
+ count--;
+ } while (count);
+ } else {
+ void *ptr = page_address(page);
+
+ memset(ptr, 0, size);
+ dma_wbinv_range((unsigned long)ptr, (unsigned long)ptr + size);
+ }
+}
+
+static void *csky_dma_alloc_nonatomic(struct device *dev, size_t size,
+ dma_addr_t *dma_handle, gfp_t gfp,
+ unsigned long attrs)
+{
+ void *vaddr;
+ struct page *page;
+ unsigned int count = PAGE_ALIGN(size) >> PAGE_SHIFT;
+
+ if (DMA_ATTR_NON_CONSISTENT & attrs) {
+ pr_err("csky %s can't support DMA_ATTR_NON_CONSISTENT.\n", __func__);
+ return NULL;
+ }
+
+ if (IS_ENABLED(CONFIG_DMA_CMA))
+ page = dma_alloc_from_contiguous(dev, count, get_order(size),
+ gfp);
+ else
+ page = alloc_pages(gfp, get_order(size));
+
+ if (!page) {
+ pr_err("csky %s no more free pages.\n", __func__);
+ return NULL;
+ }
+
+ *dma_handle = page_to_phys(page);
+
+ __dma_clear_buffer(page, size);
+
+ if (attrs & DMA_ATTR_NO_KERNEL_MAPPING)
+ return page;
+
+ vaddr = dma_common_contiguous_remap(page, PAGE_ALIGN(size), VM_USERMAP,
+ pgprot_noncached(PAGE_KERNEL), __builtin_return_address(0));
+ if (!vaddr)
+ BUG();
+
+ return vaddr;
+}
+
+static void csky_dma_free_nonatomic(
+ struct device *dev,
+ size_t size,
+ void *vaddr,
+ dma_addr_t dma_handle,
+ unsigned long attrs
+ )
+{
+ struct page *page = phys_to_page(dma_handle);
+ unsigned int count = PAGE_ALIGN(size) >> PAGE_SHIFT;
+
+ if ((unsigned int)vaddr >= VMALLOC_START)
+ dma_common_free_remap(vaddr, size, VM_USERMAP);
+
+ if (IS_ENABLED(CONFIG_DMA_CMA))
+ dma_release_from_contiguous(dev, page, count);
+ else
+ __free_pages(page, get_order(size));
+}
+
+void *arch_dma_alloc(struct device *dev, size_t size, dma_addr_t *dma_handle,
+ gfp_t gfp, unsigned long attrs)
+{
+ if (gfpflags_allow_blocking(gfp))
+ return csky_dma_alloc_nonatomic(dev, size, dma_handle, gfp,
+ attrs);
+ else
+ return csky_dma_alloc_atomic(dev, size, dma_handle);
+}
+
+void arch_dma_free(struct device *dev, size_t size, void *vaddr,
+ dma_addr_t dma_handle, unsigned long attrs)
+{
+ if (!addr_in_gen_pool(atomic_pool, (unsigned int) vaddr, size))
+ csky_dma_free_nonatomic(dev, size, vaddr, dma_handle, attrs);
+ else
+ csky_dma_free_atomic(dev, size, vaddr, dma_handle, attrs);
+}
+
+static inline void cache_op(phys_addr_t paddr, size_t size,
+ void (*fn)(unsigned long start, unsigned long end))
+{
+ struct page *page = pfn_to_page(paddr >> PAGE_SHIFT);
+ unsigned int offset = paddr & ~PAGE_MASK;
+ size_t left = size;
+ unsigned long start;
+
+ do {
+ size_t len = left;
+
+ if (PageHighMem(page)) {
+ void *addr;
+
+ if (offset + len > PAGE_SIZE) {
+ if (offset >= PAGE_SIZE) {
+ page += offset >> PAGE_SHIFT;
+ offset &= ~PAGE_MASK;
+ }
+ len = PAGE_SIZE - offset;
+ }
+
+ addr = kmap_atomic(page);
+ start = (unsigned long)(addr + offset);
+ fn(start, start + len);
+ kunmap_atomic(addr);
+ } else {
+ start = (unsigned long)phys_to_virt(paddr);
+ fn(start, start + size);
+ }
+ offset = 0;
+ page++;
+ left -= len;
+ } while (left);
+}
+
+void arch_sync_dma_for_device(struct device *dev, phys_addr_t paddr,
+ size_t size, enum dma_data_direction dir)
+{
+ switch (dir) {
+ case DMA_TO_DEVICE:
+ cache_op(paddr, size, dma_wb_range);
+ break;
+ case DMA_FROM_DEVICE:
+ case DMA_BIDIRECTIONAL:
+ cache_op(paddr, size, dma_wbinv_range);
+ break;
+ default:
+ BUG();
+ }
+}
+
+void arch_sync_dma_for_cpu(struct device *dev, phys_addr_t paddr,
+ size_t size, enum dma_data_direction dir)
+{
+ switch (dir) {
+ case DMA_TO_DEVICE:
+ cache_op(paddr, size, dma_wb_range);
+ break;
+ case DMA_FROM_DEVICE:
+ case DMA_BIDIRECTIONAL:
+ cache_op(paddr, size, dma_wbinv_range);
+ break;
+ default:
+ BUG();
+ }
+}
diff --git a/arch/csky/mm/highmem.c b/arch/csky/mm/highmem.c
new file mode 100644
index 0000000..e168ac0
--- /dev/null
+++ b/arch/csky/mm/highmem.c
@@ -0,0 +1,198 @@
+// SPDX-License-Identifier: GPL-2.0
+// Copyright (C) 2018 Hangzhou C-SKY Microsystems co.,ltd.
+
+#include <linux/module.h>
+#include <linux/highmem.h>
+#include <linux/smp.h>
+#include <linux/bootmem.h>
+#include <asm/fixmap.h>
+#include <asm/tlbflush.h>
+#include <asm/cacheflush.h>
+
+static pte_t *kmap_pte;
+
+unsigned long highstart_pfn, highend_pfn;
+
+void *kmap(struct page *page)
+{
+ void *addr;
+
+ might_sleep();
+ if (!PageHighMem(page))
+ return page_address(page);
+ addr = kmap_high(page);
+ flush_tlb_one((unsigned long)addr);
+
+ return addr;
+}
+EXPORT_SYMBOL(kmap);
+
+void kunmap(struct page *page)
+{
+ BUG_ON(in_interrupt());
+ if (!PageHighMem(page))
+ return;
+ kunmap_high(page);
+}
+EXPORT_SYMBOL(kunmap);
+
+void *kmap_atomic(struct page *page)
+{
+ unsigned long vaddr;
+ int idx, type;
+
+ preempt_disable();
+ pagefault_disable();
+ if (!PageHighMem(page))
+ return page_address(page);
+
+ type = kmap_atomic_idx_push();
+ idx = type + KM_TYPE_NR*smp_processor_id();
+ vaddr = __fix_to_virt(FIX_KMAP_BEGIN + idx);
+#ifdef CONFIG_DEBUG_HIGHMEM
+ BUG_ON(!pte_none(*(kmap_pte - idx)));
+#endif
+ set_pte(kmap_pte-idx, mk_pte(page, PAGE_KERNEL));
+ flush_tlb_one((unsigned long)vaddr);
+
+ return (void *)vaddr;
+}
+EXPORT_SYMBOL(kmap_atomic);
+
+void __kunmap_atomic(void *kvaddr)
+{
+ unsigned long vaddr = (unsigned long) kvaddr & PAGE_MASK;
+ int idx;
+
+ if (vaddr < FIXADDR_START)
+ goto out;
+
+#ifdef CONFIG_DEBUG_HIGHMEM
+ idx = KM_TYPE_NR*smp_processor_id() + kmap_atomic_idx();
+
+ BUG_ON(vaddr != __fix_to_virt(FIX_KMAP_BEGIN + idx));
+
+ pte_clear(&init_mm, vaddr, kmap_pte - idx);
+ flush_tlb_one(vaddr);
+#else
+ (void) idx; /* to kill a warning */
+#endif
+ kmap_atomic_idx_pop();
+out:
+ pagefault_enable();
+ preempt_enable();
+}
+EXPORT_SYMBOL(__kunmap_atomic);
+
+/*
+ * This is the same as kmap_atomic() but can map memory that doesn't
+ * have a struct page associated with it.
+ */
+void *kmap_atomic_pfn(unsigned long pfn)
+{
+ unsigned long vaddr;
+ int idx, type;
+
+ pagefault_disable();
+
+ type = kmap_atomic_idx_push();
+ idx = type + KM_TYPE_NR*smp_processor_id();
+ vaddr = __fix_to_virt(FIX_KMAP_BEGIN + idx);
+ set_pte(kmap_pte-idx, pfn_pte(pfn, PAGE_KERNEL));
+ flush_tlb_one(vaddr);
+
+ return (void *) vaddr;
+}
+
+struct page *kmap_atomic_to_page(void *ptr)
+{
+ unsigned long idx, vaddr = (unsigned long)ptr;
+ pte_t *pte;
+
+ if (vaddr < FIXADDR_START)
+ return virt_to_page(ptr);
+
+ idx = virt_to_fix(vaddr);
+ pte = kmap_pte - (idx - FIX_KMAP_BEGIN);
+ return pte_page(*pte);
+}
+
+static void __init fixrange_init(unsigned long start, unsigned long end,
+ pgd_t *pgd_base)
+{
+#ifdef CONFIG_HIGHMEM
+ pgd_t *pgd;
+ pud_t *pud;
+ pmd_t *pmd;
+ pte_t *pte;
+ int i, j, k;
+ unsigned long vaddr;
+
+ vaddr = start;
+ i = __pgd_offset(vaddr);
+ j = __pud_offset(vaddr);
+ k = __pmd_offset(vaddr);
+ pgd = pgd_base + i;
+
+ for ( ; (i < PTRS_PER_PGD) && (vaddr != end); pgd++, i++) {
+ pud = (pud_t *)pgd;
+ for ( ; (j < PTRS_PER_PUD) && (vaddr != end); pud++, j++) {
+ pmd = (pmd_t *)pud;
+ for (; (k < PTRS_PER_PMD) && (vaddr != end); pmd++, k++) {
+ if (pmd_none(*pmd)) {
+ pte = (pte_t *) alloc_bootmem_low_pages(PAGE_SIZE);
+ set_pmd(pmd, __pmd(__pa(pte)));
+ BUG_ON(pte != pte_offset_kernel(pmd, 0));
+ }
+ vaddr += PMD_SIZE;
+ }
+ k = 0;
+ }
+ j = 0;
+ }
+#endif
+}
+
+void __init fixaddr_kmap_pages_init(void)
+{
+ unsigned long vaddr;
+ pgd_t *pgd_base;
+#ifdef CONFIG_HIGHMEM
+ pgd_t *pgd;
+ pmd_t *pmd;
+ pud_t *pud;
+ pte_t *pte;
+#endif
+ pgd_base = swapper_pg_dir;
+
+ /*
+ * Fixed mappings:
+ */
+ vaddr = __fix_to_virt(__end_of_fixed_addresses - 1) & PMD_MASK;
+ fixrange_init(vaddr, 0, pgd_base);
+
+#ifdef CONFIG_HIGHMEM
+ /*
+ * Permanent kmaps:
+ */
+ vaddr = PKMAP_BASE;
+ fixrange_init(vaddr, vaddr + PAGE_SIZE*LAST_PKMAP, pgd_base);
+
+ pgd = swapper_pg_dir + __pgd_offset(vaddr);
+ pud = (pud_t *)pgd;
+ pmd = pmd_offset(pud, vaddr);
+ pte = pte_offset_kernel(pmd, vaddr);
+ pkmap_page_table = pte;
+#endif
+}
+
+void __init kmap_init(void)
+{
+ unsigned long vaddr;
+
+ fixaddr_kmap_pages_init();
+
+ vaddr = __fix_to_virt(FIX_KMAP_BEGIN);
+
+ kmap_pte = pte_offset_kernel((pmd_t *)pgd_offset_k(vaddr), vaddr);
+}
diff --git a/arch/csky/mm/init.c b/arch/csky/mm/init.c
new file mode 100644
index 0000000..ce2711e
--- /dev/null
+++ b/arch/csky/mm/init.c
@@ -0,0 +1,122 @@
+// SPDX-License-Identifier: GPL-2.0
+// Copyright (C) 2018 Hangzhou C-SKY Microsystems co.,ltd.
+
+#include <linux/bug.h>
+#include <linux/module.h>
+#include <linux/init.h>
+#include <linux/signal.h>
+#include <linux/sched.h>
+#include <linux/kernel.h>
+#include <linux/errno.h>
+#include <linux/string.h>
+#include <linux/types.h>
+#include <linux/pagemap.h>
+#include <linux/ptrace.h>
+#include <linux/mman.h>
+#include <linux/mm.h>
+#include <linux/bootmem.h>
+#include <linux/highmem.h>
+#include <linux/memblock.h>
+#include <linux/swap.h>
+#include <linux/proc_fs.h>
+#include <linux/pfn.h>
+
+#include <asm/setup.h>
+#include <asm/cachectl.h>
+#include <asm/dma.h>
+#include <asm/pgtable.h>
+#include <asm/pgalloc.h>
+#include <asm/mmu_context.h>
+#include <asm/sections.h>
+#include <asm/tlb.h>
+
+pgd_t swapper_pg_dir[PTRS_PER_PGD] __page_aligned_bss;
+pte_t invalid_pte_table[PTRS_PER_PTE] __page_aligned_bss;
+unsigned long empty_zero_page[PAGE_SIZE / sizeof(unsigned long)]
+ __page_aligned_bss;
+EXPORT_SYMBOL(empty_zero_page);
+
+void __init mem_init(void)
+{
+#ifdef CONFIG_HIGHMEM
+ unsigned long tmp;
+
+ max_mapnr = highend_pfn;
+#else
+ max_mapnr = max_low_pfn;
+#endif
+ high_memory = (void *) __va(max_low_pfn << PAGE_SHIFT);
+
+ free_all_bootmem();
+
+#ifdef CONFIG_HIGHMEM
+ for (tmp = highstart_pfn; tmp < highend_pfn; tmp++) {
+ struct page *page = pfn_to_page(tmp);
+
+ /* FIXME not sure about */
+ if (!memblock_is_reserved(tmp << PAGE_SHIFT))
+ free_highmem_page(page);
+ }
+#endif
+ mem_init_print_info(NULL);
+}
+
+#ifdef CONFIG_BLK_DEV_INITRD
+void free_initrd_mem(unsigned long start, unsigned long end)
+{
+ if (start < end)
+ pr_info("Freeing initrd memory: %ldk freed\n",
+ (end - start) >> 10);
+
+ for (; start < end; start += PAGE_SIZE) {
+ ClearPageReserved(virt_to_page(start));
+ init_page_count(virt_to_page(start));
+ free_page(start);
+ totalram_pages++;
+ }
+}
+#endif
+
+extern char __init_begin[], __init_end[];
+
+void free_initmem(void)
+{
+ unsigned long addr;
+
+ addr = (unsigned long) &__init_begin;
+
+ while (addr < (unsigned long) &__init_end) {
+ ClearPageReserved(virt_to_page(addr));
+ init_page_count(virt_to_page(addr));
+ free_page(addr);
+ totalram_pages++;
+ addr += PAGE_SIZE;
+ }
+
+ pr_info("Freeing unused kernel memory: %dk freed\n",
+ ((unsigned int)&__init_end - (unsigned int)&__init_begin) >> 10);
+}
+
+void pgd_init(unsigned long *p)
+{
+ int i;
+
+ for (i = 0; i < PTRS_PER_PGD; i++)
+ p[i] = __pa(invalid_pte_table);
+}
+
+void __init pre_mmu_init(void)
+{
+ /*
+ * Setup page-table and enable TLB-hardrefill
+ */
+ flush_tlb_all();
+ pgd_init((unsigned long *)swapper_pg_dir);
+ TLBMISS_HANDLER_SETUP_PGD(swapper_pg_dir);
+ TLBMISS_HANDLER_SETUP_PGD_KERNEL(swapper_pg_dir);
+
+ asid_cache(smp_processor_id()) = ASID_FIRST_VERSION;
+
+ /* Setup page mask to 4k */
+ write_mmu_pagemask(0);
+}
diff --git a/arch/csky/mm/ioremap.c b/arch/csky/mm/ioremap.c
new file mode 100644
index 0000000..7ad3ff1
--- /dev/null
+++ b/arch/csky/mm/ioremap.c
@@ -0,0 +1,48 @@
+// SPDX-License-Identifier: GPL-2.0
+// Copyright (C) 2018 Hangzhou C-SKY Microsystems co.,ltd.
+
+#include <linux/export.h>
+#include <linux/mm.h>
+#include <linux/vmalloc.h>
+#include <linux/io.h>
+
+#include <asm/pgtable.h>
+
+void __iomem *ioremap(phys_addr_t addr, size_t size)
+{
+ phys_addr_t last_addr;
+ unsigned long offset, vaddr;
+ struct vm_struct *area;
+ pgprot_t prot;
+
+ last_addr = addr + size - 1;
+ if (!size || last_addr < addr)
+ return NULL;
+
+ offset = addr & (~PAGE_MASK);
+ addr &= PAGE_MASK;
+ size = PAGE_ALIGN(size + offset);
+
+ area = get_vm_area_caller(size, VM_ALLOC, __builtin_return_address(0));
+ if (!area)
+ return NULL;
+
+ vaddr = (unsigned long)area->addr;
+
+ prot = __pgprot(_PAGE_PRESENT | __READABLE | __WRITEABLE |
+ _PAGE_GLOBAL | _CACHE_UNCACHED);
+
+ if (ioremap_page_range(vaddr, vaddr + size, addr, prot)) {
+ free_vm_area(area);
+ return NULL;
+ }
+
+ return (void __iomem *)(vaddr + offset);
+}
+EXPORT_SYMBOL(ioremap);
+
+void iounmap(void __iomem *addr)
+{
+ vunmap((void *)((unsigned long)addr & PAGE_MASK));
+}
+EXPORT_SYMBOL(iounmap);
--
2.7.4
^ permalink raw reply related [flat|nested] 68+ messages in thread* Re: [PATCH V9 07/21] csky: MMU and page table management
2018-10-16 2:58 ` [PATCH V9 07/21] csky: MMU and page table management Guo Ren
@ 2018-10-17 15:06 ` Arnd Bergmann
2018-10-18 2:05 ` Guo Ren
0 siblings, 1 reply; 68+ messages in thread
From: Arnd Bergmann @ 2018-10-17 15:06 UTC (permalink / raw)
To: Guo Ren
Cc: Andrew Morton, Daniel Lezcano, David Miller, gregkh,
Christoph Hellwig, Marc Zyngier, Mark Rutland, Peter Zijlstra,
Rob Herring, Thomas Gleixner, Linux Kernel Mailing List,
linux-arch, DTML, Rob Herring, c-sky_gcc_upstream
On Tue, Oct 16, 2018 at 5:01 AM Guo Ren <ren_guo@c-sky.com> wrote:
>
> This patch adds files related to memory management and here is our
> memory-layout:
>
> Fixmap : 0xffc02000 – 0xfffff000 (4 MB - 12KB)
> Pkmap : 0xff800000 – 0xffc00000 (4 MB)
> Vmalloc : 0xf0200000 – 0xff000000 (238 MB)
> Lowmem : 0x80000000 – 0xc0000000 (1GB)
>
> abiv1 CPU (CK610) is VIPT cache and it doesn't support highmem.
> abiv2 CPUs are all PIPT cache and they could support highmem.
>
> Lowmem is directly mapped by msa0 & msa1 reg, and we needn't setup
> memory page table for it.
>
> Link:https://lore.kernel.org/lkml/20180518215548.GH17671@n2100.armlinux.org.uk/
> Signed-off-by: Guo Ren <ren_guo@c-sky.com>
> Cc: Christoph Hellwig <hch@infradead.org>
> Cc: Arnd Bergmann <arnd@arndb.de>
Reviewed-by: Arnd Bergmann <arnd@arndb.de>
Christoph had all the useful comments on this one, I just checked that I didn't
spot anything beyond that.
Arnd
^ permalink raw reply [flat|nested] 68+ messages in thread* Re: [PATCH V9 07/21] csky: MMU and page table management
2018-10-17 15:06 ` Arnd Bergmann
@ 2018-10-18 2:05 ` Guo Ren
0 siblings, 0 replies; 68+ messages in thread
From: Guo Ren @ 2018-10-18 2:05 UTC (permalink / raw)
To: Arnd Bergmann
Cc: Andrew Morton, Daniel Lezcano, David Miller, gregkh,
Christoph Hellwig, Marc Zyngier, Mark Rutland, Peter Zijlstra,
Rob Herring, Thomas Gleixner, Linux Kernel Mailing List,
linux-arch, DTML, Rob Herring, c-sky_gcc_upstream
On Wed, Oct 17, 2018 at 05:06:56PM +0200, Arnd Bergmann wrote:
> On Tue, Oct 16, 2018 at 5:01 AM Guo Ren <ren_guo@c-sky.com> wrote:
> >
> > This patch adds files related to memory management and here is our
> > memory-layout:
> >
> > Fixmap : 0xffc02000 – 0xfffff000 (4 MB - 12KB)
> > Pkmap : 0xff800000 – 0xffc00000 (4 MB)
> > Vmalloc : 0xf0200000 – 0xff000000 (238 MB)
> > Lowmem : 0x80000000 – 0xc0000000 (1GB)
> >
> > abiv1 CPU (CK610) is VIPT cache and it doesn't support highmem.
> > abiv2 CPUs are all PIPT cache and they could support highmem.
> >
> > Lowmem is directly mapped by msa0 & msa1 reg, and we needn't setup
> > memory page table for it.
> >
> > Link:https://lore.kernel.org/lkml/20180518215548.GH17671@n2100.armlinux.org.uk/
> > Signed-off-by: Guo Ren <ren_guo@c-sky.com>
> > Cc: Christoph Hellwig <hch@infradead.org>
> > Cc: Arnd Bergmann <arnd@arndb.de>
>
> Reviewed-by: Arnd Bergmann <arnd@arndb.de>
>
> Christoph had all the useful comments on this one, I just checked that I didn't
> spot anything beyond that.
Yes, Christoph help a lot and find a critical bug of "sync_dma > 4K".
Best Regards
Guo Ren
^ permalink raw reply [flat|nested] 68+ messages in thread
* [PATCH V9 08/21] csky: Process management and Signal
2018-10-16 2:58 [PATCH V9 00/21] C-SKY(csky) Linux Kernel Port Guo Ren
` (6 preceding siblings ...)
2018-10-16 2:58 ` [PATCH V9 07/21] csky: MMU and page table management Guo Ren
@ 2018-10-16 2:58 ` Guo Ren
2018-10-17 15:11 ` Arnd Bergmann
2018-10-16 2:58 ` [PATCH V9 09/21] csky: VDSO and rt_sigreturn Guo Ren
` (13 subsequent siblings)
21 siblings, 1 reply; 68+ messages in thread
From: Guo Ren @ 2018-10-16 2:58 UTC (permalink / raw)
To: akpm, arnd, daniel.lezcano, davem, gregkh, hch, marc.zyngier,
mark.rutland, peterz, robh, tglx
Cc: linux-kernel, linux-arch, devicetree, robh+dt,
c-sky_gcc_upstream, Guo Ren
This patch adds files related to task_switch, sigcontext, signal,
fpu context switch.
Signed-off-by: Guo Ren <ren_guo@c-sky.com>
Cc: Arnd Bergmann <arnd@arndb.de>
---
Changelog:
- abiv2/fpu.c: Userspace should never be sent NSIGXXX as a si_code.
Use FPE_FLTUNK instead.
- abiv2/fpu.c: Use force_sig_fault instead.
---
---
arch/csky/abiv2/fpu.c | 275 +++++++++++++++++++++++++
arch/csky/abiv2/inc/abi/fpu.h | 66 ++++++
arch/csky/include/asm/mmu_context.h | 150 ++++++++++++++
arch/csky/include/asm/processor.h | 121 +++++++++++
arch/csky/include/asm/switch_to.h | 36 ++++
arch/csky/include/asm/thread_info.h | 75 +++++++
arch/csky/include/uapi/asm/sigcontext.h | 14 ++
arch/csky/kernel/process.c | 136 ++++++++++++
arch/csky/kernel/signal.c | 355 ++++++++++++++++++++++++++++++++
arch/csky/kernel/time.c | 11 +
10 files changed, 1239 insertions(+)
create mode 100644 arch/csky/abiv2/fpu.c
create mode 100644 arch/csky/abiv2/inc/abi/fpu.h
create mode 100644 arch/csky/include/asm/mmu_context.h
create mode 100644 arch/csky/include/asm/processor.h
create mode 100644 arch/csky/include/asm/switch_to.h
create mode 100644 arch/csky/include/asm/thread_info.h
create mode 100644 arch/csky/include/uapi/asm/sigcontext.h
create mode 100644 arch/csky/kernel/process.c
create mode 100644 arch/csky/kernel/signal.c
create mode 100644 arch/csky/kernel/time.c
diff --git a/arch/csky/abiv2/fpu.c b/arch/csky/abiv2/fpu.c
new file mode 100644
index 0000000..e7e1134
--- /dev/null
+++ b/arch/csky/abiv2/fpu.c
@@ -0,0 +1,275 @@
+// SPDX-License-Identifier: GPL-2.0
+// Copyright (C) 2018 Hangzhou C-SKY Microsystems co.,ltd.
+
+#include <linux/ptrace.h>
+#include <linux/uaccess.h>
+#include <abi/reg_ops.h>
+
+#define MTCR_MASK 0xFC00FFE0
+#define MFCR_MASK 0xFC00FFE0
+#define MTCR_DIST 0xC0006420
+#define MFCR_DIST 0xC0006020
+
+void __init init_fpu(void)
+{
+ mtcr("cr<1, 2>", 0);
+}
+
+/*
+ * fpu_libc_helper() is to help libc to excute:
+ * - mfcr %a, cr<1, 2>
+ * - mfcr %a, cr<2, 2>
+ * - mtcr %a, cr<1, 2>
+ * - mtcr %a, cr<2, 2>
+ */
+int fpu_libc_helper(struct pt_regs *regs)
+{
+ int fault;
+ unsigned long instrptr, regx = 0;
+ unsigned long index = 0, tmp = 0;
+ unsigned long tinstr = 0;
+ u16 instr_hi, instr_low;
+
+ instrptr = instruction_pointer(regs);
+ if (instrptr & 1)
+ return 0;
+
+ fault = __get_user(instr_low, (u16 *)instrptr);
+ if (fault)
+ return 0;
+
+ fault = __get_user(instr_hi, (u16 *)(instrptr + 2));
+ if (fault)
+ return 0;
+
+ tinstr = instr_hi | ((unsigned long)instr_low << 16);
+
+ if (((tinstr >> 21) & 0x1F) != 2)
+ return 0;
+
+ if ((tinstr & MTCR_MASK) == MTCR_DIST) {
+ index = (tinstr >> 16) & 0x1F;
+ if (index > 13)
+ return 0;
+
+ tmp = tinstr & 0x1F;
+ if (tmp > 2)
+ return 0;
+
+ regx = *(®s->a0 + index);
+
+ if (tmp == 1)
+ mtcr("cr<1, 2>", regx);
+ else if (tmp == 2)
+ mtcr("cr<2, 2>", regx);
+ else
+ return 0;
+
+ regs->pc += 4;
+ return 1;
+ }
+
+ if ((tinstr & MFCR_MASK) == MFCR_DIST) {
+ index = tinstr & 0x1F;
+ if (index > 13)
+ return 0;
+
+ tmp = ((tinstr >> 16) & 0x1F);
+ if (tmp > 2)
+ return 0;
+
+ if (tmp == 1)
+ regx = mfcr("cr<1, 2>");
+ else if (tmp == 2)
+ regx = mfcr("cr<2, 2>");
+ else
+ return 0;
+
+ *(®s->a0 + index) = regx;
+
+ regs->pc += 4;
+ return 1;
+ }
+
+ return 0;
+}
+
+void fpu_fpe(struct pt_regs *regs)
+{
+ int sig, code;
+ unsigned int fesr;
+
+ fesr = mfcr("cr<2, 2>");
+
+ sig = SIGFPE;
+ code = FPE_FLTUNK;
+
+ if (fesr & FPE_ILLE) {
+ sig = SIGILL;
+ code = ILL_ILLOPC;
+ } else if (fesr & FPE_IDC) {
+ sig = SIGILL;
+ code = ILL_ILLOPN;
+ } else if (fesr & FPE_FEC) {
+ sig = SIGFPE;
+ if (fesr & FPE_IOC)
+ code = FPE_FLTINV;
+ else if (fesr & FPE_DZC)
+ code = FPE_FLTDIV;
+ else if (fesr & FPE_UFC)
+ code = FPE_FLTUND;
+ else if (fesr & FPE_OFC)
+ code = FPE_FLTOVF;
+ else if (fesr & FPE_IXC)
+ code = FPE_FLTRES;
+ }
+
+ force_sig_fault(sig, code, (void __user *)regs->pc, current);
+}
+
+#define FMFVR_FPU_REGS(vrx, vry) \
+ "fmfvrl %0, "#vrx"\n" \
+ "fmfvrh %1, "#vrx"\n" \
+ "fmfvrl %2, "#vry"\n" \
+ "fmfvrh %3, "#vry"\n"
+
+#define FMTVR_FPU_REGS(vrx, vry) \
+ "fmtvrl "#vrx", %0\n" \
+ "fmtvrh "#vrx", %1\n" \
+ "fmtvrl "#vry", %2\n" \
+ "fmtvrh "#vry", %3\n"
+
+#define STW_FPU_REGS(a, b, c, d) \
+ "stw %0, (%4, "#a")\n" \
+ "stw %1, (%4, "#b")\n" \
+ "stw %2, (%4, "#c")\n" \
+ "stw %3, (%4, "#d")\n"
+
+#define LDW_FPU_REGS(a, b, c, d) \
+ "ldw %0, (%4, "#a")\n" \
+ "ldw %1, (%4, "#b")\n" \
+ "ldw %2, (%4, "#c")\n" \
+ "ldw %3, (%4, "#d")\n"
+
+void save_to_user_fp(struct user_fp *user_fp)
+{
+ unsigned long flg;
+ unsigned long tmp1, tmp2;
+ unsigned long *fpregs;
+
+ local_irq_save(flg);
+
+ tmp1 = mfcr("cr<1, 2>");
+ tmp2 = mfcr("cr<2, 2>");
+
+ user_fp->fcr = tmp1;
+ user_fp->fesr = tmp2;
+
+ fpregs = &user_fp->vr[0];
+#ifdef CONFIG_CPU_HAS_FPUV2
+#ifdef CONFIG_CPU_HAS_VDSP
+ asm volatile(
+ "vstmu.32 vr0-vr3, (%0)\n"
+ "vstmu.32 vr4-vr7, (%0)\n"
+ "vstmu.32 vr8-vr11, (%0)\n"
+ "vstmu.32 vr12-vr15, (%0)\n"
+ "fstmu.64 vr16-vr31, (%0)\n"
+ : "+a"(fpregs)
+ ::"memory");
+#else
+ asm volatile(
+ "fstmu.64 vr0-vr31, (%0)\n"
+ : "+a"(fpregs)
+ ::"memory");
+#endif
+#else
+ {
+ unsigned long tmp3, tmp4;
+
+ asm volatile(
+ FMFVR_FPU_REGS(vr0, vr1)
+ STW_FPU_REGS(0, 4, 16, 20)
+ FMFVR_FPU_REGS(vr2, vr3)
+ STW_FPU_REGS(32, 36, 48, 52)
+ FMFVR_FPU_REGS(vr4, vr5)
+ STW_FPU_REGS(64, 68, 80, 84)
+ FMFVR_FPU_REGS(vr6, vr7)
+ STW_FPU_REGS(96, 100, 112, 116)
+ "addi %4, 128\n"
+ FMFVR_FPU_REGS(vr8, vr9)
+ STW_FPU_REGS(0, 4, 16, 20)
+ FMFVR_FPU_REGS(vr10, vr11)
+ STW_FPU_REGS(32, 36, 48, 52)
+ FMFVR_FPU_REGS(vr12, vr13)
+ STW_FPU_REGS(64, 68, 80, 84)
+ FMFVR_FPU_REGS(vr14, vr15)
+ STW_FPU_REGS(96, 100, 112, 116)
+ : "=a"(tmp1), "=a"(tmp2), "=a"(tmp3),
+ "=a"(tmp4), "+a"(fpregs)
+ ::"memory");
+ }
+#endif
+
+ local_irq_restore(flg);
+}
+
+void restore_from_user_fp(struct user_fp *user_fp)
+{
+ unsigned long flg;
+ unsigned long tmp1, tmp2;
+ unsigned long *fpregs;
+
+ local_irq_save(flg);
+
+ tmp1 = user_fp->fcr;
+ tmp2 = user_fp->fesr;
+
+ mtcr("cr<1, 2>", tmp1);
+ mtcr("cr<2, 2>", tmp2);
+
+ fpregs = &user_fp->vr[0];
+#ifdef CONFIG_CPU_HAS_FPUV2
+#ifdef CONFIG_CPU_HAS_VDSP
+ asm volatile(
+ "vldmu.32 vr0-vr3, (%0)\n"
+ "vldmu.32 vr4-vr7, (%0)\n"
+ "vldmu.32 vr8-vr11, (%0)\n"
+ "vldmu.32 vr12-vr15, (%0)\n"
+ "fldmu.64 vr16-vr31, (%0)\n"
+ : "+a"(fpregs)
+ ::"memory");
+#else
+ asm volatile(
+ "fldmu.64 vr0-vr31, (%0)\n"
+ : "+a"(fpregs)
+ ::"memory");
+#endif
+#else
+ {
+ unsigned long tmp3, tmp4;
+
+ asm volatile(
+ LDW_FPU_REGS(0, 4, 16, 20)
+ FMTVR_FPU_REGS(vr0, vr1)
+ LDW_FPU_REGS(32, 36, 48, 52)
+ FMTVR_FPU_REGS(vr2, vr3)
+ LDW_FPU_REGS(64, 68, 80, 84)
+ FMTVR_FPU_REGS(vr4, vr5)
+ LDW_FPU_REGS(96, 100, 112, 116)
+ FMTVR_FPU_REGS(vr6, vr7)
+ "addi %4, 128\n"
+ LDW_FPU_REGS(0, 4, 16, 20)
+ FMTVR_FPU_REGS(vr8, vr9)
+ LDW_FPU_REGS(32, 36, 48, 52)
+ FMTVR_FPU_REGS(vr10, vr11)
+ LDW_FPU_REGS(64, 68, 80, 84)
+ FMTVR_FPU_REGS(vr12, vr13)
+ LDW_FPU_REGS(96, 100, 112, 116)
+ FMTVR_FPU_REGS(vr14, vr15)
+ : "=a"(tmp1), "=a"(tmp2), "=a"(tmp3),
+ "=a"(tmp4), "+a"(fpregs)
+ ::"memory");
+ }
+#endif
+ local_irq_restore(flg);
+}
diff --git a/arch/csky/abiv2/inc/abi/fpu.h b/arch/csky/abiv2/inc/abi/fpu.h
new file mode 100644
index 0000000..22ca3cf
--- /dev/null
+++ b/arch/csky/abiv2/inc/abi/fpu.h
@@ -0,0 +1,66 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+// Copyright (C) 2018 Hangzhou C-SKY Microsystems co.,ltd.
+
+#ifndef __ASM_CSKY_FPU_H
+#define __ASM_CSKY_FPU_H
+
+#include <asm/sigcontext.h>
+#include <asm/ptrace.h>
+
+int fpu_libc_helper(struct pt_regs *regs);
+void fpu_fpe(struct pt_regs *regs);
+void __init init_fpu(void);
+
+void save_to_user_fp(struct user_fp *user_fp);
+void restore_from_user_fp(struct user_fp *user_fp);
+
+/*
+ * Define the fesr bit for fpe handle.
+ */
+#define FPE_ILLE (1 << 16) /* Illegal instruction */
+#define FPE_FEC (1 << 7) /* Input float-point arithmetic exception */
+#define FPE_IDC (1 << 5) /* Input denormalized exception */
+#define FPE_IXC (1 << 4) /* Inexact exception */
+#define FPE_UFC (1 << 3) /* Underflow exception */
+#define FPE_OFC (1 << 2) /* Overflow exception */
+#define FPE_DZC (1 << 1) /* Divide by zero exception */
+#define FPE_IOC (1 << 0) /* Invalid operation exception */
+#define FPE_REGULAR_EXCEPTION (FPE_IXC | FPE_UFC | FPE_OFC | FPE_DZC | FPE_IOC)
+
+#ifdef CONFIG_OPEN_FPU_IDE
+#define IDE_STAT (1 << 5)
+#else
+#define IDE_STAT 0
+#endif
+
+#ifdef CONFIG_OPEN_FPU_IXE
+#define IXE_STAT (1 << 4)
+#else
+#define IXE_STAT 0
+#endif
+
+#ifdef CONFIG_OPEN_FPU_UFE
+#define UFE_STAT (1 << 3)
+#else
+#define UFE_STAT 0
+#endif
+
+#ifdef CONFIG_OPEN_FPU_OFE
+#define OFE_STAT (1 << 2)
+#else
+#define OFE_STAT 0
+#endif
+
+#ifdef CONFIG_OPEN_FPU_DZE
+#define DZE_STAT (1 << 1)
+#else
+#define DZE_STAT 0
+#endif
+
+#ifdef CONFIG_OPEN_FPU_IOE
+#define IOE_STAT (1 << 0)
+#else
+#define IOE_STAT 0
+#endif
+
+#endif /* __ASM_CSKY_FPU_H */
diff --git a/arch/csky/include/asm/mmu_context.h b/arch/csky/include/asm/mmu_context.h
new file mode 100644
index 0000000..c410aa4
--- /dev/null
+++ b/arch/csky/include/asm/mmu_context.h
@@ -0,0 +1,150 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+// Copyright (C) 2018 Hangzhou C-SKY Microsystems co.,ltd.
+
+#ifndef __ASM_CSKY_MMU_CONTEXT_H
+#define __ASM_CSKY_MMU_CONTEXT_H
+
+#include <asm-generic/mm_hooks.h>
+#include <asm/setup.h>
+#include <asm/page.h>
+#include <asm/cacheflush.h>
+#include <asm/tlbflush.h>
+
+#include <linux/errno.h>
+#include <linux/sched.h>
+#include <abi/ckmmu.h>
+
+static inline void tlbmiss_handler_setup_pgd(unsigned long pgd, bool kernel)
+{
+ pgd &= ~(1<<31);
+ pgd += PHYS_OFFSET;
+ pgd |= 1;
+ setup_pgd(pgd, kernel);
+}
+
+#define TLBMISS_HANDLER_SETUP_PGD(pgd) \
+ tlbmiss_handler_setup_pgd((unsigned long)pgd, 0)
+#define TLBMISS_HANDLER_SETUP_PGD_KERNEL(pgd) \
+ tlbmiss_handler_setup_pgd((unsigned long)pgd, 1)
+
+static inline unsigned long tlb_get_pgd(void)
+{
+ return ((get_pgd()|(1<<31)) - PHYS_OFFSET) & ~1;
+}
+
+#define cpu_context(cpu, mm) ((mm)->context.asid[cpu])
+#define cpu_asid(cpu, mm) (cpu_context((cpu), (mm)) & ASID_MASK)
+#define asid_cache(cpu) (cpu_data[cpu].asid_cache)
+
+#define ASID_FIRST_VERSION (1 << CONFIG_CPU_ASID_BITS)
+#define ASID_INC 0x1
+#define ASID_MASK (ASID_FIRST_VERSION - 1)
+#define ASID_VERSION_MASK ~ASID_MASK
+
+#define destroy_context(mm) do {} while (0)
+#define enter_lazy_tlb(mm, tsk) do {} while (0)
+#define deactivate_mm(tsk, mm) do {} while (0)
+
+/*
+ * All unused by hardware upper bits will be considered
+ * as a software asid extension.
+ */
+static inline void
+get_new_mmu_context(struct mm_struct *mm, unsigned long cpu)
+{
+ unsigned long asid = asid_cache(cpu);
+
+ asid += ASID_INC;
+ if (!(asid & ASID_MASK)) {
+ flush_tlb_all(); /* start new asid cycle */
+ if (!asid) /* fix version if needed */
+ asid = ASID_FIRST_VERSION;
+ }
+ cpu_context(cpu, mm) = asid_cache(cpu) = asid;
+}
+
+/*
+ * Initialize the context related info for a new mm_struct
+ * instance.
+ */
+static inline int
+init_new_context(struct task_struct *tsk, struct mm_struct *mm)
+{
+ int i;
+
+ for_each_online_cpu(i)
+ cpu_context(i, mm) = 0;
+ return 0;
+}
+
+static inline void switch_mm(struct mm_struct *prev, struct mm_struct *next,
+ struct task_struct *tsk)
+{
+ unsigned int cpu = smp_processor_id();
+ unsigned long flags;
+
+ local_irq_save(flags);
+ /* Check if our ASID is of an older version and thus invalid */
+ if ((cpu_context(cpu, next) ^ asid_cache(cpu)) & ASID_VERSION_MASK)
+ get_new_mmu_context(next, cpu);
+ write_mmu_entryhi(cpu_asid(cpu, next));
+ TLBMISS_HANDLER_SETUP_PGD(next->pgd);
+
+ /*
+ * Mark current->active_mm as not "active" anymore.
+ * We don't want to mislead possible IPI tlb flush routines.
+ */
+ cpumask_clear_cpu(cpu, mm_cpumask(prev));
+ cpumask_set_cpu(cpu, mm_cpumask(next));
+
+ local_irq_restore(flags);
+}
+
+/*
+ * After we have set current->mm to a new value, this activates
+ * the context for the new mm so we see the new mappings.
+ */
+static inline void
+activate_mm(struct mm_struct *prev, struct mm_struct *next)
+{
+ unsigned long flags;
+ int cpu = smp_processor_id();
+
+ local_irq_save(flags);
+
+ /* Unconditionally get a new ASID. */
+ get_new_mmu_context(next, cpu);
+
+ write_mmu_entryhi(cpu_asid(cpu, next));
+ TLBMISS_HANDLER_SETUP_PGD(next->pgd);
+
+ /* mark mmu ownership change */
+ cpumask_clear_cpu(cpu, mm_cpumask(prev));
+ cpumask_set_cpu(cpu, mm_cpumask(next));
+
+ local_irq_restore(flags);
+}
+
+/*
+ * If mm is currently active_mm, we can't really drop it. Instead,
+ * we will get a new one for it.
+ */
+static inline void
+drop_mmu_context(struct mm_struct *mm, unsigned int cpu)
+{
+ unsigned long flags;
+
+ local_irq_save(flags);
+
+ if (cpumask_test_cpu(cpu, mm_cpumask(mm))) {
+ get_new_mmu_context(mm, cpu);
+ write_mmu_entryhi(cpu_asid(cpu, mm));
+ } else {
+ /* will get a new context next time */
+ cpu_context(cpu, mm) = 0;
+ }
+
+ local_irq_restore(flags);
+}
+
+#endif /* __ASM_CSKY_MMU_CONTEXT_H */
diff --git a/arch/csky/include/asm/processor.h b/arch/csky/include/asm/processor.h
new file mode 100644
index 0000000..5ad4f0b
--- /dev/null
+++ b/arch/csky/include/asm/processor.h
@@ -0,0 +1,121 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+// Copyright (C) 2018 Hangzhou C-SKY Microsystems co.,ltd.
+
+#ifndef __ASM_CSKY_PROCESSOR_H
+#define __ASM_CSKY_PROCESSOR_H
+
+/*
+ * Default implementation of macro that returns current
+ * instruction pointer ("program counter").
+ */
+#define current_text_addr() ({ __label__ _l; _l: &&_l; })
+
+#include <linux/bitops.h>
+#include <asm/segment.h>
+#include <asm/ptrace.h>
+#include <asm/current.h>
+#include <asm/cache.h>
+#include <abi/reg_ops.h>
+#include <abi/regdef.h>
+#ifdef CONFIG_CPU_HAS_FPU
+#include <abi/fpu.h>
+#endif
+
+struct cpuinfo_csky {
+ unsigned long udelay_val;
+ unsigned long asid_cache;
+ /*
+ * Capability and feature descriptor structure for CSKY CPU
+ */
+ unsigned long options;
+ unsigned int processor_id[4];
+ unsigned int fpu_id;
+} __aligned(SMP_CACHE_BYTES);
+
+extern struct cpuinfo_csky cpu_data[];
+
+/*
+ * User space process size: 2GB. This is hardcoded into a few places,
+ * so don't change it unless you know what you are doing. TASK_SIZE
+ * for a 64 bit kernel expandable to 8192EB, of which the current CSKY
+ * implementations will "only" be able to use 1TB ...
+ */
+#define TASK_SIZE 0x7fff8000UL
+
+#ifdef __KERNEL__
+#define STACK_TOP TASK_SIZE
+#define STACK_TOP_MAX STACK_TOP
+#endif
+
+/* This decides where the kernel will search for a free chunk of vm
+ * space during mmap's.
+ */
+#define TASK_UNMAPPED_BASE (TASK_SIZE / 3)
+
+struct thread_struct {
+ unsigned long ksp; /* kernel stack pointer */
+ unsigned long sr; /* saved status register */
+ unsigned long esp0; /* points to SR of stack frame */
+ unsigned long hi;
+ unsigned long lo;
+
+ /* Other stuff associated with the thread. */
+ unsigned long address; /* Last user fault */
+ unsigned long error_code;
+
+ /* FPU regs */
+ struct user_fp __aligned(16) user_fp;
+};
+
+#define INIT_THREAD { \
+ .ksp = (unsigned long) init_thread_union.stack + THREAD_SIZE, \
+ .sr = DEFAULT_PSR_VALUE, \
+}
+
+/*
+ * Do necessary setup to start up a newly executed thread.
+ *
+ * pass the data segment into user programs if it exists,
+ * it can't hurt anything as far as I can tell
+ */
+#define start_thread(_regs, _pc, _usp) \
+do { \
+ set_fs(USER_DS); /* reads from user space */ \
+ (_regs)->pc = (_pc); \
+ (_regs)->regs[1] = 0; /* ABIV1 is R7, uClibc_main rtdl arg */ \
+ (_regs)->regs[2] = 0; \
+ (_regs)->regs[3] = 0; /* ABIV2 is R7, use it? */ \
+ (_regs)->sr &= ~PS_S; \
+ (_regs)->usp = (_usp); \
+} while (0)
+
+/* Forward declaration, a strange C thing */
+struct task_struct;
+
+/* Free all resources held by a thread. */
+static inline void release_thread(struct task_struct *dead_task)
+{
+}
+
+/* Prepare to copy thread state - unlazy all lazy status */
+#define prepare_to_copy(tsk) do { } while (0)
+
+extern int kernel_thread(int (*fn)(void *), void *arg, unsigned long flags);
+
+#define copy_segments(tsk, mm) do { } while (0)
+#define release_segments(mm) do { } while (0)
+#define forget_segments() do { } while (0)
+
+extern unsigned long thread_saved_pc(struct task_struct *tsk);
+
+unsigned long get_wchan(struct task_struct *p);
+
+#define KSTK_EIP(tsk) (task_pt_regs(tsk)->pc)
+#define KSTK_ESP(tsk) (task_pt_regs(tsk)->usp)
+
+#define task_pt_regs(p) \
+ ((struct pt_regs *)(THREAD_SIZE + p->stack) - 1)
+
+#define cpu_relax() barrier()
+
+#endif /* __ASM_CSKY_PROCESSOR_H */
diff --git a/arch/csky/include/asm/switch_to.h b/arch/csky/include/asm/switch_to.h
new file mode 100644
index 0000000..35a39e8
--- /dev/null
+++ b/arch/csky/include/asm/switch_to.h
@@ -0,0 +1,36 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+// Copyright (C) 2018 Hangzhou C-SKY Microsystems co.,ltd.
+
+#ifndef __ASM_CSKY_SWITCH_TO_H
+#define __ASM_CSKY_SWITCH_TO_H
+
+#include <linux/thread_info.h>
+#ifdef CONFIG_CPU_HAS_FPU
+#include <abi/fpu.h>
+static inline void __switch_to_fpu(struct task_struct *prev,
+ struct task_struct *next)
+{
+ save_to_user_fp(&prev->thread.user_fp);
+ restore_from_user_fp(&next->thread.user_fp);
+}
+#else
+static inline void __switch_to_fpu(struct task_struct *prev,
+ struct task_struct *next)
+{}
+#endif
+
+/*
+ * Context switching is now performed out-of-line in switch_to.S
+ */
+extern struct task_struct *__switch_to(struct task_struct *,
+ struct task_struct *);
+
+#define switch_to(prev, next, last) \
+ do { \
+ struct task_struct *__prev = (prev); \
+ struct task_struct *__next = (next); \
+ __switch_to_fpu(__prev, __next); \
+ ((last) = __switch_to((prev), (next))); \
+ } while (0)
+
+#endif /* __ASM_CSKY_SWITCH_TO_H */
diff --git a/arch/csky/include/asm/thread_info.h b/arch/csky/include/asm/thread_info.h
new file mode 100644
index 0000000..a2c69a7
--- /dev/null
+++ b/arch/csky/include/asm/thread_info.h
@@ -0,0 +1,75 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+// Copyright (C) 2018 Hangzhou C-SKY Microsystems co.,ltd.
+
+#ifndef _ASM_CSKY_THREAD_INFO_H
+#define _ASM_CSKY_THREAD_INFO_H
+
+#ifndef __ASSEMBLY__
+
+#include <linux/version.h>
+#include <asm/types.h>
+#include <asm/page.h>
+#include <asm/processor.h>
+
+struct thread_info {
+ struct task_struct *task;
+ void *dump_exec_domain;
+ unsigned long flags;
+ int preempt_count;
+ unsigned long tp_value;
+ mm_segment_t addr_limit;
+ struct restart_block restart_block;
+ struct pt_regs *regs;
+ unsigned int cpu;
+};
+
+#define INIT_THREAD_INFO(tsk) \
+{ \
+ .task = &tsk, \
+ .preempt_count = INIT_PREEMPT_COUNT, \
+ .addr_limit = KERNEL_DS, \
+ .cpu = 0, \
+ .restart_block = { \
+ .fn = do_no_restart_syscall, \
+ }, \
+}
+
+#define THREAD_SIZE_ORDER (THREAD_SHIFT - PAGE_SHIFT)
+
+static inline struct thread_info *current_thread_info(void)
+{
+ unsigned long sp;
+
+ asm volatile("mov %0, sp\n":"=r"(sp));
+
+ return (struct thread_info *)(sp & ~(THREAD_SIZE - 1));
+}
+
+#endif /* !__ASSEMBLY__ */
+
+/* entry.S relies on these definitions!
+ * bits 0-5 are tested at every exception exit
+ */
+#define TIF_SIGPENDING 0 /* signal pending */
+#define TIF_NOTIFY_RESUME 1 /* callback before returning to user */
+#define TIF_NEED_RESCHED 2 /* rescheduling necessary */
+#define TIF_SYSCALL_TRACE 5 /* syscall trace active */
+#define TIF_DELAYED_TRACE 14 /* single step a syscall */
+#define TIF_POLLING_NRFLAG 16 /* poll_idle() is TIF_NEED_RESCHED */
+#define TIF_MEMDIE 18 /* is terminating due to OOM killer */
+#define TIF_FREEZE 19 /* thread is freezing for suspend */
+#define TIF_RESTORE_SIGMASK 20 /* restore signal mask in do_signal() */
+#define TIF_SECCOMP 21 /* secure computing */
+
+#define _TIF_SIGPENDING (1 << TIF_SIGPENDING)
+#define _TIF_NOTIFY_RESUME (1 << TIF_NOTIFY_RESUME)
+#define _TIF_NEED_RESCHED (1 << TIF_NEED_RESCHED)
+#define _TIF_SYSCALL_TRACE (1 << TIF_SYSCALL_TRACE)
+#define _TIF_DELAYED_TRACE (1 << TIF_DELAYED_TRACE)
+#define _TIF_POLLING_NRFLAG (1 << TIF_POLLING_NRFLAG)
+#define _TIF_MEMDIE (1 << TIF_MEMDIE)
+#define _TIF_FREEZE (1 << TIF_FREEZE)
+#define _TIF_RESTORE_SIGMASK (1 << TIF_RESTORE_SIGMASK)
+#define _TIF_SECCOMP (1 << TIF_SECCOMP)
+
+#endif /* _ASM_CSKY_THREAD_INFO_H */
diff --git a/arch/csky/include/uapi/asm/sigcontext.h b/arch/csky/include/uapi/asm/sigcontext.h
new file mode 100644
index 0000000..e81e7ff
--- /dev/null
+++ b/arch/csky/include/uapi/asm/sigcontext.h
@@ -0,0 +1,14 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+// Copyright (C) 2018 Hangzhou C-SKY Microsystems co.,ltd.
+
+#ifndef __ASM_CSKY_SIGCONTEXT_H
+#define __ASM_CSKY_SIGCONTEXT_H
+
+#include <asm/ptrace.h>
+
+struct sigcontext {
+ struct pt_regs sc_pt_regs;
+ struct user_fp sc_user_fp;
+};
+
+#endif /* __ASM_CSKY_SIGCONTEXT_H */
diff --git a/arch/csky/kernel/process.c b/arch/csky/kernel/process.c
new file mode 100644
index 0000000..8ed2002
--- /dev/null
+++ b/arch/csky/kernel/process.c
@@ -0,0 +1,136 @@
+// SPDX-License-Identifier: GPL-2.0
+// Copyright (C) 2018 Hangzhou C-SKY Microsystems co.,ltd.
+
+#include <linux/module.h>
+#include <linux/version.h>
+#include <linux/sched.h>
+#include <linux/sched/task_stack.h>
+#include <linux/sched/debug.h>
+#include <linux/delay.h>
+#include <linux/kallsyms.h>
+#include <linux/uaccess.h>
+#include <linux/ptrace.h>
+
+#include <asm/elf.h>
+#include <abi/reg_ops.h>
+
+struct cpuinfo_csky cpu_data[NR_CPUS];
+
+asmlinkage void ret_from_fork(void);
+asmlinkage void ret_from_kernel_thread(void);
+
+/*
+ * Some archs flush debug and FPU info here
+ */
+void flush_thread(void){}
+
+/*
+ * Return saved PC from a blocked thread
+ */
+unsigned long thread_saved_pc(struct task_struct *tsk)
+{
+ struct switch_stack *sw = (struct switch_stack *)tsk->thread.ksp;
+
+ return sw->r15;
+}
+
+int copy_thread(unsigned long clone_flags,
+ unsigned long usp,
+ unsigned long kthread_arg,
+ struct task_struct *p)
+{
+ struct switch_stack *childstack;
+ struct pt_regs *childregs = task_pt_regs(p);
+
+#ifdef CONFIG_CPU_HAS_FPU
+ save_to_user_fp(&p->thread.user_fp);
+#endif
+
+ childstack = ((struct switch_stack *) childregs) - 1;
+ memset(childstack, 0, sizeof(struct switch_stack));
+
+ /* setup ksp for switch_to !!! */
+ p->thread.ksp = (unsigned long)childstack;
+
+ if (unlikely(p->flags & PF_KTHREAD)) {
+ memset(childregs, 0, sizeof(struct pt_regs));
+ childstack->r15 = (unsigned long) ret_from_kernel_thread;
+ childstack->r8 = kthread_arg;
+ childstack->r9 = usp;
+ childregs->sr = mfcr("psr");
+ } else {
+ *childregs = *(current_pt_regs());
+ if (usp)
+ childregs->usp = usp;
+ if (clone_flags & CLONE_SETTLS)
+ task_thread_info(p)->tp_value = childregs->tls
+ = childregs->regs[0];
+
+ childregs->a0 = 0;
+ childstack->r15 = (unsigned long) ret_from_fork;
+ }
+
+ return 0;
+}
+
+/* Fill in the fpu structure for a core dump. */
+int dump_fpu(struct pt_regs *regs, struct user_fp *fpu)
+{
+ memcpy(fpu, ¤t->thread.user_fp, sizeof(*fpu));
+ return 1;
+}
+EXPORT_SYMBOL(dump_fpu);
+
+int dump_task_regs(struct task_struct *tsk, elf_gregset_t *pr_regs)
+{
+ struct pt_regs *regs = task_pt_regs(tsk);
+
+ /* NOTE: usp is error value. */
+ ELF_CORE_COPY_REGS((*pr_regs), regs)
+
+ return 1;
+}
+
+unsigned long get_wchan(struct task_struct *p)
+{
+ unsigned long esp, pc;
+ unsigned long stack_page;
+ int count = 0;
+
+ if (!p || p == current || p->state == TASK_RUNNING)
+ return 0;
+
+ stack_page = (unsigned long)p;
+ esp = p->thread.esp0;
+ do {
+ if (esp < stack_page+sizeof(struct task_struct) ||
+ esp >= 8184+stack_page)
+ return 0;
+ /*FIXME: There's may be error here!*/
+ pc = ((unsigned long *)esp)[1];
+ /* FIXME: This depends on the order of these functions. */
+ if (!in_sched_functions(pc))
+ return pc;
+ esp = *(unsigned long *) esp;
+ } while (count++ < 16);
+ return 0;
+}
+EXPORT_SYMBOL(get_wchan);
+
+#ifndef CONFIG_CPU_PM_NONE
+void arch_cpu_idle(void)
+{
+#ifdef CONFIG_CPU_PM_WAIT
+ asm volatile("wait\n");
+#endif
+
+#ifdef CONFIG_CPU_PM_DOZE
+ asm volatile("doze\n");
+#endif
+
+#ifdef CONFIG_CPU_PM_STOP
+ asm volatile("stop\n");
+#endif
+ local_irq_enable();
+}
+#endif
diff --git a/arch/csky/kernel/signal.c b/arch/csky/kernel/signal.c
new file mode 100644
index 0000000..b981d6e
--- /dev/null
+++ b/arch/csky/kernel/signal.c
@@ -0,0 +1,355 @@
+// SPDX-License-Identifier: GPL-2.0
+// Copyright (C) 2018 Hangzhou C-SKY Microsystems co.,ltd.
+
+#include <linux/sched.h>
+#include <linux/mm.h>
+#include <linux/kernel.h>
+#include <linux/signal.h>
+#include <linux/syscalls.h>
+#include <linux/errno.h>
+#include <linux/wait.h>
+#include <linux/ptrace.h>
+#include <linux/unistd.h>
+#include <linux/stddef.h>
+#include <linux/highuid.h>
+#include <linux/personality.h>
+#include <linux/tty.h>
+#include <linux/binfmts.h>
+#include <linux/tracehook.h>
+#include <linux/freezer.h>
+#include <linux/uaccess.h>
+
+#include <asm/setup.h>
+#include <asm/pgtable.h>
+#include <asm/traps.h>
+#include <asm/ucontext.h>
+#include <asm/vdso.h>
+
+#include <abi/regdef.h>
+
+#ifdef CONFIG_CPU_HAS_FPU
+#include <abi/fpu.h>
+
+static int restore_fpu_state(struct sigcontext *sc)
+{
+ int err = 0;
+ struct user_fp user_fp;
+
+ err = copy_from_user(&user_fp, &sc->sc_user_fp, sizeof(user_fp));
+
+ restore_from_user_fp(&user_fp);
+
+ return err;
+}
+
+static int save_fpu_state(struct sigcontext *sc)
+{
+ struct user_fp user_fp;
+
+ save_to_user_fp(&user_fp);
+
+ return copy_to_user(&sc->sc_user_fp, &user_fp, sizeof(user_fp));
+}
+#else
+static inline int restore_fpu_state(struct sigcontext *sc) { return 0; }
+static inline int save_fpu_state(struct sigcontext *sc) { return 0; }
+#endif
+
+struct rt_sigframe {
+ int sig;
+ struct siginfo *pinfo;
+ void *puc;
+ struct siginfo info;
+ struct ucontext uc;
+};
+
+static int
+restore_sigframe(struct pt_regs *regs,
+ struct sigcontext *sc, int *pr2)
+{
+ int err = 0;
+
+ /* Always make any pending restarted system calls return -EINTR */
+ current_thread_info()->task->restart_block.fn = do_no_restart_syscall;
+
+ err |= copy_from_user(regs, &sc->sc_pt_regs, sizeof(struct pt_regs));
+
+ err |= restore_fpu_state(sc);
+
+ *pr2 = regs->a0;
+ return err;
+}
+
+asmlinkage int
+do_rt_sigreturn(void)
+{
+ sigset_t set;
+ int a0;
+ struct pt_regs *regs = current_pt_regs();
+ struct rt_sigframe *frame = (struct rt_sigframe *)(regs->usp);
+
+ if (verify_area(VERIFY_READ, frame, sizeof(*frame)))
+ goto badframe;
+ if (__copy_from_user(&set, &frame->uc.uc_sigmask, sizeof(set)))
+ goto badframe;
+
+ sigdelsetmask(&set, (sigmask(SIGKILL) | sigmask(SIGSTOP)));
+ spin_lock_irq(¤t->sighand->siglock);
+ current->blocked = set;
+ recalc_sigpending();
+ spin_unlock_irq(¤t->sighand->siglock);
+
+ if (restore_sigframe(regs, &frame->uc.uc_mcontext, &a0))
+ goto badframe;
+
+ return a0;
+
+badframe:
+ force_sig(SIGSEGV, current);
+ return 0;
+}
+
+static int setup_sigframe(struct sigcontext *sc, struct pt_regs *regs)
+{
+ int err = 0;
+
+ err |= copy_to_user(&sc->sc_pt_regs, regs, sizeof(struct pt_regs));
+ err |= save_fpu_state(sc);
+
+ return err;
+}
+
+static inline void *
+get_sigframe(struct k_sigaction *ka, struct pt_regs *regs, size_t frame_size)
+{
+ unsigned long usp;
+
+ /* Default to using normal stack. */
+ usp = regs->usp;
+
+ /* This is the X/Open sanctioned signal stack switching. */
+ if ((ka->sa.sa_flags & SA_ONSTACK) && !sas_ss_flags(usp)) {
+ if (!on_sig_stack(usp))
+ usp = current->sas_ss_sp + current->sas_ss_size;
+ }
+ return (void *)((usp - frame_size) & -8UL);
+}
+
+static int setup_rt_frame(int sig, struct k_sigaction *ka, siginfo_t *info,
+ sigset_t *set, struct pt_regs *regs)
+{
+ struct rt_sigframe *frame;
+ int err = 0;
+
+ struct csky_vdso *vdso = current->mm->context.vdso;
+
+ frame = get_sigframe(ka, regs, sizeof(*frame));
+ if (!frame)
+ return 1;
+
+ err |= __put_user(sig, &frame->sig);
+ err |= __put_user(&frame->info, &frame->pinfo);
+ err |= __put_user(&frame->uc, &frame->puc);
+ err |= copy_siginfo_to_user(&frame->info, info);
+
+ /* Create the ucontext. */
+ err |= __put_user(0, &frame->uc.uc_flags);
+ err |= __put_user(0, &frame->uc.uc_link);
+ err |= __put_user((void *)current->sas_ss_sp,
+ &frame->uc.uc_stack.ss_sp);
+ err |= __put_user(sas_ss_flags(regs->usp),
+ &frame->uc.uc_stack.ss_flags);
+ err |= __put_user(current->sas_ss_size, &frame->uc.uc_stack.ss_size);
+ err |= setup_sigframe(&frame->uc.uc_mcontext, regs);
+ err |= copy_to_user(&frame->uc.uc_sigmask, set, sizeof(*set));
+
+ if (err)
+ goto give_sigsegv;
+
+ /* Set up registers for signal handler */
+ regs->usp = (unsigned long) frame;
+ regs->pc = (unsigned long) ka->sa.sa_handler;
+ regs->lr = (unsigned long)vdso->rt_signal_retcode;
+
+adjust_stack:
+ regs->a0 = sig; /* first arg is signo */
+ regs->a1 = (unsigned long)(&(frame->info));
+ regs->a2 = (unsigned long)(&(frame->uc));
+ return err;
+
+give_sigsegv:
+ if (sig == SIGSEGV)
+ ka->sa.sa_handler = SIG_DFL;
+ force_sig(SIGSEGV, current);
+ goto adjust_stack;
+}
+
+/*
+ * OK, we're invoking a handler
+ */
+static int
+handle_signal(int sig, struct k_sigaction *ka, siginfo_t *info,
+ sigset_t *oldset, struct pt_regs *regs)
+{
+ struct task_struct *tsk = current;
+ int ret;
+
+ /*
+ * set up the stack frame, regardless of SA_SIGINFO,
+ * and pass info anyway.
+ */
+ ret = setup_rt_frame(sig, ka, info, oldset, regs);
+
+ if (ret != 0) {
+ force_sigsegv(sig, tsk);
+ return ret;
+ }
+
+ /* Block the signal if we were successful. */
+ spin_lock_irq(¤t->sighand->siglock);
+ sigorsets(¤t->blocked, ¤t->blocked, &ka->sa.sa_mask);
+ if (!(ka->sa.sa_flags & SA_NODEFER))
+ sigaddset(¤t->blocked, sig);
+ recalc_sigpending();
+ spin_unlock_irq(¤t->sighand->siglock);
+
+ return 0;
+}
+
+/*
+ * Note that 'init' is a special process: it doesn't get signals it doesn't
+ * want to handle. Thus you cannot kill init even with a SIGKILL even by
+ * mistake.
+ *
+ * Note that we go through the signals twice: once to check the signals
+ * that the kernel can handle, and then we build all the user-level signal
+ * handling stack-frames in one go after that.
+ */
+static void do_signal(struct pt_regs *regs, int syscall)
+{
+ unsigned int retval = 0, continue_addr = 0, restart_addr = 0;
+ struct ksignal ksig;
+
+ /*
+ * We want the common case to go fast, which
+ * is why we may in certain cases get here from
+ * kernel mode. Just return without doing anything
+ * if so.
+ */
+ if (!user_mode(regs))
+ return;
+
+ current->thread.esp0 = (unsigned long)regs;
+
+ /*
+ * If we were from a system call, check for system call restarting...
+ */
+ if (syscall) {
+ continue_addr = regs->pc;
+#if defined(__CSKYABIV2__)
+ restart_addr = continue_addr - 4;
+#else
+ restart_addr = continue_addr - 2;
+#endif
+ retval = regs->a0;
+
+ /*
+ * Prepare for system call restart. We do this here so that a
+ * debugger will see the already changed.
+ */
+ switch (retval) {
+ case -ERESTARTNOHAND:
+ case -ERESTARTSYS:
+ case -ERESTARTNOINTR:
+ regs->a0 = regs->orig_a0;
+ regs->pc = restart_addr;
+ break;
+ case -ERESTART_RESTARTBLOCK:
+ regs->a0 = -EINTR;
+ break;
+ }
+ }
+
+ if (try_to_freeze())
+ goto no_signal;
+
+ /*
+ * Get the signal to deliver. When running under ptrace, at this
+ * point the debugger may change all our registers ...
+ */
+ if (get_signal(&ksig)) {
+ sigset_t *oldset;
+
+ /*
+ * Depending on the signal settings we may need to revert the
+ * decision to restart the system call. But skip this if a
+ * debugger has chosen to restart at a different PC.
+ */
+ if (regs->pc == restart_addr) {
+ if (retval == -ERESTARTNOHAND ||
+ (retval == -ERESTARTSYS &&
+ !(ksig.ka.sa.sa_flags & SA_RESTART))) {
+ regs->a0 = -EINTR;
+ regs->pc = continue_addr;
+ }
+ }
+
+ if (test_thread_flag(TIF_RESTORE_SIGMASK))
+ oldset = ¤t->saved_sigmask;
+ else
+ oldset = ¤t->blocked;
+ /* Whee! Actually deliver the signal. */
+ if (handle_signal(ksig.sig, &ksig.ka,
+ &ksig.info, oldset, regs) == 0) {
+ /*
+ * A signal was successfully delivered; the saved
+ * sigmask will have been stored in the signal frame,
+ * and will be restored by sigreturn, so we can simply
+ * clear the TIF_RESTORE_SIGMASK flag.
+ */
+ if (test_thread_flag(TIF_RESTORE_SIGMASK))
+ clear_thread_flag(TIF_RESTORE_SIGMASK);
+ }
+ return;
+ }
+
+no_signal:
+ if (syscall) {
+ /*
+ * Handle restarting a different system call. As above,
+ * if a debugger has chosen to restart at a different PC,
+ * ignore the restart.
+ */
+ if (retval == -ERESTART_RESTARTBLOCK
+ && regs->pc == continue_addr) {
+#if defined(__CSKYABIV2__)
+ regs->regs[3] = __NR_restart_syscall;
+ regs->pc -= 4;
+#else
+ regs->regs[9] = __NR_restart_syscall;
+ regs->pc -= 2;
+#endif
+ }
+
+ /*
+ * If there's no signal to deliver, we just put the saved
+ * sigmask back.
+ */
+ if (test_thread_flag(TIF_RESTORE_SIGMASK)) {
+ clear_thread_flag(TIF_RESTORE_SIGMASK);
+ sigprocmask(SIG_SETMASK, ¤t->saved_sigmask, NULL);
+ }
+ }
+}
+
+asmlinkage void
+do_notify_resume(unsigned int thread_flags, struct pt_regs *regs, int syscall)
+{
+ if (thread_flags & _TIF_SIGPENDING)
+ do_signal(regs, syscall);
+
+ if (thread_flags & _TIF_NOTIFY_RESUME) {
+ clear_thread_flag(TIF_NOTIFY_RESUME);
+ tracehook_notify_resume(regs);
+ }
+}
diff --git a/arch/csky/kernel/time.c b/arch/csky/kernel/time.c
new file mode 100644
index 0000000..b5fc944
--- /dev/null
+++ b/arch/csky/kernel/time.c
@@ -0,0 +1,11 @@
+// SPDX-License-Identifier: GPL-2.0
+// Copyright (C) 2018 Hangzhou C-SKY Microsystems co.,ltd.
+
+#include <linux/clk-provider.h>
+#include <linux/clocksource.h>
+
+void __init time_init(void)
+{
+ of_clk_init(NULL);
+ timer_probe();
+}
--
2.7.4
^ permalink raw reply related [flat|nested] 68+ messages in thread* Re: [PATCH V9 08/21] csky: Process management and Signal
2018-10-16 2:58 ` [PATCH V9 08/21] csky: Process management and Signal Guo Ren
@ 2018-10-17 15:11 ` Arnd Bergmann
2018-10-18 2:37 ` Guo Ren
0 siblings, 1 reply; 68+ messages in thread
From: Arnd Bergmann @ 2018-10-17 15:11 UTC (permalink / raw)
To: Guo Ren
Cc: Andrew Morton, Daniel Lezcano, David Miller, gregkh,
Christoph Hellwig, Marc Zyngier, Mark Rutland, Peter Zijlstra,
Rob Herring, Thomas Gleixner, Linux Kernel Mailing List,
linux-arch, DTML, Rob Herring, c-sky_gcc_upstream
On Tue, Oct 16, 2018 at 5:01 AM Guo Ren <ren_guo@c-sky.com> wrote:
>
> This patch adds files related to task_switch, sigcontext, signal,
> fpu context switch.
>
> Signed-off-by: Guo Ren <ren_guo@c-sky.com>
> Cc: Arnd Bergmann <arnd@arndb.de>
> ---
> Changelog:
> - abiv2/fpu.c: Userspace should never be sent NSIGXXX as a si_code.
> Use FPE_FLTUNK instead.
> - abiv2/fpu.c: Use force_sig_fault instead.
I'm hesitant to provide a Reviewed-by: tag for this one, as I don't
understand the signal handling enough myself, and it's something
that should be reviewed properly because of the ABI implications.
Maybe someone else can have a closer look at this?
Arnd
^ permalink raw reply [flat|nested] 68+ messages in thread* Re: [PATCH V9 08/21] csky: Process management and Signal
2018-10-17 15:11 ` Arnd Bergmann
@ 2018-10-18 2:37 ` Guo Ren
0 siblings, 0 replies; 68+ messages in thread
From: Guo Ren @ 2018-10-18 2:37 UTC (permalink / raw)
To: Arnd Bergmann
Cc: Andrew Morton, Daniel Lezcano, David Miller, gregkh,
Christoph Hellwig, Marc Zyngier, Mark Rutland, Peter Zijlstra,
Rob Herring, Thomas Gleixner, Linux Kernel Mailing List,
linux-arch, DTML, Rob Herring, c-sky_gcc_upstream
On Wed, Oct 17, 2018 at 05:11:00PM +0200, Arnd Bergmann wrote:
> On Tue, Oct 16, 2018 at 5:01 AM Guo Ren <ren_guo@c-sky.com> wrote:
> >
> > This patch adds files related to task_switch, sigcontext, signal,
> > fpu context switch.
> >
> > Signed-off-by: Guo Ren <ren_guo@c-sky.com>
> > Cc: Arnd Bergmann <arnd@arndb.de>
> > ---
> > Changelog:
> > - abiv2/fpu.c: Userspace should never be sent NSIGXXX as a si_code.
> > Use FPE_FLTUNK instead.
> > - abiv2/fpu.c: Use force_sig_fault instead.
>
> I'm hesitant to provide a Reviewed-by: tag for this one, as I don't
> understand the signal handling enough myself, and it's something
> that should be reviewed properly because of the ABI implications.
>
> Maybe someone else can have a closer look at this?
NSIGXXX and force_sig_fault modification is from Eric W. Biederman
in PATCH V6 and I forget add:
Cc: "Eric W. Biederman" <ebiederm@xmission.com>
Best Regards
Guo Ren
^ permalink raw reply [flat|nested] 68+ messages in thread
* [PATCH V9 09/21] csky: VDSO and rt_sigreturn
2018-10-16 2:58 [PATCH V9 00/21] C-SKY(csky) Linux Kernel Port Guo Ren
` (7 preceding siblings ...)
2018-10-16 2:58 ` [PATCH V9 08/21] csky: Process management and Signal Guo Ren
@ 2018-10-16 2:58 ` Guo Ren
2018-10-17 15:13 ` Arnd Bergmann
2018-10-16 2:58 ` [PATCH V9 10/21] csky: IRQ handling Guo Ren
` (12 subsequent siblings)
21 siblings, 1 reply; 68+ messages in thread
From: Guo Ren @ 2018-10-16 2:58 UTC (permalink / raw)
To: akpm, arnd, daniel.lezcano, davem, gregkh, hch, marc.zyngier,
mark.rutland, peterz, robh, tglx
Cc: linux-kernel, linux-arch, devicetree, robh+dt,
c-sky_gcc_upstream, Guo Ren
This patch adds files related to VDSO and our VDSO only support
rt_sigreturn.
Signed-off-by: Guo Ren <ren_guo@c-sky.com>
Cc: Arnd Bergmann <arnd@arndb.de>
---
arch/csky/abiv1/inc/abi/vdso.h | 17 +++++++++
arch/csky/abiv2/inc/abi/vdso.h | 23 +++++++++++
arch/csky/include/asm/vdso.h | 12 ++++++
arch/csky/kernel/vdso.c | 86 ++++++++++++++++++++++++++++++++++++++++++
4 files changed, 138 insertions(+)
create mode 100644 arch/csky/abiv1/inc/abi/vdso.h
create mode 100644 arch/csky/abiv2/inc/abi/vdso.h
create mode 100644 arch/csky/include/asm/vdso.h
create mode 100644 arch/csky/kernel/vdso.c
diff --git a/arch/csky/abiv1/inc/abi/vdso.h b/arch/csky/abiv1/inc/abi/vdso.h
new file mode 100644
index 0000000..14352f5
--- /dev/null
+++ b/arch/csky/abiv1/inc/abi/vdso.h
@@ -0,0 +1,17 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+
+#include <linux/uaccess.h>
+
+static inline int setup_vdso_page(unsigned short *ptr)
+{
+ int err = 0;
+
+ /* movi r1, 127 */
+ err |= __put_user(0x67f1, ptr + 0);
+ /* addi r1, (139 - 127) */
+ err |= __put_user(0x20b1, ptr + 1);
+ /* trap 0 */
+ err |= __put_user(0x0008, ptr + 2);
+
+ return err;
+}
diff --git a/arch/csky/abiv2/inc/abi/vdso.h b/arch/csky/abiv2/inc/abi/vdso.h
new file mode 100644
index 0000000..b60d4a0
--- /dev/null
+++ b/arch/csky/abiv2/inc/abi/vdso.h
@@ -0,0 +1,23 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+
+#ifndef __ABI_CSKY_VDSO_H
+#define __ABI_CSKY_VDSO_H
+
+#include <linux/uaccess.h>
+
+static inline int setup_vdso_page(unsigned short *ptr)
+{
+ int err = 0;
+
+ /* movi r7, 173 */
+ err |= __put_user(0xea07, ptr);
+ err |= __put_user(0x008b, ptr+1);
+
+ /* trap 0 */
+ err |= __put_user(0xc000, ptr+2);
+ err |= __put_user(0x2020, ptr+3);
+
+ return err;
+}
+
+#endif /* __ABI_CSKY_STRING_H */
diff --git a/arch/csky/include/asm/vdso.h b/arch/csky/include/asm/vdso.h
new file mode 100644
index 0000000..d963d69
--- /dev/null
+++ b/arch/csky/include/asm/vdso.h
@@ -0,0 +1,12 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+
+#ifndef __ASM_CSKY_VDSO_H
+#define __ASM_CSKY_VDSO_H
+
+#include <abi/vdso.h>
+
+struct csky_vdso {
+ unsigned short rt_signal_retcode[4];
+};
+
+#endif /* __ASM_CSKY_VDSO_H */
diff --git a/arch/csky/kernel/vdso.c b/arch/csky/kernel/vdso.c
new file mode 100644
index 0000000..60ff7ad
--- /dev/null
+++ b/arch/csky/kernel/vdso.c
@@ -0,0 +1,86 @@
+// SPDX-License-Identifier: GPL-2.0
+// Copyright (C) 2018 Hangzhou C-SKY Microsystems co.,ltd.
+
+#include <linux/kernel.h>
+#include <linux/err.h>
+#include <linux/sched.h>
+#include <linux/mm.h>
+#include <linux/init.h>
+#include <linux/binfmts.h>
+#include <linux/elf.h>
+#include <linux/vmalloc.h>
+#include <linux/unistd.h>
+#include <linux/uaccess.h>
+
+#include <asm/vdso.h>
+#include <asm/cacheflush.h>
+
+static struct page *vdso_page;
+
+static int __init init_vdso(void)
+{
+ struct csky_vdso *vdso;
+ int err = 0;
+
+ vdso_page = alloc_page(GFP_KERNEL);
+ if (!vdso_page)
+ panic("Cannot allocate vdso");
+
+ vdso = vmap(&vdso_page, 1, 0, PAGE_KERNEL);
+ if (!vdso)
+ panic("Cannot map vdso");
+
+ clear_page(vdso);
+
+ err = setup_vdso_page(vdso->rt_signal_retcode);
+ if (err)
+ panic("Cannot set signal return code, err: %x.", err);
+
+ dcache_wb_range((unsigned long)vdso, (unsigned long)vdso + 16);
+
+ vunmap(vdso);
+
+ return 0;
+}
+subsys_initcall(init_vdso);
+
+int arch_setup_additional_pages(struct linux_binprm *bprm, int uses_interp)
+{
+ int ret;
+ unsigned long addr;
+ struct mm_struct *mm = current->mm;
+
+ down_write(&mm->mmap_sem);
+
+ addr = get_unmapped_area(NULL, STACK_TOP, PAGE_SIZE, 0, 0);
+ if (IS_ERR_VALUE(addr)) {
+ ret = addr;
+ goto up_fail;
+ }
+
+ ret = install_special_mapping(
+ mm,
+ addr,
+ PAGE_SIZE,
+ VM_READ|VM_EXEC|VM_MAYREAD|VM_MAYWRITE|VM_MAYEXEC,
+ &vdso_page);
+ if (ret)
+ goto up_fail;
+
+ mm->context.vdso = (void *)addr;
+
+up_fail:
+ up_write(&mm->mmap_sem);
+ return ret;
+}
+
+const char *arch_vma_name(struct vm_area_struct *vma)
+{
+ if (vma->vm_mm == NULL)
+ return NULL;
+
+ if (vma->vm_start == (long)vma->vm_mm->context.vdso)
+ return "[vdso]";
+ else
+ return NULL;
+}
--
2.7.4
^ permalink raw reply related [flat|nested] 68+ messages in thread* Re: [PATCH V9 09/21] csky: VDSO and rt_sigreturn
2018-10-16 2:58 ` [PATCH V9 09/21] csky: VDSO and rt_sigreturn Guo Ren
@ 2018-10-17 15:13 ` Arnd Bergmann
0 siblings, 0 replies; 68+ messages in thread
From: Arnd Bergmann @ 2018-10-17 15:13 UTC (permalink / raw)
To: Guo Ren
Cc: Andrew Morton, Daniel Lezcano, David Miller, gregkh,
Christoph Hellwig, Marc Zyngier, Mark Rutland, Peter Zijlstra,
Rob Herring, Thomas Gleixner, Linux Kernel Mailing List,
linux-arch, DTML, Rob Herring, c-sky_gcc_upstream
On Tue, Oct 16, 2018 at 5:01 AM Guo Ren <ren_guo@c-sky.com> wrote:
>
> This patch adds files related to VDSO and our VDSO only support
> rt_sigreturn.
>
> Signed-off-by: Guo Ren <ren_guo@c-sky.com>
> Cc: Arnd Bergmann <arnd@arndb.de>
I was a little worried about the VDSO because of the discussion we had about
64-bit time_t. Since there is no gettimeofday call in here, that's fine, so
Reviewed-by: Arnd Bergmann <arnd@arndb.de>
^ permalink raw reply [flat|nested] 68+ messages in thread
* [PATCH V9 10/21] csky: IRQ handling
2018-10-16 2:58 [PATCH V9 00/21] C-SKY(csky) Linux Kernel Port Guo Ren
` (8 preceding siblings ...)
2018-10-16 2:58 ` [PATCH V9 09/21] csky: VDSO and rt_sigreturn Guo Ren
@ 2018-10-16 2:58 ` Guo Ren
2018-10-17 15:14 ` Arnd Bergmann
2018-10-16 2:58 ` [PATCH V9 11/21] csky: Atomic operations Guo Ren
` (11 subsequent siblings)
21 siblings, 1 reply; 68+ messages in thread
From: Guo Ren @ 2018-10-16 2:58 UTC (permalink / raw)
To: akpm, arnd, daniel.lezcano, davem, gregkh, hch, marc.zyngier,
mark.rutland, peterz, robh, tglx
Cc: linux-kernel, linux-arch, devicetree, robh+dt,
c-sky_gcc_upstream, Guo Ren
This patch adds IRQ handling files.
Signed-off-by: Guo Ren <ren_guo@c-sky.com>
Cc: Arnd Bergmann <arnd@arndb.de>
Cc: Thomas Gleixner <tglx@linutronix.de>
---
Changelog:
- Use CONFIG_GENERIC_IRQ_MULTI_HANDLER
---
---
arch/csky/include/asm/irqflags.h | 49 ++++++++++++++++++++++++++++++++++++++++
arch/csky/kernel/irq.c | 22 ++++++++++++++++++
2 files changed, 71 insertions(+)
create mode 100644 arch/csky/include/asm/irqflags.h
create mode 100644 arch/csky/kernel/irq.c
diff --git a/arch/csky/include/asm/irqflags.h b/arch/csky/include/asm/irqflags.h
new file mode 100644
index 0000000..9e3a569
--- /dev/null
+++ b/arch/csky/include/asm/irqflags.h
@@ -0,0 +1,49 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+
+#ifndef __ASM_CSKY_IRQFLAGS_H
+#define __ASM_CSKY_IRQFLAGS_H
+#include <abi/reg_ops.h>
+
+static inline unsigned long arch_local_irq_save(void)
+{
+ unsigned long flags;
+
+ flags = mfcr("psr");
+ asm volatile("psrclr ie\n":::"memory");
+ return flags;
+}
+#define arch_local_irq_save arch_local_irq_save
+
+static inline void arch_local_irq_enable(void)
+{
+ asm volatile("psrset ee, ie\n":::"memory");
+}
+#define arch_local_irq_enable arch_local_irq_enable
+
+static inline void arch_local_irq_disable(void)
+{
+ asm volatile("psrclr ie\n":::"memory");
+}
+#define arch_local_irq_disable arch_local_irq_disable
+
+static inline unsigned long arch_local_save_flags(void)
+{
+ return mfcr("psr");
+}
+#define arch_local_save_flags arch_local_save_flags
+
+static inline void arch_local_irq_restore(unsigned long flags)
+{
+ mtcr("psr", flags);
+}
+#define arch_local_irq_restore arch_local_irq_restore
+
+static inline int arch_irqs_disabled_flags(unsigned long flags)
+{
+ return !(flags & (1<<6));
+}
+#define arch_irqs_disabled_flags arch_irqs_disabled_flags
+
+#include <asm-generic/irqflags.h>
+
+#endif /* __ASM_CSKY_IRQFLAGS_H */
diff --git a/arch/csky/kernel/irq.c b/arch/csky/kernel/irq.c
new file mode 100644
index 0000000..03a1930
--- /dev/null
+++ b/arch/csky/kernel/irq.c
@@ -0,0 +1,22 @@
+// SPDX-License-Identifier: GPL-2.0
+// Copyright (C) 2018 Hangzhou C-SKY Microsystems co.,ltd.
+
+#include <linux/init.h>
+#include <linux/interrupt.h>
+#include <linux/irq.h>
+#include <linux/irqchip.h>
+#include <asm/traps.h>
+#include <asm/smp.h>
+
+void __init init_IRQ(void)
+{
+ irqchip_init();
+#ifdef CONFIG_SMP
+ setup_smp_ipi();
+#endif
+}
+
+asmlinkage void __irq_entry csky_do_IRQ(struct pt_regs *regs)
+{
+ handle_arch_irq(regs);
+}
--
2.7.4
^ permalink raw reply related [flat|nested] 68+ messages in thread* Re: [PATCH V9 10/21] csky: IRQ handling
2018-10-16 2:58 ` [PATCH V9 10/21] csky: IRQ handling Guo Ren
@ 2018-10-17 15:14 ` Arnd Bergmann
2018-10-18 2:39 ` Guo Ren
0 siblings, 1 reply; 68+ messages in thread
From: Arnd Bergmann @ 2018-10-17 15:14 UTC (permalink / raw)
To: Guo Ren
Cc: Andrew Morton, Daniel Lezcano, David Miller, gregkh,
Christoph Hellwig, Marc Zyngier, Mark Rutland, Peter Zijlstra,
Rob Herring, Thomas Gleixner, Linux Kernel Mailing List,
linux-arch, DTML, Rob Herring, c-sky_gcc_upstream
On Tue, Oct 16, 2018 at 5:02 AM Guo Ren <ren_guo@c-sky.com> wrote:
>
> This patch adds IRQ handling files.
>
> Signed-off-by: Guo Ren <ren_guo@c-sky.com>
> Cc: Arnd Bergmann <arnd@arndb.de>
> Cc: Thomas Gleixner <tglx@linutronix.de>
> ---
> Changelog:
> - Use CONFIG_GENERIC_IRQ_MULTI_HANDLER
Reviewed-by: Arnd Bergmann <arnd@arndb.de>
Using CONFIG_GENERIC_IRQ_MULTI_HANDLER makes this really
short, nice!
^ permalink raw reply [flat|nested] 68+ messages in thread
* Re: [PATCH V9 10/21] csky: IRQ handling
2018-10-17 15:14 ` Arnd Bergmann
@ 2018-10-18 2:39 ` Guo Ren
0 siblings, 0 replies; 68+ messages in thread
From: Guo Ren @ 2018-10-18 2:39 UTC (permalink / raw)
To: Arnd Bergmann
Cc: Andrew Morton, Daniel Lezcano, David Miller, gregkh,
Christoph Hellwig, Marc Zyngier, Mark Rutland, Peter Zijlstra,
Rob Herring, Thomas Gleixner, Linux Kernel Mailing List,
linux-arch, DTML, Rob Herring, c-sky_gcc_upstream
On Wed, Oct 17, 2018 at 05:14:48PM +0200, Arnd Bergmann wrote:
> On Tue, Oct 16, 2018 at 5:02 AM Guo Ren <ren_guo@c-sky.com> wrote:
> >
> > This patch adds IRQ handling files.
> >
> > Signed-off-by: Guo Ren <ren_guo@c-sky.com>
> > Cc: Arnd Bergmann <arnd@arndb.de>
> > Cc: Thomas Gleixner <tglx@linutronix.de>
> > ---
> > Changelog:
> > - Use CONFIG_GENERIC_IRQ_MULTI_HANDLER
>
> Reviewed-by: Arnd Bergmann <arnd@arndb.de>
>
> Using CONFIG_GENERIC_IRQ_MULTI_HANDLER makes this really
> short, nice!
Yes, this is tglx's suggestion.
Best Regards
Guo Ren
^ permalink raw reply [flat|nested] 68+ messages in thread
* [PATCH V9 11/21] csky: Atomic operations
2018-10-16 2:58 [PATCH V9 00/21] C-SKY(csky) Linux Kernel Port Guo Ren
` (9 preceding siblings ...)
2018-10-16 2:58 ` [PATCH V9 10/21] csky: IRQ handling Guo Ren
@ 2018-10-16 2:58 ` Guo Ren
2018-10-17 15:17 ` Arnd Bergmann
2018-10-21 20:55 ` Peter Zijlstra
2018-10-16 2:58 ` [PATCH V9 12/21] csky: ELF and module probe Guo Ren
` (10 subsequent siblings)
21 siblings, 2 replies; 68+ messages in thread
From: Guo Ren @ 2018-10-16 2:58 UTC (permalink / raw)
To: akpm, arnd, daniel.lezcano, davem, gregkh, hch, marc.zyngier,
mark.rutland, peterz, robh, tglx
Cc: linux-kernel, linux-arch, devicetree, robh+dt,
c-sky_gcc_upstream, Guo Ren, Andrea Parri
This patch adds atomic, cmpxchg, spinlock files.
Signed-off-by: Guo Ren <ren_guo@c-sky.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Andrea Parri <andrea.parri@amarulasolutions.com>
Cc: Arnd Bergmann <arnd@arndb.de>
---
Changlog:
- SMP supported
- ticklock supported
- queue-rwlock supported
---
---
arch/csky/include/asm/atomic.h | 212 +++++++++++++++++++++++++
arch/csky/include/asm/cmpxchg.h | 73 +++++++++
arch/csky/include/asm/spinlock.h | 274 +++++++++++++++++++++++++++++++++
arch/csky/include/asm/spinlock_types.h | 37 +++++
arch/csky/kernel/atomic.S | 87 +++++++++++
5 files changed, 683 insertions(+)
create mode 100644 arch/csky/include/asm/atomic.h
create mode 100644 arch/csky/include/asm/cmpxchg.h
create mode 100644 arch/csky/include/asm/spinlock.h
create mode 100644 arch/csky/include/asm/spinlock_types.h
create mode 100644 arch/csky/kernel/atomic.S
diff --git a/arch/csky/include/asm/atomic.h b/arch/csky/include/asm/atomic.h
new file mode 100644
index 0000000..e369d73
--- /dev/null
+++ b/arch/csky/include/asm/atomic.h
@@ -0,0 +1,212 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+
+#ifndef __ASM_CSKY_ATOMIC_H
+#define __ASM_CSKY_ATOMIC_H
+
+#include <linux/version.h>
+#include <asm/cmpxchg.h>
+#include <asm/barrier.h>
+
+#ifdef CONFIG_CPU_HAS_LDSTEX
+
+#define __atomic_add_unless __atomic_add_unless
+static inline int __atomic_add_unless(atomic_t *v, int a, int u)
+{
+ unsigned long tmp, ret;
+
+ smp_mb();
+
+ asm volatile (
+ "1: ldex.w %0, (%3) \n"
+ " mov %1, %0 \n"
+ " cmpne %0, %4 \n"
+ " bf 2f \n"
+ " add %0, %2 \n"
+ " stex.w %0, (%3) \n"
+ " bez %0, 1b \n"
+ "2: \n"
+ : "=&r" (tmp), "=&r" (ret)
+ : "r" (a), "r"(&v->counter), "r"(u)
+ : "memory");
+
+ if (ret != u)
+ smp_mb();
+
+ return ret;
+}
+
+#define ATOMIC_OP(op, c_op) \
+static inline void atomic_##op(int i, atomic_t *v) \
+{ \
+ unsigned long tmp; \
+ \
+ asm volatile ( \
+ "1: ldex.w %0, (%2) \n" \
+ " " #op " %0, %1 \n" \
+ " stex.w %0, (%2) \n" \
+ " bez %0, 1b \n" \
+ : "=&r" (tmp) \
+ : "r" (i), "r"(&v->counter) \
+ : "memory"); \
+}
+
+#define ATOMIC_OP_RETURN(op, c_op) \
+static inline int atomic_##op##_return(int i, atomic_t *v) \
+{ \
+ unsigned long tmp, ret; \
+ \
+ smp_mb(); \
+ asm volatile ( \
+ "1: ldex.w %0, (%3) \n" \
+ " " #op " %0, %2 \n" \
+ " mov %1, %0 \n" \
+ " stex.w %0, (%3) \n" \
+ " bez %0, 1b \n" \
+ : "=&r" (tmp), "=&r" (ret) \
+ : "r" (i), "r"(&v->counter) \
+ : "memory"); \
+ smp_mb(); \
+ \
+ return ret; \
+}
+
+#define ATOMIC_FETCH_OP(op, c_op) \
+static inline int atomic_fetch_##op(int i, atomic_t *v) \
+{ \
+ unsigned long tmp, ret; \
+ \
+ smp_mb(); \
+ asm volatile ( \
+ "1: ldex.w %0, (%3) \n" \
+ " mov %1, %0 \n" \
+ " " #op " %0, %2 \n" \
+ " stex.w %0, (%3) \n" \
+ " bez %0, 1b \n" \
+ : "=&r" (tmp), "=&r" (ret) \
+ : "r" (i), "r"(&v->counter) \
+ : "memory"); \
+ smp_mb(); \
+ \
+ return ret; \
+}
+
+#else /* CONFIG_CPU_HAS_LDSTEX */
+
+#include <linux/irqflags.h>
+
+#define __atomic_add_unless __atomic_add_unless
+static inline int __atomic_add_unless(atomic_t *v, int a, int u)
+{
+ unsigned long tmp, ret, flags;
+
+ raw_local_irq_save(flags);
+
+ asm volatile (
+ " ldw %0, (%3) \n"
+ " mov %1, %0 \n"
+ " cmpne %0, %4 \n"
+ " bf 2f \n"
+ " add %0, %2 \n"
+ " stw %0, (%3) \n"
+ "2: \n"
+ : "=&r" (tmp), "=&r" (ret)
+ : "r" (a), "r"(&v->counter), "r"(u)
+ : "memory");
+
+ raw_local_irq_restore(flags);
+
+ return ret;
+}
+
+#define ATOMIC_OP(op, c_op) \
+static inline void atomic_##op(int i, atomic_t *v) \
+{ \
+ unsigned long tmp, flags; \
+ \
+ raw_local_irq_save(flags); \
+ \
+ asm volatile ( \
+ " ldw %0, (%2) \n" \
+ " " #op " %0, %1 \n" \
+ " stw %0, (%2) \n" \
+ : "=&r" (tmp) \
+ : "r" (i), "r"(&v->counter) \
+ : "memory"); \
+ \
+ raw_local_irq_restore(flags); \
+}
+
+#define ATOMIC_OP_RETURN(op, c_op) \
+static inline int atomic_##op##_return(int i, atomic_t *v) \
+{ \
+ unsigned long tmp, ret, flags; \
+ \
+ raw_local_irq_save(flags); \
+ \
+ asm volatile ( \
+ " ldw %0, (%3) \n" \
+ " " #op " %0, %2 \n" \
+ " stw %0, (%3) \n" \
+ " mov %1, %0 \n" \
+ : "=&r" (tmp), "=&r" (ret) \
+ : "r" (i), "r"(&v->counter) \
+ : "memory"); \
+ \
+ raw_local_irq_restore(flags); \
+ \
+ return ret; \
+}
+
+#define ATOMIC_FETCH_OP(op, c_op) \
+static inline int atomic_fetch_##op(int i, atomic_t *v) \
+{ \
+ unsigned long tmp, ret, flags; \
+ \
+ raw_local_irq_save(flags); \
+ \
+ asm volatile ( \
+ " ldw %0, (%3) \n" \
+ " mov %1, %0 \n" \
+ " " #op " %0, %2 \n" \
+ " stw %0, (%3) \n" \
+ : "=&r" (tmp), "=&r" (ret) \
+ : "r" (i), "r"(&v->counter) \
+ : "memory"); \
+ \
+ raw_local_irq_restore(flags); \
+ \
+ return ret; \
+}
+
+#endif /* CONFIG_CPU_HAS_LDSTEX */
+
+#define atomic_add_return atomic_add_return
+ATOMIC_OP_RETURN(add, +)
+#define atomic_sub_return atomic_sub_return
+ATOMIC_OP_RETURN(sub, -)
+
+#define atomic_fetch_add atomic_fetch_add
+ATOMIC_FETCH_OP(add, +)
+#define atomic_fetch_sub atomic_fetch_sub
+ATOMIC_FETCH_OP(sub, -)
+#define atomic_fetch_and atomic_fetch_and
+ATOMIC_FETCH_OP(and, &)
+#define atomic_fetch_or atomic_fetch_or
+ATOMIC_FETCH_OP(or, |)
+#define atomic_fetch_xor atomic_fetch_xor
+ATOMIC_FETCH_OP(xor, ^)
+
+#define atomic_and atomic_and
+ATOMIC_OP(and, &)
+#define atomic_or atomic_or
+ATOMIC_OP(or, |)
+#define atomic_xor atomic_xor
+ATOMIC_OP(xor, ^)
+
+#undef ATOMIC_FETCH_OP
+#undef ATOMIC_OP_RETURN
+#undef ATOMIC_OP
+
+#include <asm-generic/atomic.h>
+
+#endif /* __ASM_CSKY_ATOMIC_H */
diff --git a/arch/csky/include/asm/cmpxchg.h b/arch/csky/include/asm/cmpxchg.h
new file mode 100644
index 0000000..8922453
--- /dev/null
+++ b/arch/csky/include/asm/cmpxchg.h
@@ -0,0 +1,73 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+
+#ifndef __ASM_CSKY_CMPXCHG_H
+#define __ASM_CSKY_CMPXCHG_H
+
+#ifdef CONFIG_CPU_HAS_LDSTEX
+#include <asm/barrier.h>
+
+extern void __bad_xchg(void);
+
+#define __xchg(new, ptr, size) \
+({ \
+ __typeof__(ptr) __ptr = (ptr); \
+ __typeof__(new) __new = (new); \
+ __typeof__(*(ptr)) __ret; \
+ unsigned long tmp; \
+ switch (size) { \
+ case 4: \
+ smp_mb(); \
+ asm volatile ( \
+ "1: ldex.w %0, (%3) \n" \
+ " mov %1, %2 \n" \
+ " stex.w %1, (%3) \n" \
+ " bez %1, 1b \n" \
+ : "=&r" (__ret), "=&r" (tmp) \
+ : "r" (__new), "r"(__ptr) \
+ :); \
+ smp_mb(); \
+ break; \
+ default: \
+ __bad_xchg(); \
+ } \
+ __ret; \
+})
+
+#define xchg(ptr, x) (__xchg((x), (ptr), sizeof(*(ptr))))
+
+#define __cmpxchg(ptr, old, new, size) \
+({ \
+ __typeof__(ptr) __ptr = (ptr); \
+ __typeof__(new) __new = (new); \
+ __typeof__(new) __tmp; \
+ __typeof__(old) __old = (old); \
+ __typeof__(*(ptr)) __ret; \
+ switch (size) { \
+ case 4: \
+ smp_mb(); \
+ asm volatile ( \
+ "1: ldex.w %0, (%3) \n" \
+ " cmpne %0, %4 \n" \
+ " bt 2f \n" \
+ " mov %1, %2 \n" \
+ " stex.w %1, (%3) \n" \
+ " bez %1, 1b \n" \
+ "2: \n" \
+ : "=&r" (__ret), "=&r" (__tmp) \
+ : "r" (__new), "r"(__ptr), "r"(__old) \
+ :); \
+ smp_mb(); \
+ break; \
+ default: \
+ __bad_xchg(); \
+ } \
+ __ret; \
+})
+
+#define cmpxchg(ptr, o, n) \
+ (__cmpxchg((ptr), (o), (n), sizeof(*(ptr))))
+#else
+#include <asm-generic/cmpxchg.h>
+#endif
+
+#endif /* __ASM_CSKY_CMPXCHG_H */
diff --git a/arch/csky/include/asm/spinlock.h b/arch/csky/include/asm/spinlock.h
new file mode 100644
index 0000000..0474603
--- /dev/null
+++ b/arch/csky/include/asm/spinlock.h
@@ -0,0 +1,274 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+
+#ifndef __ASM_CSKY_SPINLOCK_H
+#define __ASM_CSKY_SPINLOCK_H
+
+#include <linux/spinlock_types.h>
+#include <asm/barrier.h>
+
+#ifdef CONFIG_QUEUED_RWLOCKS
+
+/*
+ * Ticket-based spin-locking.
+ */
+static inline void arch_spin_lock(arch_spinlock_t *lock)
+{
+ arch_spinlock_t lockval;
+ u32 ticket_next = 1 << TICKET_NEXT;
+ u32 *p = &lock->lock;
+ u32 tmp;
+
+ asm volatile (
+ "1: ldex.w %0, (%2) \n"
+ " mov %1, %0 \n"
+ " add %0, %3 \n"
+ " stex.w %0, (%2) \n"
+ " bez %0, 1b \n"
+ : "=&r" (tmp), "=&r" (lockval)
+ : "r"(p), "r"(ticket_next)
+ : "cc");
+
+ while (lockval.tickets.next != lockval.tickets.owner)
+ lockval.tickets.owner = READ_ONCE(lock->tickets.owner);
+
+ smp_mb();
+}
+
+static inline int arch_spin_trylock(arch_spinlock_t *lock)
+{
+ u32 tmp, contended, res;
+ u32 ticket_next = 1 << TICKET_NEXT;
+ u32 *p = &lock->lock;
+
+ do {
+ asm volatile (
+ " ldex.w %0, (%3) \n"
+ " movi %2, 1 \n"
+ " rotli %1, %0, 16 \n"
+ " cmpne %1, %0 \n"
+ " bt 1f \n"
+ " movi %2, 0 \n"
+ " add %0, %0, %4 \n"
+ " stex.w %0, (%3) \n"
+ "1: \n"
+ : "=&r" (res), "=&r" (tmp), "=&r" (contended)
+ : "r"(p), "r"(ticket_next)
+ : "cc");
+ } while (!res);
+
+ if (!contended)
+ smp_mb();
+
+ return !contended;
+}
+
+static inline void arch_spin_unlock(arch_spinlock_t *lock)
+{
+ smp_mb();
+ lock->tickets.owner++;
+}
+
+static inline int arch_spin_value_unlocked(arch_spinlock_t lock)
+{
+ return lock.tickets.owner == lock.tickets.next;
+}
+
+static inline int arch_spin_is_locked(arch_spinlock_t *lock)
+{
+ return !arch_spin_value_unlocked(READ_ONCE(*lock));
+}
+
+static inline int arch_spin_is_contended(arch_spinlock_t *lock)
+{
+ struct __raw_tickets tickets = READ_ONCE(lock->tickets);
+
+ return (tickets.next - tickets.owner) > 1;
+}
+#define arch_spin_is_contended arch_spin_is_contended
+
+#include <asm/qrwlock.h>
+
+/* See include/linux/spinlock.h */
+#define smp_mb__after_spinlock() smp_mb()
+
+#else /* CONFIG_QUEUED_RWLOCKS */
+
+/*
+ * Test-and-set spin-locking.
+ */
+static inline void arch_spin_lock(arch_spinlock_t *lock)
+{
+ u32 *p = &lock->lock;
+ u32 tmp;
+
+ asm volatile (
+ "1: ldex.w %0, (%1) \n"
+ " bnez %0, 1b \n"
+ " movi %0, 1 \n"
+ " stex.w %0, (%1) \n"
+ " bez %0, 1b \n"
+ : "=&r" (tmp)
+ : "r"(p)
+ : "cc");
+ smp_mb();
+}
+
+static inline void arch_spin_unlock(arch_spinlock_t *lock)
+{
+ u32 *p = &lock->lock;
+ u32 tmp;
+
+ smp_mb();
+ asm volatile (
+ " movi %0, 0 \n"
+ " stw %0, (%1) \n"
+ : "=&r" (tmp)
+ : "r"(p)
+ : "cc");
+}
+
+static inline int arch_spin_trylock(arch_spinlock_t *lock)
+{
+ u32 *p = &lock->lock;
+ u32 tmp;
+
+ asm volatile (
+ "1: ldex.w %0, (%1) \n"
+ " bnez %0, 2f \n"
+ " movi %0, 1 \n"
+ " stex.w %0, (%1) \n"
+ " bez %0, 1b \n"
+ " movi %0, 0 \n"
+ "2: \n"
+ : "=&r" (tmp)
+ : "r"(p)
+ : "cc");
+
+ if (!tmp)
+ smp_mb();
+
+ return !tmp;
+}
+
+#define arch_spin_is_locked(x) (READ_ONCE((x)->lock) != 0)
+
+/*
+ * read lock/unlock/trylock
+ */
+static inline void arch_read_lock(arch_rwlock_t *lock)
+{
+ u32 *p = &lock->lock;
+ u32 tmp;
+
+ asm volatile (
+ "1: ldex.w %0, (%1) \n"
+ " blz %0, 1b \n"
+ " addi %0, 1 \n"
+ " stex.w %0, (%1) \n"
+ " bez %0, 1b \n"
+ : "=&r" (tmp)
+ : "r"(p)
+ : "cc");
+ smp_mb();
+}
+
+static inline void arch_read_unlock(arch_rwlock_t *lock)
+{
+ u32 *p = &lock->lock;
+ u32 tmp;
+
+ smp_mb();
+ asm volatile (
+ "1: ldex.w %0, (%1) \n"
+ " subi %0, 1 \n"
+ " stex.w %0, (%1) \n"
+ " bez %0, 1b \n"
+ : "=&r" (tmp)
+ : "r"(p)
+ : "cc");
+}
+
+static inline int arch_read_trylock(arch_rwlock_t *lock)
+{
+ u32 *p = &lock->lock;
+ u32 tmp;
+
+ asm volatile (
+ "1: ldex.w %0, (%1) \n"
+ " blz %0, 2f \n"
+ " addi %0, 1 \n"
+ " stex.w %0, (%1) \n"
+ " bez %0, 1b \n"
+ " movi %0, 0 \n"
+ "2: \n"
+ : "=&r" (tmp)
+ : "r"(p)
+ : "cc");
+
+ if (!tmp)
+ smp_mb();
+
+ return !tmp;
+}
+
+/*
+ * write lock/unlock/trylock
+ */
+static inline void arch_write_lock(arch_rwlock_t *lock)
+{
+ u32 *p = &lock->lock;
+ u32 tmp;
+
+ asm volatile (
+ "1: ldex.w %0, (%1) \n"
+ " bnez %0, 1b \n"
+ " subi %0, 1 \n"
+ " stex.w %0, (%1) \n"
+ " bez %0, 1b \n"
+ : "=&r" (tmp)
+ : "r"(p)
+ : "cc");
+ smp_mb();
+}
+
+static inline void arch_write_unlock(arch_rwlock_t *lock)
+{
+ u32 *p = &lock->lock;
+ u32 tmp;
+
+ smp_mb();
+ asm volatile (
+ "1: ldex.w %0, (%1) \n"
+ " movi %0, 0 \n"
+ " stex.w %0, (%1) \n"
+ " bez %0, 1b \n"
+ : "=&r" (tmp)
+ : "r"(p)
+ : "cc");
+}
+
+static inline int arch_write_trylock(arch_rwlock_t *lock)
+{
+ u32 *p = &lock->lock;
+ u32 tmp;
+
+ asm volatile (
+ "1: ldex.w %0, (%1) \n"
+ " bnez %0, 2f \n"
+ " subi %0, 1 \n"
+ " stex.w %0, (%1) \n"
+ " bez %0, 1b \n"
+ " movi %0, 0 \n"
+ "2: \n"
+ : "=&r" (tmp)
+ : "r"(p)
+ : "cc");
+
+ if (!tmp)
+ smp_mb();
+
+ return !tmp;
+}
+
+#endif /* CONFIG_QUEUED_RWLOCKS */
+#endif /* __ASM_CSKY_SPINLOCK_H */
diff --git a/arch/csky/include/asm/spinlock_types.h b/arch/csky/include/asm/spinlock_types.h
new file mode 100644
index 0000000..88b8243
--- /dev/null
+++ b/arch/csky/include/asm/spinlock_types.h
@@ -0,0 +1,37 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+
+#ifndef __ASM_CSKY_SPINLOCK_TYPES_H
+#define __ASM_CSKY_SPINLOCK_TYPES_H
+
+#ifndef __LINUX_SPINLOCK_TYPES_H
+# error "please don't include this file directly"
+#endif
+
+#define TICKET_NEXT 16
+
+typedef struct {
+ union {
+ u32 lock;
+ struct __raw_tickets {
+ /* little endian */
+ u16 owner;
+ u16 next;
+ } tickets;
+ };
+} arch_spinlock_t;
+
+#define __ARCH_SPIN_LOCK_UNLOCKED { { 0 } }
+
+#ifdef CONFIG_QUEUED_RWLOCKS
+#include <asm-generic/qrwlock_types.h>
+
+#else /* CONFIG_NR_CPUS > 2 */
+
+typedef struct {
+ u32 lock;
+} arch_rwlock_t;
+
+#define __ARCH_RW_LOCK_UNLOCKED { 0 }
+
+#endif /* CONFIG_QUEUED_RWLOCKS */
+#endif /* __ASM_CSKY_SPINLOCK_TYPES_H */
diff --git a/arch/csky/kernel/atomic.S b/arch/csky/kernel/atomic.S
new file mode 100644
index 0000000..d2357c8
--- /dev/null
+++ b/arch/csky/kernel/atomic.S
@@ -0,0 +1,87 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+// Copyright (C) 2018 Hangzhou C-SKY Microsystems co.,ltd.
+
+#include <linux/linkage.h>
+#include <abi/entry.h>
+
+.text
+
+/*
+ * int csky_cmpxchg(int oldval, int newval, int *ptr)
+ *
+ * If *ptr != oldval && return 1,
+ * else *ptr = newval return 0.
+ */
+#ifdef CONFIG_CPU_HAS_LDSTEX
+ENTRY(csky_cmpxchg)
+ USPTOKSP
+ mfcr a3, epc
+ INCTRAP a3
+
+ subi sp, 8
+ stw a3, (sp, 0)
+ mfcr a3, epsr
+ stw a3, (sp, 4)
+
+ psrset ee
+1:
+ ldex a3, (a2)
+ cmpne a0, a3
+ bt16 2f
+ mov a3, a1
+ stex a3, (a2)
+ bez a3, 1b
+2:
+ sync.is
+ mvc a0
+ ldw a3, (sp, 0)
+ mtcr a3, epc
+ ldw a3, (sp, 4)
+ mtcr a3, epsr
+ addi sp, 8
+ KSPTOUSP
+ rte
+END(csky_cmpxchg)
+#else
+ENTRY(csky_cmpxchg)
+ USPTOKSP
+ mfcr a3, epc
+ INCTRAP a3
+
+ subi sp, 8
+ stw a3, (sp, 0)
+ mfcr a3, epsr
+ stw a3, (sp, 4)
+
+ psrset ee
+1:
+ ldw a3, (a2)
+ cmpne a0, a3
+ bt16 3f
+2:
+ stw a1, (a2)
+3:
+ mvc a0
+ ldw a3, (sp, 0)
+ mtcr a3, epc
+ ldw a3, (sp, 4)
+ mtcr a3, epsr
+ addi sp, 8
+ KSPTOUSP
+ rte
+END(csky_cmpxchg)
+
+/*
+ * Called from tlbmodified exception
+ */
+ENTRY(csky_cmpxchg_fixup)
+ mfcr a0, epc
+ lrw a1, 2b
+ cmpne a1, a0
+ bt 1f
+ subi a1, (2b - 1b)
+ stw a1, (sp, LSAVE_PC)
+1:
+ rts
+END(csky_cmpxchg_fixup)
+#endif
--
2.7.4
^ permalink raw reply related [flat|nested] 68+ messages in thread* Re: [PATCH V9 11/21] csky: Atomic operations
2018-10-16 2:58 ` [PATCH V9 11/21] csky: Atomic operations Guo Ren
@ 2018-10-17 15:17 ` Arnd Bergmann
2018-10-18 2:40 ` Guo Ren
2018-10-21 20:55 ` Peter Zijlstra
1 sibling, 1 reply; 68+ messages in thread
From: Arnd Bergmann @ 2018-10-17 15:17 UTC (permalink / raw)
To: Guo Ren
Cc: Andrew Morton, Daniel Lezcano, David Miller, gregkh,
Christoph Hellwig, Marc Zyngier, Mark Rutland, Peter Zijlstra,
Rob Herring, Thomas Gleixner, Linux Kernel Mailing List,
linux-arch, DTML, Rob Herring, c-sky_gcc_upstream, andrea.parri
On Tue, Oct 16, 2018 at 5:02 AM Guo Ren <ren_guo@c-sky.com> wrote:
>
> This patch adds atomic, cmpxchg, spinlock files.
>
> Signed-off-by: Guo Ren <ren_guo@c-sky.com>
> Cc: Peter Zijlstra <peterz@infradead.org>
> Cc: Andrea Parri <andrea.parri@amarulasolutions.com>
> Cc: Arnd Bergmann <arnd@arndb.de>
I'd wait for Peter to give his final Ack here.
Arnd
^ permalink raw reply [flat|nested] 68+ messages in thread* Re: [PATCH V9 11/21] csky: Atomic operations
2018-10-17 15:17 ` Arnd Bergmann
@ 2018-10-18 2:40 ` Guo Ren
0 siblings, 0 replies; 68+ messages in thread
From: Guo Ren @ 2018-10-18 2:40 UTC (permalink / raw)
To: Arnd Bergmann
Cc: Andrew Morton, Daniel Lezcano, David Miller, gregkh,
Christoph Hellwig, Marc Zyngier, Mark Rutland, Peter Zijlstra,
Rob Herring, Thomas Gleixner, Linux Kernel Mailing List,
linux-arch, DTML, Rob Herring, c-sky_gcc_upstream, andrea.parri
On Wed, Oct 17, 2018 at 05:17:05PM +0200, Arnd Bergmann wrote:
> On Tue, Oct 16, 2018 at 5:02 AM Guo Ren <ren_guo@c-sky.com> wrote:
> >
> > This patch adds atomic, cmpxchg, spinlock files.
> >
> > Signed-off-by: Guo Ren <ren_guo@c-sky.com>
> > Cc: Peter Zijlstra <peterz@infradead.org>
> > Cc: Andrea Parri <andrea.parri@amarulasolutions.com>
> > Cc: Arnd Bergmann <arnd@arndb.de>
>
> I'd wait for Peter to give his final Ack here.
Yes, Peter help a lot to improve the patch.
Best Regards
Guo Ren
^ permalink raw reply [flat|nested] 68+ messages in thread
* Re: [PATCH V9 11/21] csky: Atomic operations
2018-10-16 2:58 ` [PATCH V9 11/21] csky: Atomic operations Guo Ren
2018-10-17 15:17 ` Arnd Bergmann
@ 2018-10-21 20:55 ` Peter Zijlstra
2018-10-22 1:52 ` Guo Ren
1 sibling, 1 reply; 68+ messages in thread
From: Peter Zijlstra @ 2018-10-21 20:55 UTC (permalink / raw)
To: Guo Ren
Cc: akpm, arnd, daniel.lezcano, davem, gregkh, hch, marc.zyngier,
mark.rutland, robh, tglx, linux-kernel, linux-arch, devicetree,
robh+dt, c-sky_gcc_upstream, Andrea Parri
On Tue, Oct 16, 2018 at 10:58:30AM +0800, Guo Ren wrote:
> diff --git a/arch/csky/include/asm/spinlock.h b/arch/csky/include/asm/spinlock.h
> new file mode 100644
> index 0000000..0474603
> --- /dev/null
> +++ b/arch/csky/include/asm/spinlock.h
> @@ -0,0 +1,274 @@
> +/* SPDX-License-Identifier: GPL-2.0 */
> +
> +#ifndef __ASM_CSKY_SPINLOCK_H
> +#define __ASM_CSKY_SPINLOCK_H
> +
> +#include <linux/spinlock_types.h>
> +#include <asm/barrier.h>
> +
> +#ifdef CONFIG_QUEUED_RWLOCKS
> +
> +/*
> + * Ticket-based spin-locking.
> + */
> +static inline void arch_spin_lock(arch_spinlock_t *lock)
> +{
> + arch_spinlock_t lockval;
> + u32 ticket_next = 1 << TICKET_NEXT;
> + u32 *p = &lock->lock;
> + u32 tmp;
> +
> + asm volatile (
> + "1: ldex.w %0, (%2) \n"
> + " mov %1, %0 \n"
> + " add %0, %3 \n"
> + " stex.w %0, (%2) \n"
> + " bez %0, 1b \n"
> + : "=&r" (tmp), "=&r" (lockval)
> + : "r"(p), "r"(ticket_next)
> + : "cc");
> +
> + while (lockval.tickets.next != lockval.tickets.owner)
> + lockval.tickets.owner = READ_ONCE(lock->tickets.owner);
> +
> + smp_mb();
> +}
> +
> +static inline int arch_spin_trylock(arch_spinlock_t *lock)
> +{
> + u32 tmp, contended, res;
> + u32 ticket_next = 1 << TICKET_NEXT;
> + u32 *p = &lock->lock;
> +
> + do {
> + asm volatile (
> + " ldex.w %0, (%3) \n"
> + " movi %2, 1 \n"
> + " rotli %1, %0, 16 \n"
> + " cmpne %1, %0 \n"
> + " bt 1f \n"
> + " movi %2, 0 \n"
> + " add %0, %0, %4 \n"
> + " stex.w %0, (%3) \n"
> + "1: \n"
> + : "=&r" (res), "=&r" (tmp), "=&r" (contended)
> + : "r"(p), "r"(ticket_next)
> + : "cc");
> + } while (!res);
> +
> + if (!contended)
> + smp_mb();
> +
> + return !contended;
> +}
> +
> +static inline void arch_spin_unlock(arch_spinlock_t *lock)
> +{
> + smp_mb();
> + lock->tickets.owner++;
WRITE_ONCE(lock->tickets.owner, lock->tickets.owner + 1);
> +}
> +
> +static inline int arch_spin_value_unlocked(arch_spinlock_t lock)
> +{
> + return lock.tickets.owner == lock.tickets.next;
> +}
> +
> +static inline int arch_spin_is_locked(arch_spinlock_t *lock)
> +{
> + return !arch_spin_value_unlocked(READ_ONCE(*lock));
> +}
> +
> +static inline int arch_spin_is_contended(arch_spinlock_t *lock)
> +{
> + struct __raw_tickets tickets = READ_ONCE(lock->tickets);
> +
> + return (tickets.next - tickets.owner) > 1;
> +}
> +#define arch_spin_is_contended arch_spin_is_contended
> +
> +#include <asm/qrwlock.h>
> +
> +/* See include/linux/spinlock.h */
> +#define smp_mb__after_spinlock() smp_mb()
> +
> +#else /* CONFIG_QUEUED_RWLOCKS */
> +
> +/*
> + * Test-and-set spin-locking.
> + */
I'm still not entirely sure why you want to have two spinlock
implementations; to me that is just extra maintenance overhead.
> +static inline void arch_spin_lock(arch_spinlock_t *lock)
> +{
> + u32 *p = &lock->lock;
> + u32 tmp;
> +
> + asm volatile (
> + "1: ldex.w %0, (%1) \n"
> + " bnez %0, 1b \n"
> + " movi %0, 1 \n"
> + " stex.w %0, (%1) \n"
> + " bez %0, 1b \n"
> + : "=&r" (tmp)
> + : "r"(p)
> + : "cc");
> + smp_mb();
> +}
> +
> +static inline void arch_spin_unlock(arch_spinlock_t *lock)
> +{
> + u32 *p = &lock->lock;
> + u32 tmp;
> +
> + smp_mb();
> + asm volatile (
> + " movi %0, 0 \n"
> + " stw %0, (%1) \n"
> + : "=&r" (tmp)
> + : "r"(p)
> + : "cc");
WRITE_ONCE(lock->lock, 0);
?
> +}
> +
> +static inline int arch_spin_trylock(arch_spinlock_t *lock)
> +{
> + u32 *p = &lock->lock;
> + u32 tmp;
> +
> + asm volatile (
> + "1: ldex.w %0, (%1) \n"
> + " bnez %0, 2f \n"
> + " movi %0, 1 \n"
> + " stex.w %0, (%1) \n"
> + " bez %0, 1b \n"
> + " movi %0, 0 \n"
> + "2: \n"
> + : "=&r" (tmp)
> + : "r"(p)
> + : "cc");
> +
> + if (!tmp)
> + smp_mb();
> +
> + return !tmp;
> +}
> +
> +#define arch_spin_is_locked(x) (READ_ONCE((x)->lock) != 0)
> +
> +/*
> + * read lock/unlock/trylock
> + */
Idem, why do you want a second rwlock_t implementation?
> +/*
> + * write lock/unlock/trylock
> + */
> +static inline void arch_write_lock(arch_rwlock_t *lock)
> +{
> + u32 *p = &lock->lock;
> + u32 tmp;
> +
> + asm volatile (
> + "1: ldex.w %0, (%1) \n"
> + " bnez %0, 1b \n"
> + " subi %0, 1 \n"
> + " stex.w %0, (%1) \n"
> + " bez %0, 1b \n"
> + : "=&r" (tmp)
> + : "r"(p)
> + : "cc");
> + smp_mb();
> +}
> +
> +static inline void arch_write_unlock(arch_rwlock_t *lock)
> +{
> + u32 *p = &lock->lock;
> + u32 tmp;
> +
> + smp_mb();
> + asm volatile (
> + "1: ldex.w %0, (%1) \n"
> + " movi %0, 0 \n"
> + " stex.w %0, (%1) \n"
> + " bez %0, 1b \n"
> + : "=&r" (tmp)
> + : "r"(p)
> + : "cc");
Isn't that:
WRITE_ONCE(lock->lock, 0);
> +}
> +
> +static inline int arch_write_trylock(arch_rwlock_t *lock)
> +{
> + u32 *p = &lock->lock;
> + u32 tmp;
> +
> + asm volatile (
> + "1: ldex.w %0, (%1) \n"
> + " bnez %0, 2f \n"
> + " subi %0, 1 \n"
> + " stex.w %0, (%1) \n"
> + " bez %0, 1b \n"
> + " movi %0, 0 \n"
> + "2: \n"
> + : "=&r" (tmp)
> + : "r"(p)
> + : "cc");
> +
> + if (!tmp)
> + smp_mb();
> +
> + return !tmp;
> +}
> diff --git a/arch/csky/kernel/atomic.S b/arch/csky/kernel/atomic.S
> new file mode 100644
> index 0000000..d2357c8
> --- /dev/null
> +++ b/arch/csky/kernel/atomic.S
> @@ -0,0 +1,87 @@
> +/* SPDX-License-Identifier: GPL-2.0 */
> +// Copyright (C) 2018 Hangzhou C-SKY Microsystems co.,ltd.
> +
> +#include <linux/linkage.h>
> +#include <abi/entry.h>
> +
> +.text
> +
> +/*
> + * int csky_cmpxchg(int oldval, int newval, int *ptr)
> + *
> + * If *ptr != oldval && return 1,
> + * else *ptr = newval return 0.
> + */
> +#ifdef CONFIG_CPU_HAS_LDSTEX
> +ENTRY(csky_cmpxchg)
> + USPTOKSP
> + mfcr a3, epc
> + INCTRAP a3
> +
> + subi sp, 8
> + stw a3, (sp, 0)
> + mfcr a3, epsr
> + stw a3, (sp, 4)
> +
> + psrset ee
> +1:
> + ldex a3, (a2)
> + cmpne a0, a3
> + bt16 2f
> + mov a3, a1
> + stex a3, (a2)
> + bez a3, 1b
> +2:
> + sync.is
> + mvc a0
> + ldw a3, (sp, 0)
> + mtcr a3, epc
> + ldw a3, (sp, 4)
> + mtcr a3, epsr
> + addi sp, 8
> + KSPTOUSP
> + rte
> +END(csky_cmpxchg)
I don't understand why you have this; if the CPU has ll/sc, why do you
need syscall support?
In any case, nothing terminally broken; so I suppose that's good enough
for starters. I just really don't understand some decisions (like having
two lock implementations and having that cmpxchg syscall when you have
hardware ll/sc).
Acked-by: Peter Zijlstra (Intel) <peterz@infradead.org>
^ permalink raw reply [flat|nested] 68+ messages in thread* Re: [PATCH V9 11/21] csky: Atomic operations
2018-10-21 20:55 ` Peter Zijlstra
@ 2018-10-22 1:52 ` Guo Ren
0 siblings, 0 replies; 68+ messages in thread
From: Guo Ren @ 2018-10-22 1:52 UTC (permalink / raw)
To: Peter Zijlstra
Cc: akpm, arnd, daniel.lezcano, davem, gregkh, hch, marc.zyngier,
mark.rutland, robh, tglx, linux-kernel, linux-arch, devicetree,
robh+dt, c-sky_gcc_upstream, Andrea Parri
Thx Peter,
Your review has been a great help.
On Sun, Oct 21, 2018 at 10:55:08PM +0200, Peter Zijlstra wrote:
> On Tue, Oct 16, 2018 at 10:58:30AM +0800, Guo Ren wrote:
> > + smp_mb();
> > + lock->tickets.owner++;
>
> WRITE_ONCE(lock->tickets.owner, lock->tickets.owner + 1);
Yes, approve! I should use WRITE_ONCE/READ_ONCE as necessary.
> > +/*
> > + * Test-and-set spin-locking.
> > + */
>
> I'm still not entirely sure why you want to have two spinlock
> implementations; to me that is just extra maintenance overhead.
Test and set (spinlock & rwlock) is easier for debug :P, and I don't
know the details of queue-rwlock (maybe I should learn it).
From education's view, we could teach students both of them in
arch/csky :)
Anyway, I just want to keep both of them.
Thx
> > + asm volatile (
> > + " movi %0, 0 \n"
> > + " stw %0, (%1) \n"
> > + : "=&r" (tmp)
> > + : "r"(p)
> > + : "cc");
>
> WRITE_ONCE(lock->lock, 0);
> ?
Cool ... I like WRITE_ONCE style.
> > +
> > +#define arch_spin_is_locked(x) (READ_ONCE((x)->lock) != 0)
> > +
> > +/*
> > + * read lock/unlock/trylock
> > + */
>
> Idem, why do you want a second rwlock_t implementation?
The same as above of spinlock.
> > + asm volatile (
> > + "1: ldex.w %0, (%1) \n"
> > + " movi %0, 0 \n"
> > + " stex.w %0, (%1) \n"
> > + " bez %0, 1b \n"
> > + : "=&r" (tmp)
> > + : "r"(p)
> > + : "cc");
>
> Isn't that:
>
> WRITE_ONCE(lock->lock, 0);
Yes, no need ldex/stex and you've mentioned in spinlock before :P
> > diff --git a/arch/csky/kernel/atomic.S b/arch/csky/kernel/atomic.S
> > new file mode 100644
> > index 0000000..d2357c8
> > --- /dev/null
> > +++ b/arch/csky/kernel/atomic.S
> > @@ -0,0 +1,87 @@
> > +/* SPDX-License-Identifier: GPL-2.0 */
> > +// Copyright (C) 2018 Hangzhou C-SKY Microsystems co.,ltd.
> > +
> > +#include <linux/linkage.h>
> > +#include <abi/entry.h>
> > +
> > +.text
> > +
> > +/*
> > + * int csky_cmpxchg(int oldval, int newval, int *ptr)
> > + *
> > + * If *ptr != oldval && return 1,
> > + * else *ptr = newval return 0.
> > + */
> > +#ifdef CONFIG_CPU_HAS_LDSTEX
> > +ENTRY(csky_cmpxchg)
> > + USPTOKSP
> > + mfcr a3, epc
> > + INCTRAP a3
> > +
> > + subi sp, 8
> > + stw a3, (sp, 0)
> > + mfcr a3, epsr
> > + stw a3, (sp, 4)
> > +
> > + psrset ee
> > +1:
> > + ldex a3, (a2)
> > + cmpne a0, a3
> > + bt16 2f
> > + mov a3, a1
> > + stex a3, (a2)
> > + bez a3, 1b
> > +2:
> > + sync.is
> > + mvc a0
> > + ldw a3, (sp, 0)
> > + mtcr a3, epc
> > + ldw a3, (sp, 4)
> > + mtcr a3, epsr
> > + addi sp, 8
> > + KSPTOUSP
> > + rte
> > +END(csky_cmpxchg)
>
> I don't understand why you have this; if the CPU has ll/sc, why do you
> need syscall support?
I've really considered your advice before, but from abi view 610/807/810
all have csky_cmpxchg trap and we want to make them the same. Some apps
use the trap directly and not use libc api. Maybe we could delete the
trap in future version of kernel.
>
> In any case, nothing terminally broken; so I suppose that's good enough
> for starters. I just really don't understand some decisions (like having
> two lock implementations and having that cmpxchg syscall when you have
> hardware ll/sc).
>
> Acked-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Thx peter and the two questions which I've clarified in above.
Best Regards
Guo Ren
^ permalink raw reply [flat|nested] 68+ messages in thread
* [PATCH V9 12/21] csky: ELF and module probe
2018-10-16 2:58 [PATCH V9 00/21] C-SKY(csky) Linux Kernel Port Guo Ren
` (10 preceding siblings ...)
2018-10-16 2:58 ` [PATCH V9 11/21] csky: Atomic operations Guo Ren
@ 2018-10-16 2:58 ` Guo Ren
2018-10-17 15:18 ` Arnd Bergmann
2018-10-16 2:58 ` [PATCH V9 13/21] csky: Library functions Guo Ren
` (9 subsequent siblings)
21 siblings, 1 reply; 68+ messages in thread
From: Guo Ren @ 2018-10-16 2:58 UTC (permalink / raw)
To: akpm, arnd, daniel.lezcano, davem, gregkh, hch, marc.zyngier,
mark.rutland, peterz, robh, tglx
Cc: linux-kernel, linux-arch, devicetree, robh+dt,
c-sky_gcc_upstream, Guo Ren
This patch adds ELF definition and module relocate codes.
Signed-off-by: Guo Ren <ren_guo@c-sky.com>
Cc: Arnd Bergmann <arnd@arndb.de>
---
arch/csky/abiv1/inc/abi/elf.h | 26 ++++++++++++
arch/csky/abiv2/inc/abi/elf.h | 43 ++++++++++++++++++++
arch/csky/include/asm/elf.h | 85 +++++++++++++++++++++++++++++++++++++++
arch/csky/kernel/module.c | 92 +++++++++++++++++++++++++++++++++++++++++++
4 files changed, 246 insertions(+)
create mode 100644 arch/csky/abiv1/inc/abi/elf.h
create mode 100644 arch/csky/abiv2/inc/abi/elf.h
create mode 100644 arch/csky/include/asm/elf.h
create mode 100644 arch/csky/kernel/module.c
diff --git a/arch/csky/abiv1/inc/abi/elf.h b/arch/csky/abiv1/inc/abi/elf.h
new file mode 100644
index 0000000..3058cc0
--- /dev/null
+++ b/arch/csky/abiv1/inc/abi/elf.h
@@ -0,0 +1,26 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+
+#ifndef __ABI_CSKY_ELF_H
+#define __ABI_CSKY_ELF_H
+
+#define ELF_CORE_COPY_REGS(pr_reg, regs) do { \
+ pr_reg[0] = regs->pc; \
+ pr_reg[1] = regs->regs[9]; \
+ pr_reg[2] = regs->usp; \
+ pr_reg[3] = regs->sr; \
+ pr_reg[4] = regs->a0; \
+ pr_reg[5] = regs->a1; \
+ pr_reg[6] = regs->a2; \
+ pr_reg[7] = regs->a3; \
+ pr_reg[8] = regs->regs[0]; \
+ pr_reg[9] = regs->regs[1]; \
+ pr_reg[10] = regs->regs[2]; \
+ pr_reg[11] = regs->regs[3]; \
+ pr_reg[12] = regs->regs[4]; \
+ pr_reg[13] = regs->regs[5]; \
+ pr_reg[14] = regs->regs[6]; \
+ pr_reg[15] = regs->regs[7]; \
+ pr_reg[16] = regs->regs[8]; \
+ pr_reg[17] = regs->lr; \
+} while (0);
+#endif /* __ABI_CSKY_ELF_H */
diff --git a/arch/csky/abiv2/inc/abi/elf.h b/arch/csky/abiv2/inc/abi/elf.h
new file mode 100644
index 0000000..290f49e
--- /dev/null
+++ b/arch/csky/abiv2/inc/abi/elf.h
@@ -0,0 +1,43 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+
+#ifndef __ABI_CSKY_ELF_H
+#define __ABI_CSKY_ELF_H
+
+/* The member sort in array pr_reg[x] is defined by GDB. */
+#define ELF_CORE_COPY_REGS(pr_reg, regs) do { \
+ pr_reg[0] = regs->pc; \
+ pr_reg[1] = regs->a1; \
+ pr_reg[2] = regs->a0; \
+ pr_reg[3] = regs->sr; \
+ pr_reg[4] = regs->a2; \
+ pr_reg[5] = regs->a3; \
+ pr_reg[6] = regs->regs[0]; \
+ pr_reg[7] = regs->regs[1]; \
+ pr_reg[8] = regs->regs[2]; \
+ pr_reg[9] = regs->regs[3]; \
+ pr_reg[10] = regs->regs[4]; \
+ pr_reg[11] = regs->regs[5]; \
+ pr_reg[12] = regs->regs[6]; \
+ pr_reg[13] = regs->regs[7]; \
+ pr_reg[14] = regs->regs[8]; \
+ pr_reg[15] = regs->regs[9]; \
+ pr_reg[16] = regs->usp; \
+ pr_reg[17] = regs->lr; \
+ pr_reg[18] = regs->exregs[0]; \
+ pr_reg[19] = regs->exregs[1]; \
+ pr_reg[20] = regs->exregs[2]; \
+ pr_reg[21] = regs->exregs[3]; \
+ pr_reg[22] = regs->exregs[4]; \
+ pr_reg[23] = regs->exregs[5]; \
+ pr_reg[24] = regs->exregs[6]; \
+ pr_reg[25] = regs->exregs[7]; \
+ pr_reg[26] = regs->exregs[8]; \
+ pr_reg[27] = regs->exregs[9]; \
+ pr_reg[28] = regs->exregs[10]; \
+ pr_reg[29] = regs->exregs[11]; \
+ pr_reg[30] = regs->exregs[12]; \
+ pr_reg[31] = regs->exregs[13]; \
+ pr_reg[32] = regs->exregs[14]; \
+ pr_reg[33] = regs->tls; \
+} while (0);
+#endif /* __ABI_CSKY_ELF_H */
diff --git a/arch/csky/include/asm/elf.h b/arch/csky/include/asm/elf.h
new file mode 100644
index 0000000..773b133
--- /dev/null
+++ b/arch/csky/include/asm/elf.h
@@ -0,0 +1,85 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+// Copyright (C) 2018 Hangzhou C-SKY Microsystems co.,ltd.
+
+#ifndef __ASM_CSKY_ELF_H
+#define __ASM_CSKY_ELF_H
+
+#include <asm/ptrace.h>
+#include <abi/regdef.h>
+
+#define ELF_ARCH 252
+
+/* CSKY Relocations */
+#define R_CSKY_NONE 0
+#define R_CSKY_32 1
+#define R_CSKY_PCIMM8BY4 2
+#define R_CSKY_PCIMM11BY2 3
+#define R_CSKY_PCIMM4BY2 4
+#define R_CSKY_PC32 5
+#define R_CSKY_PCRELJSR_IMM11BY2 6
+#define R_CSKY_GNU_VTINHERIT 7
+#define R_CSKY_GNU_VTENTRY 8
+#define R_CSKY_RELATIVE 9
+#define R_CSKY_COPY 10
+#define R_CSKY_GLOB_DAT 11
+#define R_CSKY_JUMP_SLOT 12
+#define R_CSKY_ADDR_HI16 24
+#define R_CSKY_ADDR_LO16 25
+#define R_CSKY_PCRELJSR_IMM26BY2 40
+
+typedef unsigned long elf_greg_t;
+
+typedef struct user_fp elf_fpregset_t;
+
+#define ELF_NGREG (sizeof(struct pt_regs) / sizeof(elf_greg_t))
+
+typedef elf_greg_t elf_gregset_t[ELF_NGREG];
+
+/*
+ * This is used to ensure we don't load something for the wrong architecture.
+ */
+#define elf_check_arch(x) ((x)->e_machine == ELF_ARCH)
+
+/*
+ * These are used to set parameters in the core dumps.
+ */
+#define USE_ELF_CORE_DUMP
+#define ELF_EXEC_PAGESIZE 4096
+#define ELF_CLASS ELFCLASS32
+#define ELF_PLAT_INIT(_r, load_addr) { _r->a0 = 0; }
+
+#ifdef __cskyBE__
+#define ELF_DATA ELFDATA2MSB
+#else
+#define ELF_DATA ELFDATA2LSB
+#endif
+
+/*
+ * This is the location that an ET_DYN program is loaded if exec'ed. Typical
+ * use of this is to invoke "./ld.so someprog" to test out a new version of
+ * the loader. We need to make sure that it is out of the way of the program
+ * that it will "exec", and that there is sufficient room for the brk.
+ */
+#define ELF_ET_DYN_BASE 0x0UL
+#include <abi/elf.h>
+
+/* Similar, but for a thread other than current. */
+struct task_struct;
+extern int dump_task_regs(struct task_struct *tsk, elf_gregset_t *elf_regs);
+#define ELF_CORE_COPY_TASK_REGS(tsk, elf_regs) dump_task_regs(tsk, elf_regs)
+
+#define ELF_HWCAP (0)
+
+/*
+ * This yields a string that ld.so will use to load implementation specific
+ * libraries for optimization. This is more specific in intent than poking
+ * at uname or /proc/cpuinfo.
+ */
+#define ELF_PLATFORM (NULL)
+#define SET_PERSONALITY(ex) set_personality(PER_LINUX)
+
+#define ARCH_HAS_SETUP_ADDITIONAL_PAGES 1
+struct linux_binprm;
+extern int arch_setup_additional_pages(struct linux_binprm *bprm,
+ int uses_interp);
+#endif /* __ASM_CSKY_ELF_H */
diff --git a/arch/csky/kernel/module.c b/arch/csky/kernel/module.c
new file mode 100644
index 0000000..65abab0
--- /dev/null
+++ b/arch/csky/kernel/module.c
@@ -0,0 +1,92 @@
+// SPDX-License-Identifier: GPL-2.0
+// Copyright (C) 2018 Hangzhou C-SKY Microsystems co.,ltd.
+
+#include <linux/moduleloader.h>
+#include <linux/elf.h>
+#include <linux/mm.h>
+#include <linux/vmalloc.h>
+#include <linux/slab.h>
+#include <linux/fs.h>
+#include <linux/string.h>
+#include <linux/kernel.h>
+#include <linux/spinlock.h>
+#include <asm/pgtable.h>
+
+#if defined(__CSKYABIV2__)
+#define IS_BSR32(hi16, lo16) (((hi16) & 0xFC00) == 0xE000)
+#define IS_JSRI32(hi16, lo16) ((hi16) == 0xEAE0)
+
+#define CHANGE_JSRI_TO_LRW(addr) do { \
+ *(uint16_t *)(addr) = (*(uint16_t *)(addr) & 0xFF9F) | 0x001a; \
+ *((uint16_t *)(addr) + 1) = *((uint16_t *)(addr) + 1) & 0xFFFF; \
+} while (0)
+
+#define SET_JSR32_R26(addr) do { \
+ *(uint16_t *)(addr) = 0xE8Fa; \
+ *((uint16_t *)(addr) + 1) = 0x0000; \
+} while (0)
+#endif
+
+int apply_relocate_add(Elf32_Shdr *sechdrs, const char *strtab,
+ unsigned int symindex, unsigned int relsec, struct module *me)
+{
+ unsigned int i;
+ Elf32_Rela *rel = (void *) sechdrs[relsec].sh_addr;
+ Elf32_Sym *sym;
+ uint32_t *location;
+ short *temp;
+#if defined(__CSKYABIV2__)
+ uint16_t *location_tmp;
+#endif
+
+ for (i = 0; i < sechdrs[relsec].sh_size / sizeof(*rel); i++) {
+ /* This is where to make the change */
+ location = (void *)sechdrs[sechdrs[relsec].sh_info].sh_addr
+ + rel[i].r_offset;
+ sym = (Elf32_Sym *)sechdrs[symindex].sh_addr
+ + ELF32_R_SYM(rel[i].r_info);
+
+ switch (ELF32_R_TYPE(rel[i].r_info)) {
+ case R_CSKY_32:
+ /* We add the value into the location given */
+ *location = rel[i].r_addend + sym->st_value;
+ break;
+ case R_CSKY_PC32:
+ /* Add the value, subtract its postition */
+ *location = rel[i].r_addend + sym->st_value
+ - (uint32_t)location;
+ break;
+ case R_CSKY_PCRELJSR_IMM11BY2:
+ break;
+ case R_CSKY_PCRELJSR_IMM26BY2:
+#if defined(__CSKYABIV2__)
+ location_tmp = (uint16_t *)location;
+ if (IS_BSR32(*location_tmp, *(location_tmp + 1)))
+ break;
+
+ if (IS_JSRI32(*location_tmp, *(location_tmp + 1))) {
+ /* jsri 0x... --> lrw r26, 0x... */
+ CHANGE_JSRI_TO_LRW(location);
+ /* lsli r0, r0 --> jsr r26 */
+ SET_JSR32_R26(location + 1);
+ }
+#endif
+ break;
+ case R_CSKY_ADDR_HI16:
+ temp = ((short *)location) + 1;
+ *temp = (short)
+ ((rel[i].r_addend + sym->st_value) >> 16);
+ break;
+ case R_CSKY_ADDR_LO16:
+ temp = ((short *)location) + 1;
+ *temp = (short)
+ ((rel[i].r_addend + sym->st_value) & 0xffff);
+ break;
+ default:
+ pr_err("module %s: Unknown relocation: %u\n",
+ me->name, ELF32_R_TYPE(rel[i].r_info));
+ return -ENOEXEC;
+ }
+ }
+ return 0;
+}
--
2.7.4
^ permalink raw reply related [flat|nested] 68+ messages in thread* Re: [PATCH V9 12/21] csky: ELF and module probe
2018-10-16 2:58 ` [PATCH V9 12/21] csky: ELF and module probe Guo Ren
@ 2018-10-17 15:18 ` Arnd Bergmann
2018-10-18 2:49 ` Guo Ren
0 siblings, 1 reply; 68+ messages in thread
From: Arnd Bergmann @ 2018-10-17 15:18 UTC (permalink / raw)
To: Guo Ren
Cc: Andrew Morton, Daniel Lezcano, David Miller, gregkh,
Christoph Hellwig, Marc Zyngier, Mark Rutland, Peter Zijlstra,
Rob Herring, Thomas Gleixner, Linux Kernel Mailing List,
linux-arch, DTML, Rob Herring, c-sky_gcc_upstream
On Tue, Oct 16, 2018 at 5:02 AM Guo Ren <ren_guo@c-sky.com> wrote:
>
> This patch adds ELF definition and module relocate codes.
>
> Signed-off-by: Guo Ren <ren_guo@c-sky.com>
> Cc: Arnd Bergmann <arnd@arndb.de>
> +#ifdef __cskyBE__
> +#define ELF_DATA ELFDATA2MSB
> +#else
> +#define ELF_DATA ELFDATA2LSB
> +#endif
You removed support for big-endian, right? I guess the #ifdef can also
get removed here then. Aside from that,
Reviewed-by: Arnd Bergmann <arnd@arndb.de>
^ permalink raw reply [flat|nested] 68+ messages in thread
* Re: [PATCH V9 12/21] csky: ELF and module probe
2018-10-17 15:18 ` Arnd Bergmann
@ 2018-10-18 2:49 ` Guo Ren
2018-10-18 8:31 ` Arnd Bergmann
0 siblings, 1 reply; 68+ messages in thread
From: Guo Ren @ 2018-10-18 2:49 UTC (permalink / raw)
To: Arnd Bergmann
Cc: Andrew Morton, Daniel Lezcano, David Miller, gregkh,
Christoph Hellwig, Marc Zyngier, Mark Rutland, Peter Zijlstra,
Rob Herring, Thomas Gleixner, Linux Kernel Mailing List,
linux-arch, DTML, Rob Herring, c-sky_gcc_upstream
On Wed, Oct 17, 2018 at 05:18:49PM +0200, Arnd Bergmann wrote:
> On Tue, Oct 16, 2018 at 5:02 AM Guo Ren <ren_guo@c-sky.com> wrote:
> >
> > This patch adds ELF definition and module relocate codes.
> >
> > Signed-off-by: Guo Ren <ren_guo@c-sky.com>
> > Cc: Arnd Bergmann <arnd@arndb.de>
>
> > +#ifdef __cskyBE__
> > +#define ELF_DATA ELFDATA2MSB
> > +#else
> > +#define ELF_DATA ELFDATA2LSB
> > +#endif
>
> You removed support for big-endian, right? I guess the #ifdef can also
> get removed here then. Aside from that,
Now, we only support little-endian and it could be removed. But maybe we
need support big-endian in future, so I keep it here.
Best Regards
Guo Ren
^ permalink raw reply [flat|nested] 68+ messages in thread
* Re: [PATCH V9 12/21] csky: ELF and module probe
2018-10-18 2:49 ` Guo Ren
@ 2018-10-18 8:31 ` Arnd Bergmann
0 siblings, 0 replies; 68+ messages in thread
From: Arnd Bergmann @ 2018-10-18 8:31 UTC (permalink / raw)
To: Guo Ren
Cc: Andrew Morton, Daniel Lezcano, David Miller, gregkh,
Christoph Hellwig, Marc Zyngier, Mark Rutland, Peter Zijlstra,
Rob Herring, Thomas Gleixner, Linux Kernel Mailing List,
linux-arch, DTML, Rob Herring, c-sky_gcc_upstream
On Thu, Oct 18, 2018 at 4:49 AM Guo Ren <ren_guo@c-sky.com> wrote:
>
> On Wed, Oct 17, 2018 at 05:18:49PM +0200, Arnd Bergmann wrote:
> > On Tue, Oct 16, 2018 at 5:02 AM Guo Ren <ren_guo@c-sky.com> wrote:
> > >
> > > This patch adds ELF definition and module relocate codes.
> > >
> > > Signed-off-by: Guo Ren <ren_guo@c-sky.com>
> > > Cc: Arnd Bergmann <arnd@arndb.de>
> >
> > > +#ifdef __cskyBE__
> > > +#define ELF_DATA ELFDATA2MSB
> > > +#else
> > > +#define ELF_DATA ELFDATA2LSB
> > > +#endif
> >
> > You removed support for big-endian, right? I guess the #ifdef can also
> > get removed here then. Aside from that,
> Now, we only support little-endian and it could be removed. But maybe we
> need support big-endian in future, so I keep it here.
Ok, just made sure it was intentional.
Arnd
^ permalink raw reply [flat|nested] 68+ messages in thread
* [PATCH V9 13/21] csky: Library functions
2018-10-16 2:58 [PATCH V9 00/21] C-SKY(csky) Linux Kernel Port Guo Ren
` (11 preceding siblings ...)
2018-10-16 2:58 ` [PATCH V9 12/21] csky: ELF and module probe Guo Ren
@ 2018-10-16 2:58 ` Guo Ren
2018-10-17 15:24 ` Arnd Bergmann
2018-10-16 2:58 ` [PATCH V9 14/21] csky: User access Guo Ren
` (8 subsequent siblings)
21 siblings, 1 reply; 68+ messages in thread
From: Guo Ren @ 2018-10-16 2:58 UTC (permalink / raw)
To: akpm, arnd, daniel.lezcano, davem, gregkh, hch, marc.zyngier,
mark.rutland, peterz, robh, tglx
Cc: linux-kernel, linux-arch, devicetree, robh+dt,
c-sky_gcc_upstream, Guo Ren, Chen Linfei, Mao Han
This patch adds string optimize codes and some auxiliary codes.
Signed-off-by: Chen Linfei <linfei_chen@c-sky.com>
Signed-off-by: Mao Han <han_mao@c-sky.com>
Signed-off-by: Guo Ren <ren_guo@c-sky.com>
Cc: Arnd Bergmann <arnd@arndb.de>
---
Changelog:
- Use bt instead of jbt in asm, jbt will cause relocation problem.
- remove kernel/platform.c
---
---
arch/csky/abiv1/bswapdi.c | 12 ++
arch/csky/abiv1/bswapsi.c | 12 ++
arch/csky/abiv1/inc/abi/string.h | 13 ++
arch/csky/abiv1/memcpy.S | 347 +++++++++++++++++++++++++++++++++++++++
arch/csky/abiv1/memset.c | 37 +++++
arch/csky/abiv1/strksyms.c | 7 +
arch/csky/abiv2/inc/abi/string.h | 27 +++
arch/csky/abiv2/memcmp.S | 152 +++++++++++++++++
arch/csky/abiv2/memcpy.S | 110 +++++++++++++
arch/csky/abiv2/memmove.S | 108 ++++++++++++
arch/csky/abiv2/memset.S | 83 ++++++++++
arch/csky/abiv2/strcmp.S | 168 +++++++++++++++++++
arch/csky/abiv2/strcpy.S | 123 ++++++++++++++
arch/csky/abiv2/strksyms.c | 12 ++
arch/csky/abiv2/strlen.S | 97 +++++++++++
arch/csky/abiv2/sysdep.h | 30 ++++
arch/csky/include/asm/string.h | 13 ++
arch/csky/kernel/power.c | 30 ++++
arch/csky/lib/delay.c | 39 +++++
19 files changed, 1420 insertions(+)
create mode 100644 arch/csky/abiv1/bswapdi.c
create mode 100644 arch/csky/abiv1/bswapsi.c
create mode 100644 arch/csky/abiv1/inc/abi/string.h
create mode 100644 arch/csky/abiv1/memcpy.S
create mode 100644 arch/csky/abiv1/memset.c
create mode 100644 arch/csky/abiv1/strksyms.c
create mode 100644 arch/csky/abiv2/inc/abi/string.h
create mode 100644 arch/csky/abiv2/memcmp.S
create mode 100644 arch/csky/abiv2/memcpy.S
create mode 100644 arch/csky/abiv2/memmove.S
create mode 100644 arch/csky/abiv2/memset.S
create mode 100644 arch/csky/abiv2/strcmp.S
create mode 100644 arch/csky/abiv2/strcpy.S
create mode 100644 arch/csky/abiv2/strksyms.c
create mode 100644 arch/csky/abiv2/strlen.S
create mode 100644 arch/csky/abiv2/sysdep.h
create mode 100644 arch/csky/include/asm/string.h
create mode 100644 arch/csky/kernel/power.c
create mode 100644 arch/csky/lib/delay.c
diff --git a/arch/csky/abiv1/bswapdi.c b/arch/csky/abiv1/bswapdi.c
new file mode 100644
index 0000000..f50a1d6
--- /dev/null
+++ b/arch/csky/abiv1/bswapdi.c
@@ -0,0 +1,12 @@
+// SPDX-License-Identifier: GPL-2.0
+// Copyright (C) 2018 Hangzhou C-SKY Microsystems co.,ltd.
+
+#include <linux/export.h>
+#include <linux/compiler.h>
+#include <uapi/linux/swab.h>
+
+unsigned long long notrace __bswapdi2(unsigned long long u)
+{
+ return ___constant_swab64(u);
+}
+EXPORT_SYMBOL(__bswapdi2);
diff --git a/arch/csky/abiv1/bswapsi.c b/arch/csky/abiv1/bswapsi.c
new file mode 100644
index 0000000..0f79182
--- /dev/null
+++ b/arch/csky/abiv1/bswapsi.c
@@ -0,0 +1,12 @@
+// SPDX-License-Identifier: GPL-2.0
+// Copyright (C) 2018 Hangzhou C-SKY Microsystems co.,ltd.
+
+#include <linux/export.h>
+#include <linux/compiler.h>
+#include <uapi/linux/swab.h>
+
+unsigned int notrace __bswapsi2(unsigned int u)
+{
+ return ___constant_swab32(u);
+}
+EXPORT_SYMBOL(__bswapsi2);
diff --git a/arch/csky/abiv1/inc/abi/string.h b/arch/csky/abiv1/inc/abi/string.h
new file mode 100644
index 0000000..5abe80b
--- /dev/null
+++ b/arch/csky/abiv1/inc/abi/string.h
@@ -0,0 +1,13 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+// Copyright (C) 2018 Hangzhou C-SKY Microsystems co.,ltd.
+
+#ifndef __ABI_CSKY_STRING_H
+#define __ABI_CSKY_STRING_H
+
+#define __HAVE_ARCH_MEMCPY
+extern void *memcpy(void *, const void *, __kernel_size_t);
+
+#define __HAVE_ARCH_MEMSET
+extern void *memset(void *, int, __kernel_size_t);
+
+#endif /* __ABI_CSKY_STRING_H */
diff --git a/arch/csky/abiv1/memcpy.S b/arch/csky/abiv1/memcpy.S
new file mode 100644
index 0000000..5078eb5
--- /dev/null
+++ b/arch/csky/abiv1/memcpy.S
@@ -0,0 +1,347 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+// Copyright (C) 2018 Hangzhou C-SKY Microsystems co.,ltd.
+
+#include <linux/linkage.h>
+
+.macro GET_FRONT_BITS rx y
+#ifdef __cskyLE__
+ lsri \rx, \y
+#else
+ lsli \rx, \y
+#endif
+.endm
+
+.macro GET_AFTER_BITS rx y
+#ifdef __cskyLE__
+ lsli \rx, \y
+#else
+ lsri \rx, \y
+#endif
+.endm
+
+/* void *memcpy(void *dest, const void *src, size_t n); */
+ENTRY(memcpy)
+ mov r7, r2
+ cmplti r4, 4
+ bt .L_copy_by_byte
+ mov r6, r2
+ andi r6, 3
+ cmpnei r6, 0
+ jbt .L_dest_not_aligned
+ mov r6, r3
+ andi r6, 3
+ cmpnei r6, 0
+ jbt .L_dest_aligned_but_src_not_aligned
+.L0:
+ cmplti r4, 16
+ jbt .L_aligned_and_len_less_16bytes
+ subi sp, 8
+ stw r8, (sp, 0)
+.L_aligned_and_len_larger_16bytes:
+ ldw r1, (r3, 0)
+ ldw r5, (r3, 4)
+ ldw r8, (r3, 8)
+ stw r1, (r7, 0)
+ ldw r1, (r3, 12)
+ stw r5, (r7, 4)
+ stw r8, (r7, 8)
+ stw r1, (r7, 12)
+ subi r4, 16
+ addi r3, 16
+ addi r7, 16
+ cmplti r4, 16
+ jbf .L_aligned_and_len_larger_16bytes
+ ldw r8, (sp, 0)
+ addi sp, 8
+ cmpnei r4, 0
+ jbf .L_return
+
+.L_aligned_and_len_less_16bytes:
+ cmplti r4, 4
+ bt .L_copy_by_byte
+.L1:
+ ldw r1, (r3, 0)
+ stw r1, (r7, 0)
+ subi r4, 4
+ addi r3, 4
+ addi r7, 4
+ cmplti r4, 4
+ jbf .L1
+ br .L_copy_by_byte
+
+.L_return:
+ rts
+
+.L_copy_by_byte: /* len less than 4 bytes */
+ cmpnei r4, 0
+ jbf .L_return
+.L4:
+ ldb r1, (r3, 0)
+ stb r1, (r7, 0)
+ addi r3, 1
+ addi r7, 1
+ decne r4
+ jbt .L4
+ rts
+
+/*
+ * If dest is not aligned, just copying some bytes makes the dest align.
+ * Afther that, we judge whether the src is aligned.
+ */
+.L_dest_not_aligned:
+ mov r5, r3
+ rsub r5, r5, r7
+ abs r5, r5
+ cmplt r5, r4
+ bt .L_copy_by_byte
+ mov r5, r7
+ sub r5, r3
+ cmphs r5, r4
+ bf .L_copy_by_byte
+ mov r5, r6
+.L5:
+ ldb r1, (r3, 0) /* makes the dest align. */
+ stb r1, (r7, 0)
+ addi r5, 1
+ subi r4, 1
+ addi r3, 1
+ addi r7, 1
+ cmpnei r5, 4
+ jbt .L5
+ cmplti r4, 4
+ jbt .L_copy_by_byte
+ mov r6, r3 /* judge whether the src is aligned. */
+ andi r6, 3
+ cmpnei r6, 0
+ jbf .L0
+
+/* Judge the number of misaligned, 1, 2, 3? */
+.L_dest_aligned_but_src_not_aligned:
+ mov r5, r3
+ rsub r5, r5, r7
+ abs r5, r5
+ cmplt r5, r4
+ bt .L_copy_by_byte
+ bclri r3, 0
+ bclri r3, 1
+ ldw r1, (r3, 0)
+ addi r3, 4
+ cmpnei r6, 2
+ bf .L_dest_aligned_but_src_not_aligned_2bytes
+ cmpnei r6, 3
+ bf .L_dest_aligned_but_src_not_aligned_3bytes
+
+.L_dest_aligned_but_src_not_aligned_1byte:
+ mov r5, r7
+ sub r5, r3
+ cmphs r5, r4
+ bf .L_copy_by_byte
+ cmplti r4, 16
+ bf .L11
+.L10: /* If the len is less than 16 bytes */
+ GET_FRONT_BITS r1 8
+ mov r5, r1
+ ldw r6, (r3, 0)
+ mov r1, r6
+ GET_AFTER_BITS r6 24
+ or r5, r6
+ stw r5, (r7, 0)
+ subi r4, 4
+ addi r3, 4
+ addi r7, 4
+ cmplti r4, 4
+ bf .L10
+ subi r3, 3
+ br .L_copy_by_byte
+.L11:
+ subi sp, 16
+ stw r8, (sp, 0)
+ stw r9, (sp, 4)
+ stw r10, (sp, 8)
+ stw r11, (sp, 12)
+.L12:
+ ldw r5, (r3, 0)
+ ldw r11, (r3, 4)
+ ldw r8, (r3, 8)
+ ldw r9, (r3, 12)
+
+ GET_FRONT_BITS r1 8 /* little or big endian? */
+ mov r10, r5
+ GET_AFTER_BITS r5 24
+ or r5, r1
+
+ GET_FRONT_BITS r10 8
+ mov r1, r11
+ GET_AFTER_BITS r11 24
+ or r11, r10
+
+ GET_FRONT_BITS r1 8
+ mov r10, r8
+ GET_AFTER_BITS r8 24
+ or r8, r1
+
+ GET_FRONT_BITS r10 8
+ mov r1, r9
+ GET_AFTER_BITS r9 24
+ or r9, r10
+
+ stw r5, (r7, 0)
+ stw r11, (r7, 4)
+ stw r8, (r7, 8)
+ stw r9, (r7, 12)
+ subi r4, 16
+ addi r3, 16
+ addi r7, 16
+ cmplti r4, 16
+ jbf .L12
+ ldw r8, (sp, 0)
+ ldw r9, (sp, 4)
+ ldw r10, (sp, 8)
+ ldw r11, (sp, 12)
+ addi sp , 16
+ cmplti r4, 4
+ bf .L10
+ subi r3, 3
+ br .L_copy_by_byte
+
+.L_dest_aligned_but_src_not_aligned_2bytes:
+ cmplti r4, 16
+ bf .L21
+.L20:
+ GET_FRONT_BITS r1 16
+ mov r5, r1
+ ldw r6, (r3, 0)
+ mov r1, r6
+ GET_AFTER_BITS r6 16
+ or r5, r6
+ stw r5, (r7, 0)
+ subi r4, 4
+ addi r3, 4
+ addi r7, 4
+ cmplti r4, 4
+ bf .L20
+ subi r3, 2
+ br .L_copy_by_byte
+ rts
+
+.L21: /* n > 16 */
+ subi sp, 16
+ stw r8, (sp, 0)
+ stw r9, (sp, 4)
+ stw r10, (sp, 8)
+ stw r11, (sp, 12)
+
+.L22:
+ ldw r5, (r3, 0)
+ ldw r11, (r3, 4)
+ ldw r8, (r3, 8)
+ ldw r9, (r3, 12)
+
+ GET_FRONT_BITS r1 16
+ mov r10, r5
+ GET_AFTER_BITS r5 16
+ or r5, r1
+
+ GET_FRONT_BITS r10 16
+ mov r1, r11
+ GET_AFTER_BITS r11 16
+ or r11, r10
+
+ GET_FRONT_BITS r1 16
+ mov r10, r8
+ GET_AFTER_BITS r8 16
+ or r8, r1
+
+ GET_FRONT_BITS r10 16
+ mov r1, r9
+ GET_AFTER_BITS r9 16
+ or r9, r10
+
+ stw r5, (r7, 0)
+ stw r11, (r7, 4)
+ stw r8, (r7, 8)
+ stw r9, (r7, 12)
+ subi r4, 16
+ addi r3, 16
+ addi r7, 16
+ cmplti r4, 16
+ jbf .L22
+ ldw r8, (sp, 0)
+ ldw r9, (sp, 4)
+ ldw r10, (sp, 8)
+ ldw r11, (sp, 12)
+ addi sp, 16
+ cmplti r4, 4
+ bf .L20
+ subi r3, 2
+ br .L_copy_by_byte
+
+
+.L_dest_aligned_but_src_not_aligned_3bytes:
+ cmplti r4, 16
+ bf .L31
+.L30:
+ GET_FRONT_BITS r1 24
+ mov r5, r1
+ ldw r6, (r3, 0)
+ mov r1, r6
+ GET_AFTER_BITS r6 8
+ or r5, r6
+ stw r5, (r7, 0)
+ subi r4, 4
+ addi r3, 4
+ addi r7, 4
+ cmplti r4, 4
+ bf .L30
+ subi r3, 1
+ br .L_copy_by_byte
+.L31:
+ subi sp, 16
+ stw r8, (sp, 0)
+ stw r9, (sp, 4)
+ stw r10, (sp, 8)
+ stw r11, (sp, 12)
+.L32:
+ ldw r5, (r3, 0)
+ ldw r11, (r3, 4)
+ ldw r8, (r3, 8)
+ ldw r9, (r3, 12)
+
+ GET_FRONT_BITS r1 24
+ mov r10, r5
+ GET_AFTER_BITS r5 8
+ or r5, r1
+
+ GET_FRONT_BITS r10 24
+ mov r1, r11
+ GET_AFTER_BITS r11 8
+ or r11, r10
+
+ GET_FRONT_BITS r1 24
+ mov r10, r8
+ GET_AFTER_BITS r8 8
+ or r8, r1
+
+ GET_FRONT_BITS r10 24
+ mov r1, r9
+ GET_AFTER_BITS r9 8
+ or r9, r10
+
+ stw r5, (r7, 0)
+ stw r11, (r7, 4)
+ stw r8, (r7, 8)
+ stw r9, (r7, 12)
+ subi r4, 16
+ addi r3, 16
+ addi r7, 16
+ cmplti r4, 16
+ jbf .L32
+ ldw r8, (sp, 0)
+ ldw r9, (sp, 4)
+ ldw r10, (sp, 8)
+ ldw r11, (sp, 12)
+ addi sp, 16
+ cmplti r4, 4
+ bf .L30
+ subi r3, 1
+ br .L_copy_by_byte
diff --git a/arch/csky/abiv1/memset.c b/arch/csky/abiv1/memset.c
new file mode 100644
index 0000000..b4aa75b
--- /dev/null
+++ b/arch/csky/abiv1/memset.c
@@ -0,0 +1,37 @@
+// SPDX-License-Identifier: GPL-2.0
+// Copyright (C) 2018 Hangzhou C-SKY Microsystems co.,ltd.
+
+#include <linux/types.h>
+
+void *memset(void *dest, int c, size_t l)
+{
+ char *d = dest;
+ int ch = c & 0xff;
+ int tmp = (ch | ch << 8 | ch << 16 | ch << 24);
+
+ while (((uintptr_t)d & 0x3) && l--)
+ *d++ = ch;
+
+ while (l >= 16) {
+ *(((u32 *)d)) = tmp;
+ *(((u32 *)d)+1) = tmp;
+ *(((u32 *)d)+2) = tmp;
+ *(((u32 *)d)+3) = tmp;
+ l -= 16;
+ d += 16;
+ }
+
+ while (l > 3) {
+ *(((u32 *)d)) = tmp;
+ l -= 4;
+ d += 4;
+ }
+
+ while (l) {
+ *d = ch;
+ l--;
+ d++;
+ }
+
+ return dest;
+}
diff --git a/arch/csky/abiv1/strksyms.c b/arch/csky/abiv1/strksyms.c
new file mode 100644
index 0000000..436995c
--- /dev/null
+++ b/arch/csky/abiv1/strksyms.c
@@ -0,0 +1,7 @@
+// SPDX-License-Identifier: GPL-2.0
+// Copyright (C) 2018 Hangzhou C-SKY Microsystems co.,ltd.
+
+#include <linux/module.h>
+
+EXPORT_SYMBOL(memcpy);
+EXPORT_SYMBOL(memset);
diff --git a/arch/csky/abiv2/inc/abi/string.h b/arch/csky/abiv2/inc/abi/string.h
new file mode 100644
index 0000000..f01bad2
--- /dev/null
+++ b/arch/csky/abiv2/inc/abi/string.h
@@ -0,0 +1,27 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+
+#ifndef __ABI_CSKY_STRING_H
+#define __ABI_CSKY_STRING_H
+
+#define __HAVE_ARCH_MEMCMP
+extern int memcmp(const void *, const void *, __kernel_size_t);
+
+#define __HAVE_ARCH_MEMCPY
+extern void *memcpy(void *, const void *, __kernel_size_t);
+
+#define __HAVE_ARCH_MEMMOVE
+extern void *memmove(void *, const void *, __kernel_size_t);
+
+#define __HAVE_ARCH_MEMSET
+extern void *memset(void *, int, __kernel_size_t);
+
+#define __HAVE_ARCH_STRCMP
+extern int strcmp(const char *, const char *);
+
+#define __HAVE_ARCH_STRCPY
+extern char *strcpy(char *, const char *);
+
+#define __HAVE_ARCH_STRLEN
+extern __kernel_size_t strlen(const char *);
+
+#endif /* __ABI_CSKY_STRING_H */
diff --git a/arch/csky/abiv2/memcmp.S b/arch/csky/abiv2/memcmp.S
new file mode 100644
index 0000000..bf0d809
--- /dev/null
+++ b/arch/csky/abiv2/memcmp.S
@@ -0,0 +1,152 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+// Copyright (C) 2018 Hangzhou C-SKY Microsystems co.,ltd.
+
+#include <linux/linkage.h>
+#include "sysdep.h"
+
+ENTRY(memcmp)
+ /* Test if len less than 4 bytes. */
+ mov r3, r0
+ movi r0, 0
+ mov r12, r4
+ cmplti r2, 4
+ bt .L_compare_by_byte
+
+ andi r13, r0, 3
+ movi r19, 4
+
+ /* Test if s1 is not 4 bytes aligned. */
+ bnez r13, .L_s1_not_aligned
+
+ LABLE_ALIGN
+.L_s1_aligned:
+ /* If dest is aligned, then copy. */
+ zext r18, r2, 31, 4
+ /* Test if len less than 16 bytes. */
+ bez r18, .L_compare_by_word
+
+.L_compare_by_4word:
+ /* If aligned, load word each time. */
+ ldw r20, (r3, 0)
+ ldw r21, (r1, 0)
+ /* If s1[i] != s2[i], goto .L_byte_check. */
+ cmpne r20, r21
+ bt .L_byte_check
+
+ ldw r20, (r3, 4)
+ ldw r21, (r1, 4)
+ cmpne r20, r21
+ bt .L_byte_check
+
+ ldw r20, (r3, 8)
+ ldw r21, (r1, 8)
+ cmpne r20, r21
+ bt .L_byte_check
+
+ ldw r20, (r3, 12)
+ ldw r21, (r1, 12)
+ cmpne r20, r21
+ bt .L_byte_check
+
+ PRE_BNEZAD (r18)
+ addi a3, 16
+ addi a1, 16
+
+ BNEZAD (r18, .L_compare_by_4word)
+
+.L_compare_by_word:
+ zext r18, r2, 3, 2
+ bez r18, .L_compare_by_byte
+.L_compare_by_word_loop:
+ ldw r20, (r3, 0)
+ ldw r21, (r1, 0)
+ addi r3, 4
+ PRE_BNEZAD (r18)
+ cmpne r20, r21
+ addi r1, 4
+ bt .L_byte_check
+ BNEZAD (r18, .L_compare_by_word_loop)
+
+.L_compare_by_byte:
+ zext r18, r2, 1, 0
+ bez r18, .L_return
+.L_compare_by_byte_loop:
+ ldb r0, (r3, 0)
+ ldb r4, (r1, 0)
+ addi r3, 1
+ subu r0, r4
+ PRE_BNEZAD (r18)
+ addi r1, 1
+ bnez r0, .L_return
+ BNEZAD (r18, .L_compare_by_byte_loop)
+
+.L_return:
+ mov r4, r12
+ rts
+
+# ifdef __CSKYBE__
+/* d[i] != s[i] in word, so we check byte 0. */
+.L_byte_check:
+ xtrb0 r0, r20
+ xtrb0 r2, r21
+ subu r0, r2
+ bnez r0, .L_return
+
+ /* check byte 1 */
+ xtrb1 r0, r20
+ xtrb1 r2, r21
+ subu r0, r2
+ bnez r0, .L_return
+
+ /* check byte 2 */
+ xtrb2 r0, r20
+ xtrb2 r2, r21
+ subu r0, r2
+ bnez r0, .L_return
+
+ /* check byte 3 */
+ xtrb3 r0, r20
+ xtrb3 r2, r21
+ subu r0, r2
+# else
+/* s1[i] != s2[i] in word, so we check byte 3. */
+.L_byte_check:
+ xtrb3 r0, r20
+ xtrb3 r2, r21
+ subu r0, r2
+ bnez r0, .L_return
+
+ /* check byte 2 */
+ xtrb2 r0, r20
+ xtrb2 r2, r21
+ subu r0, r2
+ bnez r0, .L_return
+
+ /* check byte 1 */
+ xtrb1 r0, r20
+ xtrb1 r2, r21
+ subu r0, r2
+ bnez r0, .L_return
+
+ /* check byte 0 */
+ xtrb0 r0, r20
+ xtrb0 r2, r21
+ subu r0, r2
+ br .L_return
+# endif /* !__CSKYBE__ */
+
+/* Compare when s1 is not aligned. */
+.L_s1_not_aligned:
+ sub r13, r19, r13
+ sub r2, r13
+.L_s1_not_aligned_loop:
+ ldb r0, (r3, 0)
+ ldb r4, (r1, 0)
+ addi r3, 1
+ subu r0, r4
+ PRE_BNEZAD (r13)
+ addi r1, 1
+ bnez r0, .L_return
+ BNEZAD (r13, .L_s1_not_aligned_loop)
+ br .L_s1_aligned
+ENDPROC(memcmp)
diff --git a/arch/csky/abiv2/memcpy.S b/arch/csky/abiv2/memcpy.S
new file mode 100644
index 0000000..987fec6
--- /dev/null
+++ b/arch/csky/abiv2/memcpy.S
@@ -0,0 +1,110 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+// Copyright (C) 2018 Hangzhou C-SKY Microsystems co.,ltd.
+
+#include <linux/linkage.h>
+#include "sysdep.h"
+
+ENTRY(__memcpy)
+ENTRY(memcpy)
+ /* Test if len less than 4 bytes. */
+ mov r12, r0
+ cmplti r2, 4
+ bt .L_copy_by_byte
+
+ andi r13, r0, 3
+ movi r19, 4
+ /* Test if dest is not 4 bytes aligned. */
+ bnez r13, .L_dest_not_aligned
+
+/* Hardware can handle unaligned access directly. */
+.L_dest_aligned:
+ /* If dest is aligned, then copy. */
+ zext r18, r2, 31, 4
+
+ /* Test if len less than 16 bytes. */
+ bez r18, .L_len_less_16bytes
+ movi r19, 0
+
+ LABLE_ALIGN
+.L_len_larger_16bytes:
+#if defined(__CSKY_VDSPV2__)
+ vldx.8 vr0, (r1), r19
+ PRE_BNEZAD (r18)
+ addi r1, 16
+ vstx.8 vr0, (r0), r19
+ addi r0, 16
+#elif defined(__CK860__)
+ ldw r3, (r1, 0)
+ stw r3, (r0, 0)
+ ldw r3, (r1, 4)
+ stw r3, (r0, 4)
+ ldw r3, (r1, 8)
+ stw r3, (r0, 8)
+ ldw r3, (r1, 12)
+ addi r1, 16
+ stw r3, (r0, 12)
+ addi r0, 16
+#else
+ ldw r20, (r1, 0)
+ ldw r21, (r1, 4)
+ ldw r22, (r1, 8)
+ ldw r23, (r1, 12)
+ stw r20, (r0, 0)
+ stw r21, (r0, 4)
+ stw r22, (r0, 8)
+ stw r23, (r0, 12)
+ PRE_BNEZAD (r18)
+ addi r1, 16
+ addi r0, 16
+#endif
+ BNEZAD (r18, .L_len_larger_16bytes)
+
+.L_len_less_16bytes:
+ zext r18, r2, 3, 2
+ bez r18, .L_copy_by_byte
+.L_len_less_16bytes_loop:
+ ldw r3, (r1, 0)
+ PRE_BNEZAD (r18)
+ addi r1, 4
+ stw r3, (r0, 0)
+ addi r0, 4
+ BNEZAD (r18, .L_len_less_16bytes_loop)
+
+/* Test if len less than 4 bytes. */
+.L_copy_by_byte:
+ zext r18, r2, 1, 0
+ bez r18, .L_return
+.L_copy_by_byte_loop:
+ ldb r3, (r1, 0)
+ PRE_BNEZAD (r18)
+ addi r1, 1
+ stb r3, (r0, 0)
+ addi r0, 1
+ BNEZAD (r18, .L_copy_by_byte_loop)
+
+.L_return:
+ mov r0, r12
+ rts
+
+/*
+ * If dest is not aligned, just copying some bytes makes the
+ * dest align.
+ */
+.L_dest_not_aligned:
+ sub r13, r19, r13
+ sub r2, r13
+
+/* Makes the dest align. */
+.L_dest_not_aligned_loop:
+ ldb r3, (r1, 0)
+ PRE_BNEZAD (r13)
+ addi r1, 1
+ stb r3, (r0, 0)
+ addi r0, 1
+ BNEZAD (r13, .L_dest_not_aligned_loop)
+ cmplti r2, 4
+ bt .L_copy_by_byte
+
+ /* Check whether the src is aligned. */
+ jbr .L_dest_aligned
+ENDPROC(__memcpy)
diff --git a/arch/csky/abiv2/memmove.S b/arch/csky/abiv2/memmove.S
new file mode 100644
index 0000000..b0c42ec
--- /dev/null
+++ b/arch/csky/abiv2/memmove.S
@@ -0,0 +1,108 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+// Copyright (C) 2018 Hangzhou C-SKY Microsystems co.,ltd.
+
+#include <linux/linkage.h>
+#include "sysdep.h"
+
+ .weak memmove
+ENTRY(__memmove)
+ENTRY(memmove)
+ subu r3, r0, r1
+ cmphs r3, r2
+ bt memcpy
+
+ mov r12, r0
+ addu r0, r0, r2
+ addu r1, r1, r2
+
+ /* Test if len less than 4 bytes. */
+ cmplti r2, 4
+ bt .L_copy_by_byte
+
+ andi r13, r0, 3
+ /* Test if dest is not 4 bytes aligned. */
+ bnez r13, .L_dest_not_aligned
+ /* Hardware can handle unaligned access directly. */
+.L_dest_aligned:
+ /* If dest is aligned, then copy. */
+ zext r18, r2, 31, 4
+ /* Test if len less than 16 bytes. */
+ bez r18, .L_len_less_16bytes
+ movi r19, 0
+
+ /* len > 16 bytes */
+ LABLE_ALIGN
+.L_len_larger_16bytes:
+ subi r1, 16
+ subi r0, 16
+#if defined(__CSKY_VDSPV2__)
+ vldx.8 vr0, (r1), r19
+ PRE_BNEZAD (r18)
+ vstx.8 vr0, (r0), r19
+#elif defined(__CK860__)
+ ldw r3, (r1, 12)
+ stw r3, (r0, 12)
+ ldw r3, (r1, 8)
+ stw r3, (r0, 8)
+ ldw r3, (r1, 4)
+ stw r3, (r0, 4)
+ ldw r3, (r1, 0)
+ stw r3, (r0, 0)
+#else
+ ldw r20, (r1, 0)
+ ldw r21, (r1, 4)
+ ldw r22, (r1, 8)
+ ldw r23, (r1, 12)
+ stw r20, (r0, 0)
+ stw r21, (r0, 4)
+ stw r22, (r0, 8)
+ stw r23, (r0, 12)
+ PRE_BNEZAD (r18)
+#endif
+ BNEZAD (r18, .L_len_larger_16bytes)
+
+.L_len_less_16bytes:
+ zext r18, r2, 3, 2
+ bez r18, .L_copy_by_byte
+.L_len_less_16bytes_loop:
+ subi r1, 4
+ subi r0, 4
+ ldw r3, (r1, 0)
+ PRE_BNEZAD (r18)
+ stw r3, (r0, 0)
+ BNEZAD (r18, .L_len_less_16bytes_loop)
+
+ /* Test if len less than 4 bytes. */
+.L_copy_by_byte:
+ zext r18, r2, 1, 0
+ bez r18, .L_return
+.L_copy_by_byte_loop:
+ subi r1, 1
+ subi r0, 1
+ ldb r3, (r1, 0)
+ PRE_BNEZAD (r18)
+ stb r3, (r0, 0)
+ BNEZAD (r18, .L_copy_by_byte_loop)
+
+.L_return:
+ mov r0, r12
+ rts
+
+ /* If dest is not aligned, just copy some bytes makes the dest
+ align. */
+.L_dest_not_aligned:
+ sub r2, r13
+.L_dest_not_aligned_loop:
+ subi r1, 1
+ subi r0, 1
+ /* Makes the dest align. */
+ ldb r3, (r1, 0)
+ PRE_BNEZAD (r13)
+ stb r3, (r0, 0)
+ BNEZAD (r13, .L_dest_not_aligned_loop)
+ cmplti r2, 4
+ bt .L_copy_by_byte
+ /* Check whether the src is aligned. */
+ jbr .L_dest_aligned
+ENDPROC(memmove)
+ENDPROC(__memmove)
diff --git a/arch/csky/abiv2/memset.S b/arch/csky/abiv2/memset.S
new file mode 100644
index 0000000..a7e7d99
--- /dev/null
+++ b/arch/csky/abiv2/memset.S
@@ -0,0 +1,83 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+// Copyright (C) 2018 Hangzhou C-SKY Microsystems co.,ltd.
+
+#include <linux/linkage.h>
+#include "sysdep.h"
+
+ .weak memset
+ENTRY(__memset)
+ENTRY(memset)
+ /* Test if len less than 4 bytes. */
+ mov r12, r0
+ cmplti r2, 8
+ bt .L_set_by_byte
+
+ andi r13, r0, 3
+ movi r19, 4
+ /* Test if dest is not 4 bytes aligned. */
+ bnez r13, .L_dest_not_aligned
+ /* Hardware can handle unaligned access directly. */
+.L_dest_aligned:
+ zextb r3, r1
+ lsli r1, 8
+ or r1, r3
+ lsli r3, r1, 16
+ or r3, r1
+
+ /* If dest is aligned, then copy. */
+ zext r18, r2, 31, 4
+ /* Test if len less than 16 bytes. */
+ bez r18, .L_len_less_16bytes
+
+ LABLE_ALIGN
+.L_len_larger_16bytes:
+ stw r3, (r0, 0)
+ stw r3, (r0, 4)
+ stw r3, (r0, 8)
+ stw r3, (r0, 12)
+ PRE_BNEZAD (r18)
+ addi r0, 16
+ BNEZAD (r18, .L_len_larger_16bytes)
+
+.L_len_less_16bytes:
+ zext r18, r2, 3, 2
+ andi r2, 3
+ bez r18, .L_set_by_byte
+.L_len_less_16bytes_loop:
+ stw r3, (r0, 0)
+ PRE_BNEZAD (r18)
+ addi r0, 4
+ BNEZAD (r18, .L_len_less_16bytes_loop)
+
+ /* Test if len less than 4 bytes. */
+.L_set_by_byte:
+ zext r18, r2, 2, 0
+ bez r18, .L_return
+.L_set_by_byte_loop:
+ stb r1, (r0, 0)
+ PRE_BNEZAD (r18)
+ addi r0, 1
+ BNEZAD (r18, .L_set_by_byte_loop)
+
+.L_return:
+ mov r0, r12
+ rts
+
+ /* If dest is not aligned, just set some bytes makes the dest
+ align. */
+
+.L_dest_not_aligned:
+ sub r13, r19, r13
+ sub r2, r13
+.L_dest_not_aligned_loop:
+ /* Makes the dest align. */
+ stb r1, (r0, 0)
+ PRE_BNEZAD (r13)
+ addi r0, 1
+ BNEZAD (r13, .L_dest_not_aligned_loop)
+ cmplti r2, 8
+ bt .L_set_by_byte
+ /* Check whether the src is aligned. */
+ jbr .L_dest_aligned
+ENDPROC(memset)
+ENDPROC(__memset)
diff --git a/arch/csky/abiv2/strcmp.S b/arch/csky/abiv2/strcmp.S
new file mode 100644
index 0000000..f8403f4
--- /dev/null
+++ b/arch/csky/abiv2/strcmp.S
@@ -0,0 +1,168 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+// Copyright (C) 2018 Hangzhou C-SKY Microsystems co.,ltd.
+
+#include <linux/linkage.h>
+#include "sysdep.h"
+
+ENTRY(strcmp)
+ mov a3, a0
+ /* Check if the s1 addr is aligned. */
+ xor a2, a3, a1
+ andi a2, 0x3
+ bnez a2, 7f
+ andi t1, a0, 0x3
+ bnez t1, 5f
+
+1:
+ /* If aligned, load word each time. */
+ ldw t0, (a3, 0)
+ ldw t1, (a1, 0)
+ /* If s1[i] != s2[i], goto 2f. */
+ cmpne t0, t1
+ bt 2f
+ /* If s1[i] == s2[i], check if s1 or s2 is at the end. */
+ tstnbz t0
+ /* If at the end, goto 3f (finish comparing). */
+ bf 3f
+
+ ldw t0, (a3, 4)
+ ldw t1, (a1, 4)
+ cmpne t0, t1
+ bt 2f
+ tstnbz t0
+ bf 3f
+
+ ldw t0, (a3, 8)
+ ldw t1, (a1, 8)
+ cmpne t0, t1
+ bt 2f
+ tstnbz t0
+ bf 3f
+
+ ldw t0, (a3, 12)
+ ldw t1, (a1, 12)
+ cmpne t0, t1
+ bt 2f
+ tstnbz t0
+ bf 3f
+
+ ldw t0, (a3, 16)
+ ldw t1, (a1, 16)
+ cmpne t0, t1
+ bt 2f
+ tstnbz t0
+ bf 3f
+
+ ldw t0, (a3, 20)
+ ldw t1, (a1, 20)
+ cmpne t0, t1
+ bt 2f
+ tstnbz t0
+ bf 3f
+
+ ldw t0, (a3, 24)
+ ldw t1, (a1, 24)
+ cmpne t0, t1
+ bt 2f
+ tstnbz t0
+ bf 3f
+
+ ldw t0, (a3, 28)
+ ldw t1, (a1, 28)
+ cmpne t0, t1
+ bt 2f
+ tstnbz t0
+ bf 3f
+
+ addi a3, 32
+ addi a1, 32
+
+ br 1b
+
+# ifdef __CSKYBE__
+ /* d[i] != s[i] in word, so we check byte 0. */
+2:
+ xtrb0 a0, t0
+ xtrb0 a2, t1
+ subu a0, a2
+ bez a2, 4f
+ bnez a0, 4f
+
+ /* check byte 1 */
+ xtrb1 a0, t0
+ xtrb1 a2, t1
+ subu a0, a2
+ bez a2, 4f
+ bnez a0, 4f
+
+ /* check byte 2 */
+ xtrb2 a0, t0
+ xtrb2 a2, t1
+ subu a0, a2
+ bez a2, 4f
+ bnez a0, 4f
+
+ /* check byte 3 */
+ xtrb3 a0, t0
+ xtrb3 a2, t1
+ subu a0, a2
+# else
+ /* s1[i] != s2[i] in word, so we check byte 3. */
+2:
+ xtrb3 a0, t0
+ xtrb3 a2, t1
+ subu a0, a2
+ bez a2, 4f
+ bnez a0, 4f
+
+ /* check byte 2 */
+ xtrb2 a0, t0
+ xtrb2 a2, t1
+ subu a0, a2
+ bez a2, 4f
+ bnez a0, 4f
+
+ /* check byte 1 */
+ xtrb1 a0, t0
+ xtrb1 a2, t1
+ subu a0, a2
+ bez a2, 4f
+ bnez a0, 4f
+
+ /* check byte 0 */
+ xtrb0 a0, t0
+ xtrb0 a2, t1
+ subu a0, a2
+
+# endif /* !__CSKYBE__ */
+ jmp lr
+3:
+ movi a0, 0
+4:
+ jmp lr
+
+ /* Compare when s1 or s2 is not aligned. */
+5:
+ subi t1, 4
+6:
+ ldb a0, (a3, 0)
+ ldb a2, (a1, 0)
+ subu a0, a2
+ bez a2, 4b
+ bnez a0, 4b
+ addi t1, 1
+ addi a1, 1
+ addi a3, 1
+ bnez t1, 6b
+ br 1b
+
+7:
+ ldb a0, (a3, 0)
+ addi a3, 1
+ ldb a2, (a1, 0)
+ addi a1, 1
+ subu a0, a2
+ bnez a0, 4b
+ bnez a2, 7b
+ jmp r15
+ENDPROC(strcmp)
diff --git a/arch/csky/abiv2/strcpy.S b/arch/csky/abiv2/strcpy.S
new file mode 100644
index 0000000..3c6d3f6
--- /dev/null
+++ b/arch/csky/abiv2/strcpy.S
@@ -0,0 +1,123 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+// Copyright (C) 2018 Hangzhou C-SKY Microsystems co.,ltd.
+
+#include <linux/linkage.h>
+#include "sysdep.h"
+
+ENTRY(strcpy)
+ mov a3, a0
+ /* Check if the src addr is aligned. */
+ andi t0, a1, 3
+ bnez t0, 11f
+1:
+ /* Check if all the bytes in the word are not zero. */
+ ldw a2, (a1)
+ tstnbz a2
+ bf 9f
+ stw a2, (a3)
+
+ ldw a2, (a1, 4)
+ tstnbz a2
+ bf 2f
+ stw a2, (a3, 4)
+
+ ldw a2, (a1, 8)
+ tstnbz a2
+ bf 3f
+ stw a2, (a3, 8)
+
+ ldw a2, (a1, 12)
+ tstnbz a2
+ bf 4f
+ stw a2, (a3, 12)
+
+ ldw a2, (a1, 16)
+ tstnbz a2
+ bf 5f
+ stw a2, (a3, 16)
+
+ ldw a2, (a1, 20)
+ tstnbz a2
+ bf 6f
+ stw a2, (a3, 20)
+
+ ldw a2, (a1, 24)
+ tstnbz a2
+ bf 7f
+ stw a2, (a3, 24)
+
+ ldw a2, (a1, 28)
+ tstnbz a2
+ bf 8f
+ stw a2, (a3, 28)
+
+ addi a3, 32
+ addi a1, 32
+ br 1b
+
+
+2:
+ addi a3, 4
+ br 9f
+
+3:
+ addi a3, 8
+ br 9f
+
+4:
+ addi a3, 12
+ br 9f
+
+5:
+ addi a3, 16
+ br 9f
+
+6:
+ addi a3, 20
+ br 9f
+
+7:
+ addi a3, 24
+ br 9f
+
+8:
+ addi a3, 28
+9:
+# ifdef __CSKYBE__
+ xtrb0 t0, a2
+ st.b t0, (a3)
+ bez t0, 10f
+ xtrb1 t0, a2
+ st.b t0, (a3, 1)
+ bez t0, 10f
+ xtrb2 t0, a2
+ st.b t0, (a3, 2)
+ bez t0, 10f
+ stw a2, (a3)
+# else
+ xtrb3 t0, a2
+ st.b t0, (a3)
+ bez t0, 10f
+ xtrb2 t0, a2
+ st.b t0, (a3, 1)
+ bez t0, 10f
+ xtrb1 t0, a2
+ st.b t0, (a3, 2)
+ bez t0, 10f
+ stw a2, (a3)
+# endif /* !__CSKYBE__ */
+10:
+ jmp lr
+
+11:
+ subi t0, 4
+12:
+ ld.b a2, (a1)
+ st.b a2, (a3)
+ bez a2, 10b
+ addi t0, 1
+ addi a1, a1, 1
+ addi a3, a3, 1
+ bnez t0, 12b
+ jbr 1b
+ENDPROC(strcpy)
diff --git a/arch/csky/abiv2/strksyms.c b/arch/csky/abiv2/strksyms.c
new file mode 100644
index 0000000..06da723
--- /dev/null
+++ b/arch/csky/abiv2/strksyms.c
@@ -0,0 +1,12 @@
+// SPDX-License-Identifier: GPL-2.0
+// Copyright (C) 2018 Hangzhou C-SKY Microsystems co.,ltd.
+
+#include <linux/module.h>
+
+EXPORT_SYMBOL(memcpy);
+EXPORT_SYMBOL(memset);
+EXPORT_SYMBOL(memcmp);
+EXPORT_SYMBOL(memmove);
+EXPORT_SYMBOL(strcmp);
+EXPORT_SYMBOL(strcpy);
+EXPORT_SYMBOL(strlen);
diff --git a/arch/csky/abiv2/strlen.S b/arch/csky/abiv2/strlen.S
new file mode 100644
index 0000000..bcdd707
--- /dev/null
+++ b/arch/csky/abiv2/strlen.S
@@ -0,0 +1,97 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+// Copyright (C) 2018 Hangzhou C-SKY Microsystems co.,ltd.
+
+#include <linux/linkage.h>
+#include "sysdep.h"
+
+ENTRY(strlen)
+ /* Check if the start addr is aligned. */
+ mov r3, r0
+ andi r1, r0, 3
+ movi r2, 4
+ movi r0, 0
+ bnez r1, .L_start_not_aligned
+
+ LABLE_ALIGN
+.L_start_addr_aligned:
+ /* Check if all the bytes in the word are not zero. */
+ ldw r1, (r3)
+ tstnbz r1
+ bf .L_string_tail
+
+ ldw r1, (r3, 4)
+ addi r0, 4
+ tstnbz r1
+ bf .L_string_tail
+
+ ldw r1, (r3, 8)
+ addi r0, 4
+ tstnbz r1
+ bf .L_string_tail
+
+ ldw r1, (r3, 12)
+ addi r0, 4
+ tstnbz r1
+ bf .L_string_tail
+
+ ldw r1, (r3, 16)
+ addi r0, 4
+ tstnbz r1
+ bf .L_string_tail
+
+ ldw r1, (r3, 20)
+ addi r0, 4
+ tstnbz r1
+ bf .L_string_tail
+
+ ldw r1, (r3, 24)
+ addi r0, 4
+ tstnbz r1
+ bf .L_string_tail
+
+ ldw r1, (r3, 28)
+ addi r0, 4
+ tstnbz r1
+ bf .L_string_tail
+
+ addi r0, 4
+ addi r3, 32
+ br .L_start_addr_aligned
+
+.L_string_tail:
+# ifdef __CSKYBE__
+ xtrb0 r3, r1
+ bez r3, .L_return
+ addi r0, 1
+ xtrb1 r3, r1
+ bez r3, .L_return
+ addi r0, 1
+ xtrb2 r3, r1
+ bez r3, .L_return
+ addi r0, 1
+# else
+ xtrb3 r3, r1
+ bez r3, .L_return
+ addi r0, 1
+ xtrb2 r3, r1
+ bez r3, .L_return
+ addi r0, 1
+ xtrb1 r3, r1
+ bez r3, .L_return
+ addi r0, 1
+# endif /* !__CSKYBE__ */
+
+.L_return:
+ rts
+
+.L_start_not_aligned:
+ sub r2, r2, r1
+.L_start_not_aligned_loop:
+ ldb r1, (r3)
+ PRE_BNEZAD (r2)
+ addi r3, 1
+ bez r1, .L_return
+ addi r0, 1
+ BNEZAD (r2, .L_start_not_aligned_loop)
+ br .L_start_addr_aligned
+ENDPROC(strlen)
diff --git a/arch/csky/abiv2/sysdep.h b/arch/csky/abiv2/sysdep.h
new file mode 100644
index 0000000..bbbedfd
--- /dev/null
+++ b/arch/csky/abiv2/sysdep.h
@@ -0,0 +1,30 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+// Copyright (C) 2018 Hangzhou C-SKY Microsystems co.,ltd.
+
+#ifndef __SYSDEP_H
+#define __SYSDEP_H
+
+#ifdef __ASSEMBLER__
+
+#if defined(__CK860__)
+#define LABLE_ALIGN \
+ .balignw 16, 0x6c03
+
+#define PRE_BNEZAD(R)
+
+#define BNEZAD(R, L) \
+ bnezad R, L
+#else
+#define LABLE_ALIGN \
+ .balignw 8, 0x6c03
+
+#define PRE_BNEZAD(R) \
+ subi R, 1
+
+#define BNEZAD(R, L) \
+ bnez R, L
+#endif
+
+#endif
+
+#endif
diff --git a/arch/csky/include/asm/string.h b/arch/csky/include/asm/string.h
new file mode 100644
index 0000000..73142de
--- /dev/null
+++ b/arch/csky/include/asm/string.h
@@ -0,0 +1,13 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+// Copyright (C) 2018 Hangzhou C-SKY Microsystems co.,ltd.
+
+#ifndef _CSKY_STRING_MM_H_
+#define _CSKY_STRING_MM_H_
+
+#ifndef __ASSEMBLY__
+#include <linux/types.h>
+#include <linux/compiler.h>
+#include <abi/string.h>
+#endif
+
+#endif /* _CSKY_STRING_MM_H_ */
diff --git a/arch/csky/kernel/power.c b/arch/csky/kernel/power.c
new file mode 100644
index 0000000..923ee4e
--- /dev/null
+++ b/arch/csky/kernel/power.c
@@ -0,0 +1,30 @@
+// SPDX-License-Identifier: GPL-2.0
+// Copyright (C) 2018 Hangzhou C-SKY Microsystems co.,ltd.
+
+#include <linux/reboot.h>
+
+void (*pm_power_off)(void);
+EXPORT_SYMBOL(pm_power_off);
+
+void machine_power_off(void)
+{
+ local_irq_disable();
+ if (pm_power_off)
+ pm_power_off();
+ asm volatile ("bkpt");
+}
+
+void machine_halt(void)
+{
+ local_irq_disable();
+ if (pm_power_off)
+ pm_power_off();
+ asm volatile ("bkpt");
+}
+
+void machine_restart(char *cmd)
+{
+ local_irq_disable();
+ do_kernel_restart(cmd);
+ asm volatile ("bkpt");
+}
diff --git a/arch/csky/lib/delay.c b/arch/csky/lib/delay.c
new file mode 100644
index 0000000..22570b0
--- /dev/null
+++ b/arch/csky/lib/delay.c
@@ -0,0 +1,39 @@
+// SPDX-License-Identifier: GPL-2.0
+// Copyright (C) 2018 Hangzhou C-SKY Microsystems co.,ltd.
+#include <linux/kernel.h>
+#include <linux/module.h>
+#include <linux/init.h>
+#include <linux/delay.h>
+
+void __delay(unsigned long loops)
+{
+ asm volatile (
+ "mov r0, r0\n"
+ "1:declt %0\n"
+ "bf 1b"
+ : "=r"(loops)
+ : "0"(loops));
+}
+EXPORT_SYMBOL(__delay);
+
+void __const_udelay(unsigned long xloops)
+{
+ unsigned long long loops;
+
+ loops = (unsigned long long)xloops * loops_per_jiffy * HZ;
+
+ __delay(loops >> 32);
+}
+EXPORT_SYMBOL(__const_udelay);
+
+void __udelay(unsigned long usecs)
+{
+ __const_udelay(usecs * 0x10C7UL); /* 2**32 / 1000000 (rounded up) */
+}
+EXPORT_SYMBOL(__udelay);
+
+void __ndelay(unsigned long nsecs)
+{
+ __const_udelay(nsecs * 0x5UL); /* 2**32 / 1000000000 (rounded up) */
+}
+EXPORT_SYMBOL(__ndelay);
--
2.7.4
^ permalink raw reply related [flat|nested] 68+ messages in thread* Re: [PATCH V9 13/21] csky: Library functions
2018-10-16 2:58 ` [PATCH V9 13/21] csky: Library functions Guo Ren
@ 2018-10-17 15:24 ` Arnd Bergmann
2018-10-18 3:10 ` Guo Ren
0 siblings, 1 reply; 68+ messages in thread
From: Arnd Bergmann @ 2018-10-17 15:24 UTC (permalink / raw)
To: Guo Ren
Cc: Andrew Morton, Daniel Lezcano, David Miller, gregkh,
Christoph Hellwig, Marc Zyngier, Mark Rutland, Peter Zijlstra,
Rob Herring, Thomas Gleixner, Linux Kernel Mailing List,
linux-arch, DTML, Rob Herring, c-sky_gcc_upstream, linfei_chen,
Mao Han
On Tue, Oct 16, 2018 at 5:02 AM Guo Ren <ren_guo@c-sky.com> wrote:
> +++ b/arch/csky/lib/delay.c
> @@ -0,0 +1,39 @@
> +// SPDX-License-Identifier: GPL-2.0
> +// Copyright (C) 2018 Hangzhou C-SKY Microsystems co.,ltd.
> +#include <linux/kernel.h>
> +#include <linux/module.h>
> +#include <linux/init.h>
> +#include <linux/delay.h>
> +
> +void __delay(unsigned long loops)
> +{
> + asm volatile (
> + "mov r0, r0\n"
> + "1:declt %0\n"
> + "bf 1b"
> + : "=r"(loops)
> + : "0"(loops));
> +}
> +EXPORT_SYMBOL(__delay);
> +
> +void __const_udelay(unsigned long xloops)
> +{
> + unsigned long long loops;
> +
> + loops = (unsigned long long)xloops * loops_per_jiffy * HZ;
> +
> + __delay(loops >> 32);
> +}
> +EXPORT_SYMBOL(__const_udelay);
I can't find any indication we discussed this part before. Can you
explain (in the
changelog and in a reply) if there is a reliable high-resolution clocksource
available in the architecture? If you can read the elapsed time here instead of
using the loops_per_jiffy estimate, that would make this more robust, as
well as speed up the boot.
If you can't do that in general, that's fine, and you can add my
Reviewed-by: Arnd Bergmann <arnd@arndb.de>
^ permalink raw reply [flat|nested] 68+ messages in thread* Re: [PATCH V9 13/21] csky: Library functions
2018-10-17 15:24 ` Arnd Bergmann
@ 2018-10-18 3:10 ` Guo Ren
0 siblings, 0 replies; 68+ messages in thread
From: Guo Ren @ 2018-10-18 3:10 UTC (permalink / raw)
To: Arnd Bergmann
Cc: Andrew Morton, Daniel Lezcano, David Miller, gregkh,
Christoph Hellwig, Marc Zyngier, Mark Rutland, Peter Zijlstra,
Rob Herring, Thomas Gleixner, Linux Kernel Mailing List,
linux-arch, DTML, Rob Herring, c-sky_gcc_upstream, linfei_chen,
Mao Han
On Wed, Oct 17, 2018 at 05:24:36PM +0200, Arnd Bergmann wrote:
> On Tue, Oct 16, 2018 at 5:02 AM Guo Ren <ren_guo@c-sky.com> wrote:
>
> > +++ b/arch/csky/lib/delay.c
> > @@ -0,0 +1,39 @@
> > +// SPDX-License-Identifier: GPL-2.0
> > +// Copyright (C) 2018 Hangzhou C-SKY Microsystems co.,ltd.
> > +#include <linux/kernel.h>
> > +#include <linux/module.h>
> > +#include <linux/init.h>
> > +#include <linux/delay.h>
> > +
> > +void __delay(unsigned long loops)
> > +{
> > + asm volatile (
> > + "mov r0, r0\n"
> > + "1:declt %0\n"
> > + "bf 1b"
> > + : "=r"(loops)
> > + : "0"(loops));
> > +}
> > +EXPORT_SYMBOL(__delay);
> > +
> > +void __const_udelay(unsigned long xloops)
> > +{
> > + unsigned long long loops;
> > +
> > + loops = (unsigned long long)xloops * loops_per_jiffy * HZ;
> > +
> > + __delay(loops >> 32);
> > +}
> > +EXPORT_SYMBOL(__const_udelay);
>
> I can't find any indication we discussed this part before. Can you
> explain (in the
> changelog and in a reply) if there is a reliable high-resolution clocksource
> available in the architecture? If you can read the elapsed time here instead of
> using the loops_per_jiffy estimate, that would make this more robust, as
> well as speed up the boot.
>
> If you can't do that in general, that's fine, and you can add my
I'll add TODO list here, __delay should use get_cycles() and we didn't
implement __udelay() __ndelay() ... We'll improve it next.
Best Regards
Guo Ren
^ permalink raw reply [flat|nested] 68+ messages in thread
* [PATCH V9 14/21] csky: User access
2018-10-16 2:58 [PATCH V9 00/21] C-SKY(csky) Linux Kernel Port Guo Ren
` (12 preceding siblings ...)
2018-10-16 2:58 ` [PATCH V9 13/21] csky: Library functions Guo Ren
@ 2018-10-16 2:58 ` Guo Ren
2018-10-17 15:37 ` Arnd Bergmann
2018-10-16 2:58 ` [PATCH V9 15/21] csky: Debug and Ptrace GDB Guo Ren
` (7 subsequent siblings)
21 siblings, 1 reply; 68+ messages in thread
From: Guo Ren @ 2018-10-16 2:58 UTC (permalink / raw)
To: akpm, arnd, daniel.lezcano, davem, gregkh, hch, marc.zyngier,
mark.rutland, peterz, robh, tglx
Cc: linux-kernel, linux-arch, devicetree, robh+dt,
c-sky_gcc_upstream, Guo Ren
The patch adds "user access from kernel" codes.
Signed-off-by: Guo Ren <ren_guo@c-sky.com>
Cc: Arnd Bergmann <arnd@arndb.de>
---
arch/csky/include/asm/uaccess.h | 416 ++++++++++++++++++++++++++++++++++++++++
arch/csky/lib/usercopy.c | 262 +++++++++++++++++++++++++
2 files changed, 678 insertions(+)
create mode 100644 arch/csky/include/asm/uaccess.h
create mode 100644 arch/csky/lib/usercopy.c
diff --git a/arch/csky/include/asm/uaccess.h b/arch/csky/include/asm/uaccess.h
new file mode 100644
index 0000000..acaf0e210
--- /dev/null
+++ b/arch/csky/include/asm/uaccess.h
@@ -0,0 +1,416 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+// Copyright (C) 2018 Hangzhou C-SKY Microsystems co.,ltd.
+
+#ifndef __ASM_CSKY_UACCESS_H
+#define __ASM_CSKY_UACCESS_H
+
+/*
+ * User space memory access functions
+ */
+#include <linux/compiler.h>
+#include <linux/errno.h>
+#include <linux/types.h>
+#include <linux/sched.h>
+#include <linux/mm.h>
+#include <linux/string.h>
+#include <linux/version.h>
+#include <asm/segment.h>
+
+#define VERIFY_READ 0
+#define VERIFY_WRITE 1
+
+static inline int access_ok(int type, const void *addr, unsigned long size)
+{
+ unsigned long limit = current_thread_info()->addr_limit.seg;
+
+ return (((unsigned long)addr < limit) &&
+ ((unsigned long)(addr + size) < limit));
+}
+
+static inline int verify_area(int type, const void *addr, unsigned long size)
+{
+ return access_ok(type, addr, size) ? 0 : -EFAULT;
+}
+
+#define __addr_ok(addr) (access_ok(VERIFY_READ, addr, 0))
+
+extern int __put_user_bad(void);
+
+/*
+ * Tell gcc we read from memory instead of writing: this is because
+ * we do not write to any memory gcc knows about, so there are no
+ * aliasing issues.
+ */
+
+/*
+ * These are the main single-value transfer routines. They automatically
+ * use the right size if we just have the right pointer type.
+ *
+ * This gets kind of ugly. We want to return _two_ values in "get_user()"
+ * and yet we don't want to do any pointers, because that is too much
+ * of a performance impact. Thus we have a few rather ugly macros here,
+ * and hide all the ugliness from the user.
+ *
+ * The "__xxx" versions of the user access functions are versions that
+ * do not verify the address space, that must have been done previously
+ * with a separate "access_ok()" call (this is used when we do multiple
+ * accesses to the same area of user memory).
+ *
+ * As we use the same address space for kernel and user data on
+ * Ckcore, we can just do these as direct assignments. (Of course, the
+ * exception handling means that it's no longer "just"...)
+ */
+
+#define put_user(x, ptr) \
+ __put_user_check((x), (ptr), sizeof(*(ptr)))
+
+#define __put_user(x, ptr) \
+ __put_user_nocheck((x), (ptr), sizeof(*(ptr)))
+
+#define __ptr(x) ((unsigned long *)(x))
+
+#define get_user(x, ptr) \
+ __get_user_check((x), (ptr), sizeof(*(ptr)))
+
+#define __get_user(x, ptr) \
+ __get_user_nocheck((x), (ptr), sizeof(*(ptr)))
+
+#define __put_user_nocheck(x, ptr, size) \
+({ \
+ long __pu_err = 0; \
+ typeof(*(ptr)) *__pu_addr = (ptr); \
+ typeof(*(ptr)) __pu_val = (typeof(*(ptr)))(x); \
+ if (__pu_addr) \
+ __put_user_size(__pu_val, (__pu_addr), (size), \
+ __pu_err); \
+ __pu_err; \
+})
+
+#define __put_user_check(x, ptr, size) \
+({ \
+ long __pu_err = -EFAULT; \
+ typeof(*(ptr)) *__pu_addr = (ptr); \
+ typeof(*(ptr)) __pu_val = (typeof(*(ptr)))(x); \
+ if (access_ok(VERIFY_WRITE, __pu_addr, size) && __pu_addr) \
+ __put_user_size(__pu_val, __pu_addr, (size), __pu_err); \
+ __pu_err; \
+})
+
+#define __put_user_size(x, ptr, size, retval) \
+do { \
+ retval = 0; \
+ switch (size) { \
+ case 1: \
+ __put_user_asm_b(x, ptr, retval); \
+ break; \
+ case 2: \
+ __put_user_asm_h(x, ptr, retval); \
+ break; \
+ case 4: \
+ __put_user_asm_w(x, ptr, retval); \
+ break; \
+ case 8: \
+ __put_user_asm_64(x, ptr, retval); \
+ break; \
+ default: \
+ __put_user_bad(); \
+ } \
+} while (0)
+
+/*
+ * We don't tell gcc that we are accessing memory, but this is OK
+ * because we do not write to any memory gcc knows about, so there
+ * are no aliasing issues.
+ *
+ * Note that PC at a fault is the address *after* the faulting
+ * instruction.
+ */
+#define __put_user_asm_b(x, ptr, err) \
+do { \
+ int errcode; \
+ asm volatile( \
+ "1: stb %1, (%2,0) \n" \
+ " br 3f \n" \
+ "2: mov %0, %3 \n" \
+ " br 3f \n" \
+ ".section __ex_table, \"a\" \n" \
+ ".align 2 \n" \
+ ".long 1b,2b \n" \
+ ".previous \n" \
+ "3: \n" \
+ : "=r"(err), "=r"(x), "=r"(ptr), "=r"(errcode) \
+ : "0"(err), "1"(x), "2"(ptr), "3"(-EFAULT) \
+ : "memory"); \
+} while (0)
+
+#define __put_user_asm_h(x, ptr, err) \
+do { \
+ int errcode; \
+ asm volatile( \
+ "1: sth %1, (%2,0) \n" \
+ " br 3f \n" \
+ "2: mov %0, %3 \n" \
+ " br 3f \n" \
+ ".section __ex_table, \"a\" \n" \
+ ".align 2 \n" \
+ ".long 1b,2b \n" \
+ ".previous \n" \
+ "3: \n" \
+ : "=r"(err), "=r"(x), "=r"(ptr), "=r"(errcode) \
+ : "0"(err), "1"(x), "2"(ptr), "3"(-EFAULT) \
+ : "memory"); \
+} while (0)
+
+#define __put_user_asm_w(x, ptr, err) \
+do { \
+ int errcode; \
+ asm volatile( \
+ "1: stw %1, (%2,0) \n" \
+ " br 3f \n" \
+ "2: mov %0, %3 \n" \
+ " br 3f \n" \
+ ".section __ex_table,\"a\" \n" \
+ ".align 2 \n" \
+ ".long 1b, 2b \n" \
+ ".previous \n" \
+ "3: \n" \
+ : "=r"(err), "=r"(x), "=r"(ptr), "=r"(errcode) \
+ : "0"(err), "1"(x), "2"(ptr), "3"(-EFAULT) \
+ : "memory"); \
+} while (0)
+
+#define __put_user_asm_64(x, ptr, err) \
+do { \
+ int tmp; \
+ int errcode; \
+ typeof(*(ptr))src = (typeof(*(ptr)))x; \
+ typeof(*(ptr))*psrc = &src; \
+ \
+ asm volatile( \
+ " ldw %3, (%1, 0) \n" \
+ "1: stw %3, (%2, 0) \n" \
+ " ldw %3, (%1, 4) \n" \
+ "2: stw %3, (%2, 4) \n" \
+ " br 4f \n" \
+ "3: mov %0, %4 \n" \
+ " br 4f \n" \
+ ".section __ex_table, \"a\" \n" \
+ ".align 2 \n" \
+ ".long 1b, 3b \n" \
+ ".long 2b, 3b \n" \
+ ".previous \n" \
+ "4: \n" \
+ : "=r"(err), "=r"(psrc), "=r"(ptr), \
+ "=r"(tmp), "=r"(errcode) \
+ : "0"(err), "1"(psrc), "2"(ptr), "3"(0), "4"(-EFAULT) \
+ : "memory"); \
+} while (0)
+
+#define __get_user_nocheck(x, ptr, size) \
+({ \
+ long __gu_err; \
+ __get_user_size(x, (ptr), (size), __gu_err); \
+ __gu_err; \
+})
+
+#define __get_user_check(x, ptr, size) \
+({ \
+ int __gu_err = -EFAULT; \
+ const __typeof__(*(ptr)) __user *__gu_ptr = (ptr); \
+ if (access_ok(VERIFY_READ, __gu_ptr, size) && __gu_ptr) \
+ __get_user_size(x, __gu_ptr, size, __gu_err); \
+ __gu_err; \
+})
+
+#define __get_user_size(x, ptr, size, retval) \
+do { \
+ switch (size) { \
+ case 1: \
+ __get_user_asm_common((x), ptr, "ldb", retval); \
+ break; \
+ case 2: \
+ __get_user_asm_common((x), ptr, "ldh", retval); \
+ break; \
+ case 4: \
+ __get_user_asm_common((x), ptr, "ldw", retval); \
+ break; \
+ default: \
+ x = 0; \
+ (retval) = __get_user_bad(); \
+ } \
+} while (0)
+
+#define __get_user_asm_common(x, ptr, ins, err) \
+do { \
+ int errcode; \
+ asm volatile( \
+ "1: " ins " %1, (%4,0) \n" \
+ " br 3f \n" \
+ /* Fix up codes */ \
+ "2: mov %0, %2 \n" \
+ " movi %1, 0 \n" \
+ " br 3f \n" \
+ ".section __ex_table,\"a\" \n" \
+ ".align 2 \n" \
+ ".long 1b, 2b \n" \
+ ".previous \n" \
+ "3: \n" \
+ : "=r"(err), "=r"(x), "=r"(errcode) \
+ : "0"(0), "r"(ptr), "2"(-EFAULT) \
+ : "memory"); \
+} while (0)
+
+extern int __get_user_bad(void);
+
+#define __copy_user(to, from, n) \
+do { \
+ int w0, w1, w2, w3; \
+ asm volatile( \
+ "0: cmpnei %1, 0 \n" \
+ " bf 8f \n" \
+ " mov %3, %1 \n" \
+ " or %3, %2 \n" \
+ " andi %3, 3 \n" \
+ " cmpnei %3, 0 \n" \
+ " bf 1f \n" \
+ " br 5f \n" \
+ "1: cmplti %0, 16 \n" /* 4W */ \
+ " bt 3f \n" \
+ " ldw %3, (%2, 0) \n" \
+ " ldw %4, (%2, 4) \n" \
+ " ldw %5, (%2, 8) \n" \
+ " ldw %6, (%2, 12) \n" \
+ "2: stw %3, (%1, 0) \n" \
+ "9: stw %4, (%1, 4) \n" \
+ "10: stw %5, (%1, 8) \n" \
+ "11: stw %6, (%1, 12) \n" \
+ " addi %2, 16 \n" \
+ " addi %1, 16 \n" \
+ " subi %0, 16 \n" \
+ " br 1b \n" \
+ "3: cmplti %0, 4 \n" /* 1W */ \
+ " bt 5f \n" \
+ " ldw %3, (%2, 0) \n" \
+ "4: stw %3, (%1, 0) \n" \
+ " addi %2, 4 \n" \
+ " addi %1, 4 \n" \
+ " subi %0, 4 \n" \
+ " br 3b \n" \
+ "5: cmpnei %0, 0 \n" /* 1B */ \
+ " bf 8f \n" \
+ " ldb %3, (%2, 0) \n" \
+ "6: stb %3, (%1, 0) \n" \
+ " addi %2, 1 \n" \
+ " addi %1, 1 \n" \
+ " subi %0, 1 \n" \
+ " br 5b \n" \
+ "7: br 8f \n" \
+ ".section __ex_table, \"a\" \n" \
+ ".align 2 \n" \
+ ".long 2b, 7b \n" \
+ ".long 9b, 7b \n" \
+ ".long 10b, 7b \n" \
+ ".long 11b, 7b \n" \
+ ".long 4b, 7b \n" \
+ ".long 6b, 7b \n" \
+ ".previous \n" \
+ "8: \n" \
+ : "=r"(n), "=r"(to), "=r"(from), "=r"(w0), \
+ "=r"(w1), "=r"(w2), "=r"(w3) \
+ : "0"(n), "1"(to), "2"(from) \
+ : "memory"); \
+} while (0)
+
+#define __copy_user_zeroing(to, from, n) \
+do { \
+ int tmp; \
+ int nsave; \
+ asm volatile( \
+ "0: cmpnei %1, 0 \n" \
+ " bf 7f \n" \
+ " mov %3, %1 \n" \
+ " or %3, %2 \n" \
+ " andi %3, 3 \n" \
+ " cmpnei %3, 0 \n" \
+ " bf 1f \n" \
+ " br 5f \n" \
+ "1: cmplti %0, 16 \n" \
+ " bt 3f \n" \
+ "2: ldw %3, (%2, 0) \n" \
+ "10: ldw %4, (%2, 4) \n" \
+ " stw %3, (%1, 0) \n" \
+ " stw %4, (%1, 4) \n" \
+ "11: ldw %3, (%2, 8) \n" \
+ "12: ldw %4, (%2, 12) \n" \
+ " stw %3, (%1, 8) \n" \
+ " stw %4, (%1, 12) \n" \
+ " addi %2, 16 \n" \
+ " addi %1, 16 \n" \
+ " subi %0, 16 \n" \
+ " br 1b \n" \
+ "3: cmplti %0, 4 \n" \
+ " bt 5f \n" \
+ "4: ldw %3, (%2, 0) \n" \
+ " stw %3, (%1, 0) \n" \
+ " addi %2, 4 \n" \
+ " addi %1, 4 \n" \
+ " subi %0, 4 \n" \
+ " br 3b \n" \
+ "5: cmpnei %0, 0 \n" \
+ " bf 7f \n" \
+ "6: ldb %3, (%2, 0) \n" \
+ " stb %3, (%1, 0) \n" \
+ " addi %2, 1 \n" \
+ " addi %1, 1 \n" \
+ " subi %0, 1 \n" \
+ " br 5b \n" \
+ "8: mov %3, %0 \n" \
+ " movi %4, 0 \n" \
+ "9: stb %4, (%1, 0) \n" \
+ " addi %1, 1 \n" \
+ " subi %3, 1 \n" \
+ " cmpnei %3, 0 \n" \
+ " bt 9b \n" \
+ " br 7f \n" \
+ ".section __ex_table, \"a\" \n" \
+ ".align 2 \n" \
+ ".long 2b, 8b \n" \
+ ".long 10b, 8b \n" \
+ ".long 11b, 8b \n" \
+ ".long 12b, 8b \n" \
+ ".long 4b, 8b \n" \
+ ".long 6b, 8b \n" \
+ ".previous \n" \
+ "7: \n" \
+ : "=r"(n), "=r"(to), "=r"(from), "=r"(nsave), \
+ "=r"(tmp) \
+ : "0"(n), "1"(to), "2"(from) \
+ : "memory"); \
+} while (0)
+
+unsigned long raw_copy_from_user(void *to, const void *from, unsigned long n);
+unsigned long raw_copy_to_user(void *to, const void *from, unsigned long n);
+
+unsigned long clear_user(void *to, unsigned long n);
+unsigned long __clear_user(void __user *to, unsigned long n);
+
+long strncpy_from_user(char *dst, const char *src, long count);
+long __strncpy_from_user(char *dst, const char *src, long count);
+
+/*
+ * Return the size of a string (including the ending 0)
+ *
+ * Return 0 on exception, a value greater than N if too long
+ */
+long strnlen_user(const char *src, long n);
+
+#define strlen_user(str) strnlen_user(str, 32767)
+
+struct exception_table_entry {
+ unsigned long insn;
+ unsigned long nextinsn;
+};
+
+extern int fixup_exception(struct pt_regs *regs);
+
+#endif /* __ASM_CSKY_UACCESS_H */
diff --git a/arch/csky/lib/usercopy.c b/arch/csky/lib/usercopy.c
new file mode 100644
index 0000000..ac9170e
--- /dev/null
+++ b/arch/csky/lib/usercopy.c
@@ -0,0 +1,262 @@
+// SPDX-License-Identifier: GPL-2.0
+// Copyright (C) 2018 Hangzhou C-SKY Microsystems co.,ltd.
+
+#include <linux/uaccess.h>
+#include <linux/types.h>
+
+unsigned long raw_copy_from_user(void *to, const void *from,
+ unsigned long n)
+{
+ if (access_ok(VERIFY_READ, from, n))
+ __copy_user_zeroing(to, from, n);
+ else
+ memset(to, 0, n);
+ return n;
+}
+EXPORT_SYMBOL(raw_copy_from_user);
+
+unsigned long raw_copy_to_user(void *to, const void *from,
+ unsigned long n)
+{
+ if (access_ok(VERIFY_WRITE, to, n))
+ __copy_user(to, from, n);
+ return n;
+}
+EXPORT_SYMBOL(raw_copy_to_user);
+
+
+/*
+ * copy a null terminated string from userspace.
+ */
+#define __do_strncpy_from_user(dst, src, count, res) \
+do { \
+ int tmp; \
+ long faultres; \
+ asm volatile( \
+ " cmpnei %3, 0 \n" \
+ " bf 4f \n" \
+ "1: cmpnei %1, 0 \n" \
+ " bf 5f \n" \
+ "2: ldb %4, (%3, 0) \n" \
+ " stb %4, (%2, 0) \n" \
+ " cmpnei %4, 0 \n" \
+ " bf 3f \n" \
+ " addi %3, 1 \n" \
+ " addi %2, 1 \n" \
+ " subi %1, 1 \n" \
+ " br 1b \n" \
+ "3: subu %0, %1 \n" \
+ " br 5f \n" \
+ "4: mov %0, %5 \n" \
+ " br 5f \n" \
+ ".section __ex_table, \"a\" \n" \
+ ".align 2 \n" \
+ ".long 2b, 4b \n" \
+ ".previous \n" \
+ "5: \n" \
+ : "=r"(res), "=r"(count), "=r"(dst), \
+ "=r"(src), "=r"(tmp), "=r"(faultres) \
+ : "5"(-EFAULT), "0"(count), "1"(count), \
+ "2"(dst), "3"(src) \
+ : "memory", "cc"); \
+} while (0)
+
+/*
+ * __strncpy_from_user: - Copy a NUL terminated string from userspace,
+ * with less checking.
+ * @dst: Destination address, in kernel space. This buffer must be at
+ * least @count bytes long.
+ * @src: Source address, in user space.
+ * @count: Maximum number of bytes to copy, including the trailing NUL.
+ *
+ * Copies a NUL-terminated string from userspace to kernel space.
+ * Caller must check the specified block with access_ok() before calling
+ * this function.
+ *
+ * On success, returns the length of the string (not including the trailing
+ * NUL).
+ *
+ * If access to userspace fails, returns -EFAULT (some data may have been
+ * copied).
+ *
+ * If @count is smaller than the length of the string, copies @count bytes
+ * and returns @count.
+ */
+long __strncpy_from_user(char *dst, const char *src, long count)
+{
+ long res;
+
+ __do_strncpy_from_user(dst, src, count, res);
+ return res;
+}
+EXPORT_SYMBOL(__strncpy_from_user);
+
+/*
+ * strncpy_from_user: - Copy a NUL terminated string from userspace.
+ * @dst: Destination address, in kernel space. This buffer must be at
+ * least @count bytes long.
+ * @src: Source address, in user space.
+ * @count: Maximum number of bytes to copy, including the trailing NUL.
+ *
+ * Copies a NUL-terminated string from userspace to kernel space.
+ *
+ * On success, returns the length of the string (not including the trailing
+ * NUL).
+ *
+ * If access to userspace fails, returns -EFAULT (some data may have been
+ * copied).
+ *
+ * If @count is smaller than the length of the string, copies @count bytes
+ * and returns @count.
+ */
+long strncpy_from_user(char *dst, const char *src, long count)
+{
+ long res = -EFAULT;
+
+ if (access_ok(VERIFY_READ, src, 1))
+ __do_strncpy_from_user(dst, src, count, res);
+ return res;
+}
+EXPORT_SYMBOL(strncpy_from_user);
+
+/*
+ * strlen_user: - Get the size of a string in user space.
+ * @str: The string to measure.
+ * @n: The maximum valid length
+ *
+ * Get the size of a NUL-terminated string in user space.
+ *
+ * Returns the size of the string INCLUDING the terminating NUL.
+ * On exception, returns 0.
+ * If the string is too long, returns a value greater than @n.
+ */
+long strnlen_user(const char *s, long n)
+{
+ unsigned long res, tmp;
+
+ if (s == NULL)
+ return 0;
+
+ asm volatile(
+ " cmpnei %1, 0 \n"
+ " bf 3f \n"
+ "1: cmpnei %0, 0 \n"
+ " bf 3f \n"
+ "2: ldb %3, (%1, 0) \n"
+ " cmpnei %3, 0 \n"
+ " bf 3f \n"
+ " subi %0, 1 \n"
+ " addi %1, 1 \n"
+ " br 1b \n"
+ "3: subu %2, %0 \n"
+ " addi %2, 1 \n"
+ " br 5f \n"
+ "4: movi %0, 0 \n"
+ " br 5f \n"
+ ".section __ex_table, \"a\" \n"
+ ".align 2 \n"
+ ".long 2b, 4b \n"
+ ".previous \n"
+ "5: \n"
+ : "=r"(n), "=r"(s), "=r"(res), "=r"(tmp)
+ : "0"(n), "1"(s), "2"(n)
+ : "memory", "cc");
+
+ return res;
+}
+EXPORT_SYMBOL(strnlen_user);
+
+#define __do_clear_user(addr, size) \
+do { \
+ int __d0, zvalue, tmp; \
+ \
+ asm volatile( \
+ "0: cmpnei %1, 0 \n" \
+ " bf 7f \n" \
+ " mov %3, %1 \n" \
+ " andi %3, 3 \n" \
+ " cmpnei %3, 0 \n" \
+ " bf 1f \n" \
+ " br 5f \n" \
+ "1: cmplti %0, 32 \n" /* 4W */ \
+ " bt 3f \n" \
+ "8: stw %2, (%1, 0) \n" \
+ "10: stw %2, (%1, 4) \n" \
+ "11: stw %2, (%1, 8) \n" \
+ "12: stw %2, (%1, 12) \n" \
+ "13: stw %2, (%1, 16) \n" \
+ "14: stw %2, (%1, 20) \n" \
+ "15: stw %2, (%1, 24) \n" \
+ "16: stw %2, (%1, 28) \n" \
+ " addi %1, 32 \n" \
+ " subi %0, 32 \n" \
+ " br 1b \n" \
+ "3: cmplti %0, 4 \n" /* 1W */ \
+ " bt 5f \n" \
+ "4: stw %2, (%1, 0) \n" \
+ " addi %1, 4 \n" \
+ " subi %0, 4 \n" \
+ " br 3b \n" \
+ "5: cmpnei %0, 0 \n" /* 1B */ \
+ "9: bf 7f \n" \
+ "6: stb %2, (%1, 0) \n" \
+ " addi %1, 1 \n" \
+ " subi %0, 1 \n" \
+ " br 5b \n" \
+ ".section __ex_table,\"a\" \n" \
+ ".align 2 \n" \
+ ".long 8b, 9b \n" \
+ ".long 10b, 9b \n" \
+ ".long 11b, 9b \n" \
+ ".long 12b, 9b \n" \
+ ".long 13b, 9b \n" \
+ ".long 14b, 9b \n" \
+ ".long 15b, 9b \n" \
+ ".long 16b, 9b \n" \
+ ".long 4b, 9b \n" \
+ ".long 6b, 9b \n" \
+ ".previous \n" \
+ "7: \n" \
+ : "=r"(size), "=r" (__d0), \
+ "=r"(zvalue), "=r"(tmp) \
+ : "0"(size), "1"(addr), "2"(0) \
+ : "memory", "cc"); \
+} while (0)
+
+/*
+ * clear_user: - Zero a block of memory in user space.
+ * @to: Destination address, in user space.
+ * @n: Number of bytes to zero.
+ *
+ * Zero a block of memory in user space.
+ *
+ * Returns number of bytes that could not be cleared.
+ * On success, this will be zero.
+ */
+unsigned long
+clear_user(void __user *to, unsigned long n)
+{
+ if (access_ok(VERIFY_WRITE, to, n))
+ __do_clear_user(to, n);
+ return n;
+}
+EXPORT_SYMBOL(clear_user);
+
+/*
+ * __clear_user: - Zero a block of memory in user space, with less checking.
+ * @to: Destination address, in user space.
+ * @n: Number of bytes to zero.
+ *
+ * Zero a block of memory in user space. Caller must check
+ * the specified block with access_ok() before calling this function.
+ *
+ * Returns number of bytes that could not be cleared.
+ * On success, this will be zero.
+ */
+unsigned long
+__clear_user(void __user *to, unsigned long n)
+{
+ __do_clear_user(to, n);
+ return n;
+}
+EXPORT_SYMBOL(__clear_user);
--
2.7.4
^ permalink raw reply related [flat|nested] 68+ messages in thread* Re: [PATCH V9 14/21] csky: User access
2018-10-16 2:58 ` [PATCH V9 14/21] csky: User access Guo Ren
@ 2018-10-17 15:37 ` Arnd Bergmann
0 siblings, 0 replies; 68+ messages in thread
From: Arnd Bergmann @ 2018-10-17 15:37 UTC (permalink / raw)
To: Guo Ren
Cc: Andrew Morton, Daniel Lezcano, David Miller, gregkh,
Christoph Hellwig, Marc Zyngier, Mark Rutland, Peter Zijlstra,
Rob Herring, Thomas Gleixner, Linux Kernel Mailing List,
linux-arch, DTML, Rob Herring, c-sky_gcc_upstream
On Tue, Oct 16, 2018 at 5:05 AM Guo Ren <ren_guo@c-sky.com> wrote:
>
> The patch adds "user access from kernel" codes.
>
> Signed-off-by: Guo Ren <ren_guo@c-sky.com>
> Cc: Arnd Bergmann <arnd@arndb.de>
> ---
> arch/csky/include/asm/uaccess.h | 416 ++++++++++++++++++++++++++++++++++++++++
> arch/csky/lib/usercopy.c | 262 +++++++++++++++++++++++++
> 2 files changed, 678 insertions(+)
> create mode 100644 arch/csky/include/asm/uaccess.h
> create mode 100644 arch/csky/lib/usercopy.c
Reviewed-by: Arnd Bergmann <arnd@arndb.de>
The 64-bit __put_user() is usually what everyone gets wrong, but
your version seems fine. Also the access_ok() checks and
the copy_to_user() result value all look correct.
Arnd
^ permalink raw reply [flat|nested] 68+ messages in thread
* [PATCH V9 15/21] csky: Debug and Ptrace GDB
2018-10-16 2:58 [PATCH V9 00/21] C-SKY(csky) Linux Kernel Port Guo Ren
` (13 preceding siblings ...)
2018-10-16 2:58 ` [PATCH V9 14/21] csky: User access Guo Ren
@ 2018-10-16 2:58 ` Guo Ren
2018-10-17 15:46 ` Arnd Bergmann
2018-10-16 2:58 ` [PATCH V9 16/21] csky: SMP support Guo Ren
` (6 subsequent siblings)
21 siblings, 1 reply; 68+ messages in thread
From: Guo Ren @ 2018-10-16 2:58 UTC (permalink / raw)
To: akpm, arnd, daniel.lezcano, davem, gregkh, hch, marc.zyngier,
mark.rutland, peterz, robh, tglx
Cc: linux-kernel, linux-arch, devicetree, robh+dt,
c-sky_gcc_upstream, Guo Ren
This patch adds arch ptrace implementation, stack dump and bug.h.
Signed-off-by: Guo Ren <ren_guo@c-sky.com>
Cc: Arnd Bergmann <arnd@arndb.de>
---
arch/csky/include/asm/bug.h | 26 +++
arch/csky/include/uapi/asm/ptrace.h | 104 ++++++++++++
arch/csky/kernel/dumpstack.c | 66 ++++++++
arch/csky/kernel/ptrace.c | 314 ++++++++++++++++++++++++++++++++++++
4 files changed, 510 insertions(+)
create mode 100644 arch/csky/include/asm/bug.h
create mode 100644 arch/csky/include/uapi/asm/ptrace.h
create mode 100644 arch/csky/kernel/dumpstack.c
create mode 100644 arch/csky/kernel/ptrace.c
diff --git a/arch/csky/include/asm/bug.h b/arch/csky/include/asm/bug.h
new file mode 100644
index 0000000..bd7b323
--- /dev/null
+++ b/arch/csky/include/asm/bug.h
@@ -0,0 +1,26 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+// Copyright (C) 2018 Hangzhou C-SKY Microsystems co.,ltd.
+
+#ifndef __ASM_CSKY_BUG_H
+#define __ASM_CSKY_BUG_H
+
+#include <linux/compiler.h>
+#include <linux/const.h>
+#include <linux/types.h>
+
+#define BUG() \
+do { \
+ asm volatile ("bkpt\n"); \
+ unreachable(); \
+} while (0)
+
+#define HAVE_ARCH_BUG
+
+#include <asm-generic/bug.h>
+
+struct pt_regs;
+
+void die_if_kernel(char *str, struct pt_regs *regs, int nr);
+void show_regs(struct pt_regs *regs);
+
+#endif /* __ASM_CSKY_BUG_H */
diff --git a/arch/csky/include/uapi/asm/ptrace.h b/arch/csky/include/uapi/asm/ptrace.h
new file mode 100644
index 0000000..f10d02c
--- /dev/null
+++ b/arch/csky/include/uapi/asm/ptrace.h
@@ -0,0 +1,104 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+// Copyright (C) 2018 Hangzhou C-SKY Microsystems co.,ltd.
+
+#ifndef _CSKY_PTRACE_H
+#define _CSKY_PTRACE_H
+
+#ifndef __ASSEMBLY__
+
+struct pt_regs {
+ unsigned long tls;
+ unsigned long lr;
+ unsigned long pc;
+ unsigned long sr;
+ unsigned long usp;
+
+ /*
+ * a0, a1, a2, a3:
+ * abiv1: r2, r3, r4, r5
+ * abiv2: r0, r1, r2, r3
+ */
+ unsigned long orig_a0;
+ unsigned long a0;
+ unsigned long a1;
+ unsigned long a2;
+ unsigned long a3;
+
+ /*
+ * ABIV2: r4 ~ r13
+ * ABIV1: r6 ~ r14, r1
+ */
+ unsigned long regs[10];
+
+#if defined(__CSKYABIV2__)
+ /* r16 ~ r30 */
+ unsigned long exregs[15];
+
+ unsigned long rhi;
+ unsigned long rlo;
+ unsigned long pad; /* reserved */
+#endif
+};
+
+struct user_fp {
+ unsigned long vr[96];
+ unsigned long fcr;
+ unsigned long fesr;
+ unsigned long fid;
+ unsigned long reserved;
+};
+
+/*
+ * Switch stack for switch_to after push pt_regs.
+ *
+ * ABI_CSKYV2: r4 ~ r11, r15 ~ r17, r26 ~ r30;
+ * ABI_CSKYV1: r8 ~ r14, r15;
+ */
+struct switch_stack {
+#if defined(__CSKYABIV2__)
+ unsigned long r4;
+ unsigned long r5;
+ unsigned long r6;
+ unsigned long r7;
+ unsigned long r8;
+ unsigned long r9;
+ unsigned long r10;
+ unsigned long r11;
+#else
+ unsigned long r8;
+ unsigned long r9;
+ unsigned long r10;
+ unsigned long r11;
+ unsigned long r12;
+ unsigned long r13;
+ unsigned long r14;
+#endif
+ unsigned long r15;
+#if defined(__CSKYABIV2__)
+ unsigned long r16;
+ unsigned long r17;
+ unsigned long r26;
+ unsigned long r27;
+ unsigned long r28;
+ unsigned long r29;
+ unsigned long r30;
+#endif
+};
+
+#ifdef __KERNEL__
+
+#define PS_S 0x80000000 /* Supervisor Mode */
+
+#define arch_has_single_step() (1)
+#define current_pt_regs() \
+({ (struct pt_regs *)((char *)current_thread_info() + THREAD_SIZE) - 1; })
+
+#define user_stack_pointer(regs) ((regs)->usp)
+
+#define user_mode(regs) (!((regs)->sr & PS_S))
+#define instruction_pointer(regs) ((regs)->pc)
+#define profile_pc(regs) instruction_pointer(regs)
+
+#endif /* __KERNEL__ */
+#endif /* __ASSEMBLY__ */
+#endif /* _CSKY_PTRACE_H */
diff --git a/arch/csky/kernel/dumpstack.c b/arch/csky/kernel/dumpstack.c
new file mode 100644
index 0000000..a9a03ac
--- /dev/null
+++ b/arch/csky/kernel/dumpstack.c
@@ -0,0 +1,66 @@
+// SPDX-License-Identifier: GPL-2.0
+// Copyright (C) 2018 Hangzhou C-SKY Microsystems co.,ltd.
+
+#include <linux/ptrace.h>
+
+int kstack_depth_to_print = 48;
+
+void show_trace(unsigned long *stack)
+{
+ unsigned long *endstack;
+ unsigned long addr;
+ int i;
+
+ pr_info("Call Trace:\n");
+ addr = (unsigned long)stack + THREAD_SIZE - 1;
+ endstack = (unsigned long *)(addr & -THREAD_SIZE);
+ i = 0;
+ while (stack + 1 <= endstack) {
+ addr = *stack++;
+ /*
+ * If the address is either in the text segment of the
+ * kernel, or in the region which contains vmalloc'ed
+ * memory, it *may* be the address of a calling
+ * routine; if so, print it so that someone tracing
+ * down the cause of the crash will be able to figure
+ * out the call path that was taken.
+ */
+ if (__kernel_text_address(addr)) {
+#ifndef CONFIG_KALLSYMS
+ if (i % 5 == 0)
+ pr_cont("\n ");
+#endif
+ pr_cont(" [<%08lx>] %pS\n", addr, (void *)addr);
+ i++;
+ }
+ }
+ pr_cont("\n");
+}
+
+void show_stack(struct task_struct *task, unsigned long *stack)
+{
+ unsigned long *p;
+ unsigned long *endstack;
+ int i;
+
+ if (!stack) {
+ if (task)
+ stack = (unsigned long *)task->thread.esp0;
+ else
+ stack = (unsigned long *)&stack;
+ }
+ endstack = (unsigned long *)
+ (((unsigned long)stack + THREAD_SIZE - 1) & -THREAD_SIZE);
+
+ pr_info("Stack from %08lx:", (unsigned long)stack);
+ p = stack;
+ for (i = 0; i < kstack_depth_to_print; i++) {
+ if (p + 1 > endstack)
+ break;
+ if (i % 8 == 0)
+ pr_cont("\n ");
+ pr_cont(" %08lx", *p++);
+ }
+ pr_cont("\n");
+ show_trace(stack);
+}
diff --git a/arch/csky/kernel/ptrace.c b/arch/csky/kernel/ptrace.c
new file mode 100644
index 0000000..34b3025
--- /dev/null
+++ b/arch/csky/kernel/ptrace.c
@@ -0,0 +1,314 @@
+// SPDX-License-Identifier: GPL-2.0
+// Copyright (C) 2018 Hangzhou C-SKY Microsystems co.,ltd.
+
+#include <linux/elf.h>
+#include <linux/errno.h>
+#include <linux/kernel.h>
+#include <linux/mm.h>
+#include <linux/ptrace.h>
+#include <linux/regset.h>
+#include <linux/sched.h>
+#include <linux/signal.h>
+#include <linux/smp.h>
+#include <linux/uaccess.h>
+#include <linux/user.h>
+
+#include <asm/thread_info.h>
+#include <asm/page.h>
+#include <asm/pgtable.h>
+#include <asm/processor.h>
+#include <asm/asm-offsets.h>
+
+#include <abi/regdef.h>
+
+/* sets the trace bits. */
+#define TRACE_MODE_SI (1 << 14)
+#define TRACE_MODE_RUN 0
+#define TRACE_MODE_MASK ~(0x3 << 14)
+
+/*
+ * Make sure the single step bit is not set.
+ */
+static void singlestep_disable(struct task_struct *tsk)
+{
+ struct pt_regs *regs;
+
+ regs = task_pt_regs(tsk);
+ regs->sr = (regs->sr & TRACE_MODE_MASK) | TRACE_MODE_RUN;
+}
+
+static void singlestep_enable(struct task_struct *tsk)
+{
+ struct pt_regs *regs;
+
+ regs = task_pt_regs(tsk);
+ regs->sr = (regs->sr & TRACE_MODE_MASK) | TRACE_MODE_SI;
+}
+
+/*
+ * Make sure the single step bit is set.
+ */
+void user_enable_single_step(struct task_struct *child)
+{
+ if (child->thread.esp0 == 0)
+ return;
+ singlestep_enable(child);
+}
+
+void user_disable_single_step(struct task_struct *child)
+{
+ if (child->thread.esp0 == 0)
+ return;
+ singlestep_disable(child);
+}
+
+enum csky_regset {
+ REGSET_GPR,
+ REGSET_FPR,
+};
+
+static int gpr_get(struct task_struct *target,
+ const struct user_regset *regset,
+ unsigned int pos, unsigned int count,
+ void *kbuf, void __user *ubuf)
+{
+ struct pt_regs *regs;
+
+ regs = task_pt_regs(target);
+
+ /* Abiv1 regs->tls is fake and we need sync here. */
+ regs->tls = task_thread_info(target)->tp_value;
+
+ return user_regset_copyout(&pos, &count, &kbuf, &ubuf, regs, 0, -1);
+}
+
+static int gpr_set(struct task_struct *target,
+ const struct user_regset *regset,
+ unsigned int pos, unsigned int count,
+ const void *kbuf, const void __user *ubuf)
+{
+ int ret;
+ struct pt_regs regs;
+
+ ret = user_regset_copyin(&pos, &count, &kbuf, &ubuf, ®s, 0, -1);
+ if (ret)
+ return ret;
+
+ regs.sr = task_pt_regs(target)->sr;
+
+ task_thread_info(target)->tp_value = regs.tls;
+
+ *task_pt_regs(target) = regs;
+
+ return 0;
+}
+
+static int fpr_get(struct task_struct *target,
+ const struct user_regset *regset,
+ unsigned int pos, unsigned int count,
+ void *kbuf, void __user *ubuf)
+{
+ struct user_fp *regs = (struct user_fp *)&target->thread.user_fp;
+
+#if defined(CONFIG_CPU_HAS_FPUV2) && !defined(CONFIG_CPU_HAS_VDSP)
+ int i;
+ struct user_fp tmp = *regs;
+
+ for (i = 0; i < 16; i++) {
+ tmp.vr[i*4] = regs->vr[i*2];
+ tmp.vr[i*4 + 1] = regs->vr[i*2 + 1];
+ }
+
+ for (i = 0; i < 32; i++)
+ tmp.vr[64 + i] = regs->vr[32 + i];
+
+ return user_regset_copyout(&pos, &count, &kbuf, &ubuf, &tmp, 0, -1);
+#else
+ return user_regset_copyout(&pos, &count, &kbuf, &ubuf, regs, 0, -1);
+#endif
+}
+
+static int fpr_set(struct task_struct *target,
+ const struct user_regset *regset,
+ unsigned int pos, unsigned int count,
+ const void *kbuf, const void __user *ubuf)
+{
+ int ret;
+ struct user_fp *regs = (struct user_fp *)&target->thread.user_fp;
+
+#if defined(CONFIG_CPU_HAS_FPUV2) && !defined(CONFIG_CPU_HAS_VDSP)
+ int i;
+ struct user_fp tmp;
+
+ ret = user_regset_copyin(&pos, &count, &kbuf, &ubuf, &tmp, 0, -1);
+
+ *regs = tmp;
+
+ for (i = 0; i < 16; i++) {
+ regs->vr[i*2] = tmp.vr[i*4];
+ regs->vr[i*2 + 1] = tmp.vr[i*4 + 1];
+ }
+
+ for (i = 0; i < 32; i++)
+ regs->vr[32 + i] = tmp.vr[64 + i];
+#else
+ ret = user_regset_copyin(&pos, &count, &kbuf, &ubuf, regs, 0, -1);
+#endif
+
+ return ret;
+}
+
+static const struct user_regset csky_regsets[] = {
+ [REGSET_GPR] = {
+ .core_note_type = NT_PRSTATUS,
+ .n = ELF_NGREG,
+ .size = sizeof(u32),
+ .align = sizeof(u32),
+ .get = &gpr_get,
+ .set = &gpr_set,
+ },
+ [REGSET_FPR] = {
+ .core_note_type = NT_PRFPREG,
+ .n = sizeof(struct user_fp) / sizeof(u32),
+ .size = sizeof(u32),
+ .align = sizeof(u32),
+ .get = &fpr_get,
+ .set = &fpr_set,
+ },
+};
+
+static const struct user_regset_view user_csky_view = {
+ .name = "csky",
+ .e_machine = ELF_ARCH,
+ .regsets = csky_regsets,
+ .n = ARRAY_SIZE(csky_regsets),
+};
+
+const struct user_regset_view *task_user_regset_view(struct task_struct *task)
+{
+ return &user_csky_view;
+}
+
+void ptrace_disable(struct task_struct *child)
+{
+ singlestep_disable(child);
+}
+
+long arch_ptrace(struct task_struct *child, long request,
+ unsigned long addr, unsigned long data)
+{
+ long ret = -EIO;
+
+ switch (request) {
+ default:
+ ret = ptrace_request(child, request, addr, data);
+ break;
+ }
+
+ return ret;
+}
+
+/*
+ * If process's system calls is traces, do some corresponding handles in this
+ * function before entering system call function and after exiting system call
+ * function.
+ */
+asmlinkage void syscall_trace(int why, struct pt_regs *regs)
+{
+ long saved_why;
+ /*
+ * Save saved_why, why is used to denote syscall entry/exit;
+ * why = 0:entry, why = 1: exit
+ */
+ saved_why = regs->regs[SYSTRACE_SAVENUM];
+ regs->regs[SYSTRACE_SAVENUM] = why;
+
+ ptrace_notify(SIGTRAP | ((current->ptrace & PT_TRACESYSGOOD)
+ ? 0x80 : 0));
+
+ /*
+ * this isn't the same as continuing with a signal, but it will do
+ * for normal use. strace only continues with a signal if the
+ * stopping signal is not SIGTRAP. -brl
+ */
+ if (current->exit_code) {
+ send_sig(current->exit_code, current, 1);
+ current->exit_code = 0;
+ }
+
+ regs->regs[SYSTRACE_SAVENUM] = saved_why;
+}
+
+void show_regs(struct pt_regs *fp)
+{
+ unsigned long *sp;
+ unsigned char *tp;
+ int i;
+
+ pr_info("\nCURRENT PROCESS:\n\n");
+ pr_info("COMM=%s PID=%d\n", current->comm, current->pid);
+
+ if (current->mm) {
+ pr_info("TEXT=%08x-%08x DATA=%08x-%08x BSS=%08x-%08x\n",
+ (int) current->mm->start_code,
+ (int) current->mm->end_code,
+ (int) current->mm->start_data,
+ (int) current->mm->end_data,
+ (int) current->mm->end_data,
+ (int) current->mm->brk);
+ pr_info("USER-STACK=%08x KERNEL-STACK=%08x\n\n",
+ (int) current->mm->start_stack,
+ (int) (((unsigned long) current) + 2 * PAGE_SIZE));
+ }
+
+ pr_info("PC: 0x%08lx\n", (long)fp->pc);
+ pr_info("orig_a0: 0x%08lx\n", fp->orig_a0);
+ pr_info("PSR: 0x%08lx\n", (long)fp->sr);
+
+ pr_info("a0: 0x%08lx a1: 0x%08lx a2: 0x%08lx a3: 0x%08lx\n",
+ fp->a0, fp->a1, fp->a2, fp->a3);
+#if defined(__CSKYABIV2__)
+ pr_info("r4: 0x%08lx r5: 0x%08lx r6: 0x%08lx r7: 0x%08lx\n",
+ fp->regs[0], fp->regs[1], fp->regs[2], fp->regs[3]);
+ pr_info("r8: 0x%08lx r9: 0x%08lx r10: 0x%08lx r11: 0x%08lx\n",
+ fp->regs[4], fp->regs[5], fp->regs[6], fp->regs[7]);
+ pr_info("r12 0x%08lx r13: 0x%08lx r15: 0x%08lx\n",
+ fp->regs[8], fp->regs[9], fp->lr);
+ pr_info("r16:0x%08lx r17: 0x%08lx r18: 0x%08lx r19: 0x%08lx\n",
+ fp->exregs[0], fp->exregs[1], fp->exregs[2], fp->exregs[3]);
+ pr_info("r20 0x%08lx r21: 0x%08lx r22: 0x%08lx r23: 0x%08lx\n",
+ fp->exregs[4], fp->exregs[5], fp->exregs[6], fp->exregs[7]);
+ pr_info("r24 0x%08lx r25: 0x%08lx r26: 0x%08lx r27: 0x%08lx\n",
+ fp->exregs[8], fp->exregs[9], fp->exregs[10], fp->exregs[11]);
+ pr_info("r28 0x%08lx r29: 0x%08lx r30: 0x%08lx tls: 0x%08lx\n",
+ fp->exregs[12], fp->exregs[13], fp->exregs[14], fp->tls);
+ pr_info("hi 0x%08lx lo: 0x%08lx\n",
+ fp->rhi, fp->rlo);
+#else
+ pr_info("r6: 0x%08lx r7: 0x%08lx r8: 0x%08lx r9: 0x%08lx\n",
+ fp->regs[0], fp->regs[1], fp->regs[2], fp->regs[3]);
+ pr_info("r10: 0x%08lx r11: 0x%08lx r12: 0x%08lx r13: 0x%08lx\n",
+ fp->regs[4], fp->regs[5], fp->regs[6], fp->regs[7]);
+ pr_info("r14 0x%08lx r1: 0x%08lx r15: 0x%08lx\n",
+ fp->regs[8], fp->regs[9], fp->lr);
+#endif
+
+ pr_info("\nCODE:");
+ tp = ((unsigned char *) fp->pc) - 0x20;
+ tp += ((int)tp % 4) ? 2 : 0;
+ for (sp = (unsigned long *) tp, i = 0; (i < 0x40); i += 4) {
+ if ((i % 0x10) == 0)
+ pr_cont("\n%08x: ", (int) (tp + i));
+ pr_cont("%08x ", (int) *sp++);
+ }
+ pr_cont("\n");
+
+ pr_info("\nKERNEL STACK:");
+ tp = ((unsigned char *) fp) - 0x40;
+ for (sp = (unsigned long *) tp, i = 0; (i < 0xc0); i += 4) {
+ if ((i % 0x10) == 0)
+ pr_cont("\n%08x: ", (int) (tp + i));
+ pr_cont("%08x ", (int) *sp++);
+ }
+ pr_cont("\n");
+}
--
2.7.4
^ permalink raw reply related [flat|nested] 68+ messages in thread* Re: [PATCH V9 15/21] csky: Debug and Ptrace GDB
2018-10-16 2:58 ` [PATCH V9 15/21] csky: Debug and Ptrace GDB Guo Ren
@ 2018-10-17 15:46 ` Arnd Bergmann
2018-10-18 3:17 ` Guo Ren
0 siblings, 1 reply; 68+ messages in thread
From: Arnd Bergmann @ 2018-10-17 15:46 UTC (permalink / raw)
To: Guo Ren
Cc: Andrew Morton, Daniel Lezcano, David Miller, gregkh,
Christoph Hellwig, Marc Zyngier, Mark Rutland, Peter Zijlstra,
Rob Herring, Thomas Gleixner, Linux Kernel Mailing List,
linux-arch, DTML, Rob Herring, c-sky_gcc_upstream
On Tue, Oct 16, 2018 at 5:06 AM Guo Ren <ren_guo@c-sky.com> wrote:
>
> This patch adds arch ptrace implementation, stack dump and bug.h.
>
> Signed-off-by: Guo Ren <ren_guo@c-sky.com>
> Cc: Arnd Bergmann <arnd@arndb.de>
Reviewed-by: Arnd Bergmann <arnd@arndb.de>
> diff --git a/arch/csky/include/asm/bug.h b/arch/csky/include/asm/bug.h
> new file mode 100644
> index 0000000..bd7b323
> --- /dev/null
> +++ b/arch/csky/include/asm/bug.h
> @@ -0,0 +1,26 @@
> +/* SPDX-License-Identifier: GPL-2.0 */
> +// Copyright (C) 2018 Hangzhou C-SKY Microsystems co.,ltd.
> +
> +#ifndef __ASM_CSKY_BUG_H
> +#define __ASM_CSKY_BUG_H
> +
> +#include <linux/compiler.h>
> +#include <linux/const.h>
> +#include <linux/types.h>
> +
> +#define BUG() \
> +do { \
> + asm volatile ("bkpt\n"); \
> + unreachable(); \
> +} while (0)
Since you have a breakpoint instruction, a nice addition might
be to use that for implementing WARN_ON as well, whenever you
get around to doing that. It's certainly not required for the initial
merge.
^ permalink raw reply [flat|nested] 68+ messages in thread* Re: [PATCH V9 15/21] csky: Debug and Ptrace GDB
2018-10-17 15:46 ` Arnd Bergmann
@ 2018-10-18 3:17 ` Guo Ren
0 siblings, 0 replies; 68+ messages in thread
From: Guo Ren @ 2018-10-18 3:17 UTC (permalink / raw)
To: Arnd Bergmann
Cc: Andrew Morton, Daniel Lezcano, David Miller, gregkh,
Christoph Hellwig, Marc Zyngier, Mark Rutland, Peter Zijlstra,
Rob Herring, Thomas Gleixner, Linux Kernel Mailing List,
linux-arch, DTML, Rob Herring, c-sky_gcc_upstream
On Wed, Oct 17, 2018 at 05:46:56PM +0200, Arnd Bergmann wrote:
> On Tue, Oct 16, 2018 at 5:06 AM Guo Ren <ren_guo@c-sky.com> wrote:
> >
> > This patch adds arch ptrace implementation, stack dump and bug.h.
> >
> > Signed-off-by: Guo Ren <ren_guo@c-sky.com>
> > Cc: Arnd Bergmann <arnd@arndb.de>
>
> Reviewed-by: Arnd Bergmann <arnd@arndb.de>
>
> > diff --git a/arch/csky/include/asm/bug.h b/arch/csky/include/asm/bug.h
> > new file mode 100644
> > index 0000000..bd7b323
> > --- /dev/null
> > +++ b/arch/csky/include/asm/bug.h
> > @@ -0,0 +1,26 @@
> > +/* SPDX-License-Identifier: GPL-2.0 */
> > +// Copyright (C) 2018 Hangzhou C-SKY Microsystems co.,ltd.
> > +
> > +#ifndef __ASM_CSKY_BUG_H
> > +#define __ASM_CSKY_BUG_H
> > +
> > +#include <linux/compiler.h>
> > +#include <linux/const.h>
> > +#include <linux/types.h>
> > +
> > +#define BUG() \
> > +do { \
> > + asm volatile ("bkpt\n"); \
> > + unreachable(); \
> > +} while (0)
>
> Since you have a breakpoint instruction, a nice addition might
> be to use that for implementing WARN_ON as well, whenever you
> get around to doing that. It's certainly not required for the initial
> merge.
Thx for remind and I'll improve it next.
Best Regards
Guo Ren
^ permalink raw reply [flat|nested] 68+ messages in thread
* [PATCH V9 16/21] csky: SMP support
2018-10-16 2:58 [PATCH V9 00/21] C-SKY(csky) Linux Kernel Port Guo Ren
` (14 preceding siblings ...)
2018-10-16 2:58 ` [PATCH V9 15/21] csky: Debug and Ptrace GDB Guo Ren
@ 2018-10-16 2:58 ` Guo Ren
2018-10-17 15:47 ` Arnd Bergmann
2018-10-16 2:58 ` [PATCH V9 17/21] csky: Misc headers Guo Ren
` (5 subsequent siblings)
21 siblings, 1 reply; 68+ messages in thread
From: Guo Ren @ 2018-10-16 2:58 UTC (permalink / raw)
To: akpm, arnd, daniel.lezcano, davem, gregkh, hch, marc.zyngier,
mark.rutland, peterz, robh, tglx
Cc: linux-kernel, linux-arch, devicetree, robh+dt,
c-sky_gcc_upstream, Guo Ren
This patch adds boot, ipi, hotplug codes for SMP.
Signed-off-by: Guo Ren <ren_guo@c-sky.com>
Cc: Marc Zyngier <marc.zyngier@arm.com>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Arnd Bergmann <arnd@arndb.de>
---
Changelog:
- Use PER_CPU_DEFINE for ipi_data.
- Remove unused variable and add static for enable_smp_ipi().
- Add comment to explain pass args from primary CPU to secondary CPU.
- Remove set_ipi_irq_mapping callback.
- Convert the cpumask to an interrupt-controller specific representation
in driver's code, and not the SMP code's.
- csky: remove irq_mapping from smp.c
There are some feedbacks from irqchip, and we need to adjust
"smp.c & smp.h" to match the csky_mp_intc modification.
- Move IPI_IRQ define into drivers/irqchip/csky_mpintc.c, because it's a
interrupt controller specific.
- Bugfix request_irq with IPI_IRQ, we must use irq_mapping return value
not directly use IPI_IRQ. The modification also involves csky_mp_intc.
---
---
arch/csky/include/asm/smp.h | 26 +++++
arch/csky/kernel/smp.c | 237 ++++++++++++++++++++++++++++++++++++++++++++
2 files changed, 263 insertions(+)
create mode 100644 arch/csky/include/asm/smp.h
create mode 100644 arch/csky/kernel/smp.c
diff --git a/arch/csky/include/asm/smp.h b/arch/csky/include/asm/smp.h
new file mode 100644
index 0000000..4a929c4
--- /dev/null
+++ b/arch/csky/include/asm/smp.h
@@ -0,0 +1,26 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+
+#ifndef __ASM_CSKY_SMP_H
+#define __ASM_CSKY_SMP_H
+
+#include <linux/cpumask.h>
+#include <linux/irqreturn.h>
+#include <linux/threads.h>
+
+#ifdef CONFIG_SMP
+
+void __init setup_smp(void);
+
+void __init setup_smp_ipi(void);
+
+void arch_send_call_function_ipi_mask(struct cpumask *mask);
+
+void arch_send_call_function_single_ipi(int cpu);
+
+void __init set_send_ipi(void (*func)(const struct cpumask *mask), int irq);
+
+#define raw_smp_processor_id() (current_thread_info()->cpu)
+
+#endif /* CONFIG_SMP */
+
+#endif /* __ASM_CSKY_SMP_H */
diff --git a/arch/csky/kernel/smp.c b/arch/csky/kernel/smp.c
new file mode 100644
index 0000000..36ebaf9
--- /dev/null
+++ b/arch/csky/kernel/smp.c
@@ -0,0 +1,237 @@
+// SPDX-License-Identifier: GPL-2.0
+
+#include <linux/module.h>
+#include <linux/init.h>
+#include <linux/kernel.h>
+#include <linux/mm.h>
+#include <linux/sched.h>
+#include <linux/kernel_stat.h>
+#include <linux/notifier.h>
+#include <linux/cpu.h>
+#include <linux/percpu.h>
+#include <linux/delay.h>
+#include <linux/err.h>
+#include <linux/irq.h>
+#include <linux/irqdomain.h>
+#include <linux/of.h>
+#include <linux/sched/task_stack.h>
+#include <linux/sched/mm.h>
+#include <asm/irq.h>
+#include <asm/traps.h>
+#include <asm/sections.h>
+#include <asm/mmu_context.h>
+#include <asm/pgalloc.h>
+
+struct ipi_data_struct {
+ unsigned long bits ____cacheline_aligned;
+};
+static DEFINE_PER_CPU(struct ipi_data_struct, ipi_data);
+
+enum ipi_message_type {
+ IPI_EMPTY,
+ IPI_RESCHEDULE,
+ IPI_CALL_FUNC,
+ IPI_MAX
+};
+
+static irqreturn_t handle_ipi(int irq, void *dev)
+{
+ while (true) {
+ unsigned long ops;
+
+ ops = xchg(&this_cpu_ptr(&ipi_data)->bits, 0);
+ if (ops == 0)
+ return IRQ_HANDLED;
+
+ if (ops & (1 << IPI_RESCHEDULE))
+ scheduler_ipi();
+
+ if (ops & (1 << IPI_CALL_FUNC))
+ generic_smp_call_function_interrupt();
+
+ BUG_ON((ops >> IPI_MAX) != 0);
+ }
+
+ return IRQ_HANDLED;
+}
+
+static void (*send_arch_ipi)(const struct cpumask *mask);
+
+static int ipi_irq;
+void __init set_send_ipi(void (*func)(const struct cpumask *mask), int irq)
+{
+ if (send_arch_ipi)
+ return;
+
+ send_arch_ipi = func;
+ ipi_irq = irq;
+}
+
+static void
+send_ipi_message(const struct cpumask *to_whom, enum ipi_message_type operation)
+{
+ int i;
+
+ for_each_cpu(i, to_whom)
+ set_bit(operation, &per_cpu_ptr(&ipi_data, i)->bits);
+
+ smp_mb();
+ send_arch_ipi(to_whom);
+}
+
+void arch_send_call_function_ipi_mask(struct cpumask *mask)
+{
+ send_ipi_message(mask, IPI_CALL_FUNC);
+}
+
+void arch_send_call_function_single_ipi(int cpu)
+{
+ send_ipi_message(cpumask_of(cpu), IPI_CALL_FUNC);
+}
+
+static void ipi_stop(void *unused)
+{
+ while (1);
+}
+
+void smp_send_stop(void)
+{
+ on_each_cpu(ipi_stop, NULL, 1);
+}
+
+void smp_send_reschedule(int cpu)
+{
+ send_ipi_message(cpumask_of(cpu), IPI_RESCHEDULE);
+}
+
+void __init smp_prepare_boot_cpu(void)
+{
+}
+
+void __init smp_prepare_cpus(unsigned int max_cpus)
+{
+}
+
+static void __init enable_smp_ipi(void)
+{
+ enable_percpu_irq(ipi_irq, 0);
+}
+
+static int ipi_dummy_dev;
+void __init setup_smp_ipi(void)
+{
+ int rc;
+
+ if (ipi_irq == 0)
+ panic("%s IRQ mapping failed\n", __func__);
+
+ rc = request_percpu_irq(ipi_irq, handle_ipi, "IPI Interrupt",
+ &ipi_dummy_dev);
+ if (rc)
+ panic("%s IRQ request failed\n", __func__);
+
+ enable_smp_ipi();
+}
+
+void __init setup_smp(void)
+{
+ struct device_node *node = NULL;
+ int cpu;
+
+ while ((node = of_find_node_by_type(node, "cpu"))) {
+ if (!of_device_is_available(node))
+ continue;
+
+ if (of_property_read_u32(node, "reg", &cpu))
+ continue;
+
+ if (cpu >= NR_CPUS)
+ continue;
+
+ set_cpu_possible(cpu, true);
+ set_cpu_present(cpu, true);
+ }
+}
+
+extern void _start_smp_secondary(void);
+
+volatile unsigned int secondary_hint;
+volatile unsigned int secondary_ccr;
+volatile unsigned int secondary_stack;
+
+int __cpu_up(unsigned int cpu, struct task_struct *tidle)
+{
+ unsigned int tmp;
+
+ secondary_stack = (unsigned int)tidle->stack + THREAD_SIZE;
+
+ secondary_hint = mfcr("cr31");
+
+ secondary_ccr = mfcr("cr18");
+
+ /*
+ * Because other CPUs are in reset status, we must flush data
+ * from cache to out and secondary CPUs use them in
+ * csky_start_secondary(void)
+ */
+ mtcr("cr17", 0x22);
+
+ /* Enable cpu in SMP reset ctrl reg */
+ tmp = mfcr("cr<29, 0>");
+ tmp |= 1 << cpu;
+ mtcr("cr<29, 0>", tmp);
+
+ /* Wait for the cpu online */
+ while (!cpu_online(cpu));
+
+ secondary_stack = 0;
+
+ return 0;
+}
+
+void __init smp_cpus_done(unsigned int max_cpus)
+{
+}
+
+int setup_profiling_timer(unsigned int multiplier)
+{
+ return -EINVAL;
+}
+
+void csky_start_secondary(void)
+{
+ struct mm_struct *mm = &init_mm;
+ unsigned int cpu = smp_processor_id();
+
+ mtcr("cr31", secondary_hint);
+ mtcr("cr18", secondary_ccr);
+
+ mtcr("vbr", vec_base);
+
+ flush_tlb_all();
+ write_mmu_pagemask(0);
+ TLBMISS_HANDLER_SETUP_PGD(swapper_pg_dir);
+ TLBMISS_HANDLER_SETUP_PGD_KERNEL(swapper_pg_dir);
+
+ asid_cache(smp_processor_id()) = ASID_FIRST_VERSION;
+
+#ifdef CONFIG_CPU_HAS_FPU
+ init_fpu();
+#endif
+
+ enable_smp_ipi();
+
+ mmget(mm);
+ mmgrab(mm);
+ current->active_mm = mm;
+ cpumask_set_cpu(cpu, mm_cpumask(mm));
+
+ notify_cpu_starting(cpu);
+ set_cpu_online(cpu, true);
+
+ pr_info("CPU%u Online: %s...\n", cpu, __func__);
+
+ local_irq_enable();
+ preempt_disable();
+ cpu_startup_entry(CPUHP_AP_ONLINE_IDLE);
+}
--
2.7.4
^ permalink raw reply related [flat|nested] 68+ messages in thread* Re: [PATCH V9 16/21] csky: SMP support
2018-10-16 2:58 ` [PATCH V9 16/21] csky: SMP support Guo Ren
@ 2018-10-17 15:47 ` Arnd Bergmann
0 siblings, 0 replies; 68+ messages in thread
From: Arnd Bergmann @ 2018-10-17 15:47 UTC (permalink / raw)
To: Guo Ren
Cc: Andrew Morton, Daniel Lezcano, David Miller, gregkh,
Christoph Hellwig, Marc Zyngier, Mark Rutland, Peter Zijlstra,
Rob Herring, Thomas Gleixner, Linux Kernel Mailing List,
linux-arch, DTML, Rob Herring, c-sky_gcc_upstream
On Tue, Oct 16, 2018 at 5:09 AM Guo Ren <ren_guo@c-sky.com> wrote:
>
> This patch adds boot, ipi, hotplug codes for SMP.
>
> Signed-off-by: Guo Ren <ren_guo@c-sky.com>
> Cc: Marc Zyngier <marc.zyngier@arm.com>
> Cc: Mark Rutland <mark.rutland@arm.com>
> Cc: Peter Zijlstra <peterz@infradead.org>
> Cc: Arnd Bergmann <arnd@arndb.de>
Reviewed-by: Arnd Bergmann <arnd@arndb.de>
^ permalink raw reply [flat|nested] 68+ messages in thread
* [PATCH V9 17/21] csky: Misc headers
2018-10-16 2:58 [PATCH V9 00/21] C-SKY(csky) Linux Kernel Port Guo Ren
` (15 preceding siblings ...)
2018-10-16 2:58 ` [PATCH V9 16/21] csky: SMP support Guo Ren
@ 2018-10-16 2:58 ` Guo Ren
2018-10-17 15:49 ` Arnd Bergmann
2018-10-16 2:58 ` [PATCH V9 18/21] dt-bindings: csky CPU Bindings Guo Ren
` (4 subsequent siblings)
21 siblings, 1 reply; 68+ messages in thread
From: Guo Ren @ 2018-10-16 2:58 UTC (permalink / raw)
To: akpm, arnd, daniel.lezcano, davem, gregkh, hch, marc.zyngier,
mark.rutland, peterz, robh, tglx
Cc: linux-kernel, linux-arch, devicetree, robh+dt,
c-sky_gcc_upstream, Guo Ren
This patch adds csky registers' definition, bitops, byteorder,
asm-offsets codes.
Signed-off-by: Guo Ren <ren_guo@c-sky.com>
Cc: Arnd Bergmann <arnd@arndb.de>
---
arch/csky/abiv1/inc/abi/reg_ops.h | 27 +++++++++++
arch/csky/abiv1/inc/abi/regdef.h | 26 ++++++++++
arch/csky/abiv2/inc/abi/reg_ops.h | 17 +++++++
arch/csky/abiv2/inc/abi/regdef.h | 26 ++++++++++
arch/csky/include/asm/bitops.h | 82 +++++++++++++++++++++++++++++++
arch/csky/include/asm/checksum.h | 50 +++++++++++++++++++
arch/csky/include/asm/reg_ops.h | 26 ++++++++++
arch/csky/include/uapi/asm/byteorder.h | 9 ++++
arch/csky/kernel/asm-offsets.c | 88 ++++++++++++++++++++++++++++++++++
9 files changed, 351 insertions(+)
create mode 100644 arch/csky/abiv1/inc/abi/reg_ops.h
create mode 100644 arch/csky/abiv1/inc/abi/regdef.h
create mode 100644 arch/csky/abiv2/inc/abi/reg_ops.h
create mode 100644 arch/csky/abiv2/inc/abi/regdef.h
create mode 100644 arch/csky/include/asm/bitops.h
create mode 100644 arch/csky/include/asm/checksum.h
create mode 100644 arch/csky/include/asm/reg_ops.h
create mode 100644 arch/csky/include/uapi/asm/byteorder.h
create mode 100644 arch/csky/kernel/asm-offsets.c
diff --git a/arch/csky/abiv1/inc/abi/reg_ops.h b/arch/csky/abiv1/inc/abi/reg_ops.h
new file mode 100644
index 0000000..a153bd3
--- /dev/null
+++ b/arch/csky/abiv1/inc/abi/reg_ops.h
@@ -0,0 +1,27 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+// Copyright (C) 2018 Hangzhou C-SKY Microsystems co.,ltd.
+
+#ifndef __ABI_REG_OPS_H
+#define __ABI_REG_OPS_H
+#include <asm/reg_ops.h>
+
+#define cprcr(reg) \
+({ \
+ unsigned int tmp; \
+ asm volatile("cprcr %0, "reg"\n":"=b"(tmp)); \
+ tmp; \
+})
+
+#define cpwcr(reg, val) \
+({ \
+ asm volatile("cpwcr %0, "reg"\n"::"b"(val)); \
+})
+
+static inline unsigned int mfcr_hint(void)
+{
+ return mfcr("cr30");
+}
+
+static inline unsigned int mfcr_ccr2(void) { return 0; }
+
+#endif /* __ABI_REG_OPS_H */
diff --git a/arch/csky/abiv1/inc/abi/regdef.h b/arch/csky/abiv1/inc/abi/regdef.h
new file mode 100644
index 0000000..8766892
--- /dev/null
+++ b/arch/csky/abiv1/inc/abi/regdef.h
@@ -0,0 +1,26 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+// Copyright (C) 2018 Hangzhou C-SKY Microsystems co.,ltd.
+
+#ifndef __ASM_CSKY_REGDEF_H
+#define __ASM_CSKY_REGDEF_H
+
+#define syscallid r1
+#define r11_sig r11
+
+#define regs_syscallid(regs) regs->regs[9]
+
+/*
+ * PSR format:
+ * | 31 | 30-24 | 23-16 | 15 14 | 13-0 |
+ * S CPID VEC TM
+ *
+ * S: Super Mode
+ * CPID: Coprocessor id, only 15 for MMU
+ * VEC: Exception Number
+ * TM: Trace Mode
+ */
+#define DEFAULT_PSR_VALUE 0x8f000000
+
+#define SYSTRACE_SAVENUM 2
+
+#endif /* __ASM_CSKY_REGDEF_H */
diff --git a/arch/csky/abiv2/inc/abi/reg_ops.h b/arch/csky/abiv2/inc/abi/reg_ops.h
new file mode 100644
index 0000000..ae82c3f
--- /dev/null
+++ b/arch/csky/abiv2/inc/abi/reg_ops.h
@@ -0,0 +1,17 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+// Copyright (C) 2018 Hangzhou C-SKY Microsystems co.,ltd.
+
+#ifndef __ABI_REG_OPS_H
+#define __ABI_REG_OPS_H
+#include <asm/reg_ops.h>
+
+static inline unsigned int mfcr_hint(void)
+{
+ return mfcr("cr31");
+}
+
+static inline unsigned int mfcr_ccr2(void)
+{
+ return mfcr("cr23");
+}
+#endif /* __ABI_REG_OPS_H */
diff --git a/arch/csky/abiv2/inc/abi/regdef.h b/arch/csky/abiv2/inc/abi/regdef.h
new file mode 100644
index 0000000..c72abb7
--- /dev/null
+++ b/arch/csky/abiv2/inc/abi/regdef.h
@@ -0,0 +1,26 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+// Copyright (C) 2018 Hangzhou C-SKY Microsystems co.,ltd.
+
+#ifndef __ASM_CSKY_REGDEF_H
+#define __ASM_CSKY_REGDEF_H
+
+#define syscallid r7
+#define r11_sig r11
+
+#define regs_syscallid(regs) regs->regs[3]
+
+/*
+ * PSR format:
+ * | 31 | 30-24 | 23-16 | 15 14 | 13-10 | 9 | 8-0 |
+ * S VEC TM MM
+ *
+ * S: Super Mode
+ * VEC: Exception Number
+ * TM: Trace Mode
+ * MM: Memory unaligned addr access
+ */
+#define DEFAULT_PSR_VALUE 0x80000200
+
+#define SYSTRACE_SAVENUM 5
+
+#endif /* __ASM_CSKY_REGDEF_H */
diff --git a/arch/csky/include/asm/bitops.h b/arch/csky/include/asm/bitops.h
new file mode 100644
index 0000000..335f288
--- /dev/null
+++ b/arch/csky/include/asm/bitops.h
@@ -0,0 +1,82 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+// Copyright (C) 2018 Hangzhou C-SKY Microsystems co.,ltd.
+
+#ifndef __ASM_CSKY_BITOPS_H
+#define __ASM_CSKY_BITOPS_H
+
+#include <linux/compiler.h>
+#include <asm/barrier.h>
+
+/*
+ * asm-generic/bitops/ffs.h
+ */
+static inline int ffs(int x)
+{
+ if (!x)
+ return 0;
+
+ asm volatile (
+ "brev %0\n"
+ "ff1 %0\n"
+ "addi %0, 1\n"
+ : "=&r"(x)
+ : "0"(x));
+ return x;
+}
+
+/*
+ * asm-generic/bitops/__ffs.h
+ */
+static __always_inline unsigned long __ffs(unsigned long x)
+{
+ asm volatile (
+ "brev %0\n"
+ "ff1 %0\n"
+ : "=&r"(x)
+ : "0"(x));
+ return x;
+}
+
+/*
+ * asm-generic/bitops/fls.h
+ */
+static __always_inline int fls(int x)
+{
+ asm volatile(
+ "ff1 %0\n"
+ : "=&r"(x)
+ : "0"(x));
+
+ return (32 - x);
+}
+
+/*
+ * asm-generic/bitops/__fls.h
+ */
+static __always_inline unsigned long __fls(unsigned long x)
+{
+ return fls(x) - 1;
+}
+
+#include <asm-generic/bitops/ffz.h>
+#include <asm-generic/bitops/fls64.h>
+#include <asm-generic/bitops/find.h>
+
+#ifndef _LINUX_BITOPS_H
+#error only <linux/bitops.h> can be included directly
+#endif
+
+#include <asm-generic/bitops/sched.h>
+#include <asm-generic/bitops/hweight.h>
+#include <asm-generic/bitops/lock.h>
+#include <asm-generic/bitops/atomic.h>
+
+/*
+ * bug fix, why only could use atomic!!!!
+ */
+#include <asm-generic/bitops/non-atomic.h>
+#define __clear_bit(nr, vaddr) clear_bit(nr, vaddr)
+
+#include <asm-generic/bitops/le.h>
+#include <asm-generic/bitops/ext2-atomic.h>
+#endif /* __ASM_CSKY_BITOPS_H */
diff --git a/arch/csky/include/asm/checksum.h b/arch/csky/include/asm/checksum.h
new file mode 100644
index 0000000..7685824
--- /dev/null
+++ b/arch/csky/include/asm/checksum.h
@@ -0,0 +1,50 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+// Copyright (C) 2018 Hangzhou C-SKY Microsystems co.,ltd.
+
+#ifndef __ASM_CSKY_CHECKSUM_H
+#define __ASM_CSKY_CHECKSUM_H
+
+#include <linux/in6.h>
+#include <asm/byteorder.h>
+
+static inline __sum16 csum_fold(__wsum csum)
+{
+ u32 tmp;
+
+ asm volatile(
+ "mov %1, %0\n"
+ "rori %0, 16\n"
+ "addu %0, %1\n"
+ "lsri %0, 16\n"
+ : "=r"(csum), "=r"(tmp)
+ : "0"(csum));
+
+ return (__force __sum16) ~csum;
+}
+#define csum_fold csum_fold
+
+static inline __wsum csum_tcpudp_nofold(__be32 saddr, __be32 daddr,
+ unsigned short len, unsigned short proto, __wsum sum)
+{
+ asm volatile(
+ "clrc\n"
+ "addc %0, %1\n"
+ "addc %0, %2\n"
+ "addc %0, %3\n"
+ "inct %0\n"
+ : "=r"(sum)
+ : "r"((__force u32)saddr), "r"((__force u32)daddr),
+#ifdef __BIG_ENDIAN
+ "r"(proto + len),
+#else
+ "r"((proto + len) << 8),
+#endif
+ "0" ((__force unsigned long)sum)
+ : "cc");
+ return sum;
+}
+#define csum_tcpudp_nofold csum_tcpudp_nofold
+
+#include <asm-generic/checksum.h>
+
+#endif /* __ASM_CSKY_CHECKSUM_H */
diff --git a/arch/csky/include/asm/reg_ops.h b/arch/csky/include/asm/reg_ops.h
new file mode 100644
index 0000000..cccf7d5
--- /dev/null
+++ b/arch/csky/include/asm/reg_ops.h
@@ -0,0 +1,26 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+
+#ifndef __ASM_REGS_OPS_H
+#define __ASM_REGS_OPS_H
+
+#define mfcr(reg) \
+({ \
+ unsigned int tmp; \
+ asm volatile( \
+ "mfcr %0, "reg"\n" \
+ : "=r"(tmp) \
+ : \
+ : "memory"); \
+ tmp; \
+})
+
+#define mtcr(reg, val) \
+({ \
+ asm volatile( \
+ "mtcr %0, "reg"\n" \
+ : \
+ : "r"(val) \
+ : "memory"); \
+})
+
+#endif /* __ASM_REGS_OPS_H */
diff --git a/arch/csky/include/uapi/asm/byteorder.h b/arch/csky/include/uapi/asm/byteorder.h
new file mode 100644
index 0000000..b079ec7
--- /dev/null
+++ b/arch/csky/include/uapi/asm/byteorder.h
@@ -0,0 +1,9 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+// Copyright (C) 2018 Hangzhou C-SKY Microsystems co.,ltd.
+
+#ifndef __ASM_CSKY_BYTEORDER_H
+#define __ASM_CSKY_BYTEORDER_H
+
+#include <linux/byteorder/little_endian.h>
+
+#endif /* __ASM_CSKY_BYTEORDER_H */
diff --git a/arch/csky/kernel/asm-offsets.c b/arch/csky/kernel/asm-offsets.c
new file mode 100644
index 0000000..8d3ed81
--- /dev/null
+++ b/arch/csky/kernel/asm-offsets.c
@@ -0,0 +1,88 @@
+// SPDX-License-Identifier: GPL-2.0
+// Copyright (C) 2018 Hangzhou C-SKY Microsystems co.,ltd.
+
+#include <linux/sched.h>
+#include <linux/kernel_stat.h>
+#include <linux/kbuild.h>
+#include <abi/regdef.h>
+
+int main(void)
+{
+ /* offsets into the task struct */
+ DEFINE(TASK_STATE, offsetof(struct task_struct, state));
+ DEFINE(TASK_THREAD_INFO, offsetof(struct task_struct, stack));
+ DEFINE(TASK_FLAGS, offsetof(struct task_struct, flags));
+ DEFINE(TASK_PTRACE, offsetof(struct task_struct, ptrace));
+ DEFINE(TASK_THREAD, offsetof(struct task_struct, thread));
+ DEFINE(TASK_MM, offsetof(struct task_struct, mm));
+ DEFINE(TASK_ACTIVE_MM, offsetof(struct task_struct, active_mm));
+
+ /* offsets into the thread struct */
+ DEFINE(THREAD_KSP, offsetof(struct thread_struct, ksp));
+ DEFINE(THREAD_SR, offsetof(struct thread_struct, sr));
+ DEFINE(THREAD_ESP0, offsetof(struct thread_struct, esp0));
+ DEFINE(THREAD_FESR, offsetof(struct thread_struct, user_fp.fesr));
+ DEFINE(THREAD_FCR, offsetof(struct thread_struct, user_fp.fcr));
+ DEFINE(THREAD_FPREG, offsetof(struct thread_struct, user_fp.vr));
+ DEFINE(THREAD_DSPHI, offsetof(struct thread_struct, hi));
+ DEFINE(THREAD_DSPLO, offsetof(struct thread_struct, lo));
+
+ /* offsets into the thread_info struct */
+ DEFINE(TINFO_FLAGS, offsetof(struct thread_info, flags));
+ DEFINE(TINFO_PREEMPT, offsetof(struct thread_info, preempt_count));
+ DEFINE(TINFO_ADDR_LIMIT, offsetof(struct thread_info, addr_limit));
+ DEFINE(TINFO_TP_VALUE, offsetof(struct thread_info, tp_value));
+ DEFINE(TINFO_TASK, offsetof(struct thread_info, task));
+
+ /* offsets into the pt_regs */
+ DEFINE(PT_PC, offsetof(struct pt_regs, pc));
+ DEFINE(PT_ORIG_AO, offsetof(struct pt_regs, orig_a0));
+ DEFINE(PT_SR, offsetof(struct pt_regs, sr));
+
+ DEFINE(PT_A0, offsetof(struct pt_regs, a0));
+ DEFINE(PT_A1, offsetof(struct pt_regs, a1));
+ DEFINE(PT_A2, offsetof(struct pt_regs, a2));
+ DEFINE(PT_A3, offsetof(struct pt_regs, a3));
+ DEFINE(PT_REGS0, offsetof(struct pt_regs, regs[0]));
+ DEFINE(PT_REGS1, offsetof(struct pt_regs, regs[1]));
+ DEFINE(PT_REGS2, offsetof(struct pt_regs, regs[2]));
+ DEFINE(PT_REGS3, offsetof(struct pt_regs, regs[3]));
+ DEFINE(PT_REGS4, offsetof(struct pt_regs, regs[4]));
+ DEFINE(PT_REGS5, offsetof(struct pt_regs, regs[5]));
+ DEFINE(PT_REGS6, offsetof(struct pt_regs, regs[6]));
+ DEFINE(PT_REGS7, offsetof(struct pt_regs, regs[7]));
+ DEFINE(PT_REGS8, offsetof(struct pt_regs, regs[8]));
+ DEFINE(PT_REGS9, offsetof(struct pt_regs, regs[9]));
+ DEFINE(PT_R15, offsetof(struct pt_regs, lr));
+#if defined(__CSKYABIV2__)
+ DEFINE(PT_R16, offsetof(struct pt_regs, exregs[0]));
+ DEFINE(PT_R17, offsetof(struct pt_regs, exregs[1]));
+ DEFINE(PT_R18, offsetof(struct pt_regs, exregs[2]));
+ DEFINE(PT_R19, offsetof(struct pt_regs, exregs[3]));
+ DEFINE(PT_R20, offsetof(struct pt_regs, exregs[4]));
+ DEFINE(PT_R21, offsetof(struct pt_regs, exregs[5]));
+ DEFINE(PT_R22, offsetof(struct pt_regs, exregs[6]));
+ DEFINE(PT_R23, offsetof(struct pt_regs, exregs[7]));
+ DEFINE(PT_R24, offsetof(struct pt_regs, exregs[8]));
+ DEFINE(PT_R25, offsetof(struct pt_regs, exregs[9]));
+ DEFINE(PT_R26, offsetof(struct pt_regs, exregs[10]));
+ DEFINE(PT_R27, offsetof(struct pt_regs, exregs[11]));
+ DEFINE(PT_R28, offsetof(struct pt_regs, exregs[12]));
+ DEFINE(PT_R29, offsetof(struct pt_regs, exregs[13]));
+ DEFINE(PT_R30, offsetof(struct pt_regs, exregs[14]));
+ DEFINE(PT_R31, offsetof(struct pt_regs, exregs[15]));
+ DEFINE(PT_RHI, offsetof(struct pt_regs, rhi));
+ DEFINE(PT_RLO, offsetof(struct pt_regs, rlo));
+#endif
+ DEFINE(PT_USP, offsetof(struct pt_regs, usp));
+
+ /* offsets into the irq_cpustat_t struct */
+ DEFINE(CPUSTAT_SOFTIRQ_PENDING, offsetof(irq_cpustat_t,
+ __softirq_pending));
+
+ /* signal defines */
+ DEFINE(SIGSEGV, SIGSEGV);
+ DEFINE(SIGTRAP, SIGTRAP);
+
+ return 0;
+}
--
2.7.4
^ permalink raw reply related [flat|nested] 68+ messages in thread* Re: [PATCH V9 17/21] csky: Misc headers
2018-10-16 2:58 ` [PATCH V9 17/21] csky: Misc headers Guo Ren
@ 2018-10-17 15:49 ` Arnd Bergmann
0 siblings, 0 replies; 68+ messages in thread
From: Arnd Bergmann @ 2018-10-17 15:49 UTC (permalink / raw)
To: Guo Ren
Cc: Andrew Morton, Daniel Lezcano, David Miller, gregkh,
Christoph Hellwig, Marc Zyngier, Mark Rutland, Peter Zijlstra,
Rob Herring, Thomas Gleixner, Linux Kernel Mailing List,
linux-arch, DTML, Rob Herring, c-sky_gcc_upstream
On Tue, Oct 16, 2018 at 5:29 AM Guo Ren <ren_guo@c-sky.com> wrote:
>
> This patch adds csky registers' definition, bitops, byteorder,
> asm-offsets codes.
>
> Signed-off-by: Guo Ren <ren_guo@c-sky.com>
> Cc: Arnd Bergmann <arnd@arndb.de>
Reviewed-by: Arnd Bergmann <arnd@arndb.de>
^ permalink raw reply [flat|nested] 68+ messages in thread
* [PATCH V9 18/21] dt-bindings: csky CPU Bindings
2018-10-16 2:58 [PATCH V9 00/21] C-SKY(csky) Linux Kernel Port Guo Ren
` (16 preceding siblings ...)
2018-10-16 2:58 ` [PATCH V9 17/21] csky: Misc headers Guo Ren
@ 2018-10-16 2:58 ` Guo Ren
2018-10-17 15:50 ` Arnd Bergmann
2018-10-18 14:31 ` Rob Herring
2018-10-16 2:58 ` [PATCH V9 19/21] dt-bindings: Add vendor prefix for csky Guo Ren
` (3 subsequent siblings)
21 siblings, 2 replies; 68+ messages in thread
From: Guo Ren @ 2018-10-16 2:58 UTC (permalink / raw)
To: akpm, arnd, daniel.lezcano, davem, gregkh, hch, marc.zyngier,
mark.rutland, peterz, robh, tglx
Cc: linux-kernel, linux-arch, devicetree, robh+dt,
c-sky_gcc_upstream, Guo Ren
This patch adds the documentation to describe that how to add cpu nodes in
dts for SMP.
Signed-off-by: Guo Ren <ren_guo@c-sky.com>
Cc: Rob Herring <robh@kernel.org>
---
Changelog:
- Add compatible.
- Remove status part.
---
---
Documentation/devicetree/bindings/csky/cpus.txt | 73 +++++++++++++++++++++++++
1 file changed, 73 insertions(+)
create mode 100644 Documentation/devicetree/bindings/csky/cpus.txt
diff --git a/Documentation/devicetree/bindings/csky/cpus.txt b/Documentation/devicetree/bindings/csky/cpus.txt
new file mode 100644
index 0000000..ae79412
--- /dev/null
+++ b/Documentation/devicetree/bindings/csky/cpus.txt
@@ -0,0 +1,73 @@
+==================
+C-SKY CPU Bindings
+==================
+
+The device tree allows to describe the layout of CPUs in a system through
+the "cpus" node, which in turn contains a number of subnodes (ie "cpu")
+defining properties for every cpu.
+
+Only SMP system need to care about the cpus node and single processor
+needn't define cpus node at all.
+
+=====================================
+cpus and cpu node bindings definition
+=====================================
+
+- cpus node
+
+ Description: Container of cpu nodes
+
+ The node name must be "cpus".
+
+ A cpus node must define the following properties:
+
+ - #address-cells
+ Usage: required
+ Value type: <u32>
+ Definition: must be set to 1
+ - #size-cells
+ Usage: required
+ Value type: <u32>
+ Definition: must be set to 0
+
+- cpu node
+
+ Description: Describes one of SMP cores
+
+ PROPERTIES
+
+ - device_type
+ Usage: required
+ Value type: <string>
+ Definition: must be "cpu"
+ - reg
+ Usage: required
+ Value type: <u32>
+ Definition: CPU index
+ - compatible:
+ Usage: required
+ Value type: <string>
+ Definition: must contain "csky", eg:
+ "csky,610"
+ "csky,807"
+ "csky,810"
+ "csky,860"
+
+Example:
+--------
+
+ cpus {
+ #address-cells = <1>;
+ #size-cells = <0>;
+ cpu@0 {
+ device_type = "cpu";
+ reg = <0>;
+ status = "ok";
+ };
+
+ cpu@1 {
+ device_type = "cpu";
+ reg = <1>;
+ status = "ok";
+ };
+ };
--
2.7.4
^ permalink raw reply related [flat|nested] 68+ messages in thread* Re: [PATCH V9 18/21] dt-bindings: csky CPU Bindings
2018-10-16 2:58 ` [PATCH V9 18/21] dt-bindings: csky CPU Bindings Guo Ren
@ 2018-10-17 15:50 ` Arnd Bergmann
2018-10-18 3:21 ` Guo Ren
2018-10-18 14:31 ` Rob Herring
1 sibling, 1 reply; 68+ messages in thread
From: Arnd Bergmann @ 2018-10-17 15:50 UTC (permalink / raw)
To: Guo Ren
Cc: Andrew Morton, Daniel Lezcano, David Miller, gregkh,
Christoph Hellwig, Marc Zyngier, Mark Rutland, Peter Zijlstra,
Rob Herring, Thomas Gleixner, Linux Kernel Mailing List,
linux-arch, DTML, Rob Herring, c-sky_gcc_upstream
On Tue, Oct 16, 2018 at 5:30 AM Guo Ren <ren_guo@c-sky.com> wrote:
>
> This patch adds the documentation to describe that how to add cpu nodes in
> dts for SMP.
>
> Signed-off-by: Guo Ren <ren_guo@c-sky.com>
> Cc: Rob Herring <robh@kernel.org>
I don't see anything wrong, but I'll let Rob or some other DT maintainer
provide an Ack when they are happy.
Arnd
^ permalink raw reply [flat|nested] 68+ messages in thread* Re: [PATCH V9 18/21] dt-bindings: csky CPU Bindings
2018-10-17 15:50 ` Arnd Bergmann
@ 2018-10-18 3:21 ` Guo Ren
2018-10-18 3:45 ` Guo Ren
0 siblings, 1 reply; 68+ messages in thread
From: Guo Ren @ 2018-10-18 3:21 UTC (permalink / raw)
To: Arnd Bergmann
Cc: Andrew Morton, Daniel Lezcano, David Miller, gregkh,
Christoph Hellwig, Marc Zyngier, Mark Rutland, Peter Zijlstra,
Rob Herring, Thomas Gleixner, Linux Kernel Mailing List,
linux-arch, DTML, Rob Herring, c-sky_gcc_upstream
On Wed, Oct 17, 2018 at 05:50:33PM +0200, Arnd Bergmann wrote:
> On Tue, Oct 16, 2018 at 5:30 AM Guo Ren <ren_guo@c-sky.com> wrote:
> >
> > This patch adds the documentation to describe that how to add cpu nodes in
> > dts for SMP.
> >
> > Signed-off-by: Guo Ren <ren_guo@c-sky.com>
> > Cc: Rob Herring <robh@kernel.org>
>
> I don't see anything wrong, but I'll let Rob or some other DT maintainer
> provide an Ack when they are happy.
Yes, Rob help a lot for reviewing the DT-docs of csky subsystem. Thx Bob :)
Best Regards
Guo Ren
^ permalink raw reply [flat|nested] 68+ messages in thread
* Re: [PATCH V9 18/21] dt-bindings: csky CPU Bindings
2018-10-18 3:21 ` Guo Ren
@ 2018-10-18 3:45 ` Guo Ren
0 siblings, 0 replies; 68+ messages in thread
From: Guo Ren @ 2018-10-18 3:45 UTC (permalink / raw)
To: Arnd Bergmann
Cc: Andrew Morton, Daniel Lezcano, David Miller, gregkh,
Christoph Hellwig, Marc Zyngier, Mark Rutland, Peter Zijlstra,
Rob Herring, Thomas Gleixner, Linux Kernel Mailing List,
linux-arch, DTML, Rob Herring, c-sky_gcc_upstream
On Thu, Oct 18, 2018 at 11:21:30AM +0800, Guo Ren wrote:
> On Wed, Oct 17, 2018 at 05:50:33PM +0200, Arnd Bergmann wrote:
> > On Tue, Oct 16, 2018 at 5:30 AM Guo Ren <ren_guo@c-sky.com> wrote:
> > >
> > > This patch adds the documentation to describe that how to add cpu nodes in
> > > dts for SMP.
> > >
> > > Signed-off-by: Guo Ren <ren_guo@c-sky.com>
> > > Cc: Rob Herring <robh@kernel.org>
> >
> > I don't see anything wrong, but I'll let Rob or some other DT maintainer
> > provide an Ack when they are happy.
> Yes, Rob help a lot for reviewing the DT-docs of csky subsystem. Thx Bob :)
Sorry, Thx Rob.
>
> Best Regards
> Guo Ren
^ permalink raw reply [flat|nested] 68+ messages in thread
* Re: [PATCH V9 18/21] dt-bindings: csky CPU Bindings
2018-10-16 2:58 ` [PATCH V9 18/21] dt-bindings: csky CPU Bindings Guo Ren
2018-10-17 15:50 ` Arnd Bergmann
@ 2018-10-18 14:31 ` Rob Herring
2018-10-19 2:19 ` Guo Ren
1 sibling, 1 reply; 68+ messages in thread
From: Rob Herring @ 2018-10-18 14:31 UTC (permalink / raw)
To: Guo Ren
Cc: akpm, arnd, daniel.lezcano, davem, gregkh, hch, marc.zyngier,
mark.rutland, peterz, robh, tglx, linux-kernel, linux-arch,
devicetree, robh+dt, c-sky_gcc_upstream, Guo Ren
On Tue, 16 Oct 2018 10:58:37 +0800, Guo Ren wrote:
> This patch adds the documentation to describe that how to add cpu nodes in
> dts for SMP.
>
> Signed-off-by: Guo Ren <ren_guo@c-sky.com>
> Cc: Rob Herring <robh@kernel.org>
> ---
> Changelog:
> - Add compatible.
> - Remove status part.
> ---
> ---
> Documentation/devicetree/bindings/csky/cpus.txt | 73 +++++++++++++++++++++++++
> 1 file changed, 73 insertions(+)
> create mode 100644 Documentation/devicetree/bindings/csky/cpus.txt
>
Reviewed-by: Rob Herring <robh@kernel.org>
^ permalink raw reply [flat|nested] 68+ messages in thread
* Re: [PATCH V9 18/21] dt-bindings: csky CPU Bindings
2018-10-18 14:31 ` Rob Herring
@ 2018-10-19 2:19 ` Guo Ren
0 siblings, 0 replies; 68+ messages in thread
From: Guo Ren @ 2018-10-19 2:19 UTC (permalink / raw)
To: Rob Herring
Cc: akpm, arnd, daniel.lezcano, davem, gregkh, hch, marc.zyngier,
mark.rutland, peterz, tglx, linux-kernel, linux-arch, devicetree,
robh+dt, c-sky_gcc_upstream
On Thu, Oct 18, 2018 at 09:31:35AM -0500, Rob Herring wrote:
> On Tue, 16 Oct 2018 10:58:37 +0800, Guo Ren wrote:
> > This patch adds the documentation to describe that how to add cpu nodes in
> > dts for SMP.
> >
> > Signed-off-by: Guo Ren <ren_guo@c-sky.com>
> > Cc: Rob Herring <robh@kernel.org>
> > ---
> > Changelog:
> > - Add compatible.
> > - Remove status part.
> > ---
> > ---
> > Documentation/devicetree/bindings/csky/cpus.txt | 73 +++++++++++++++++++++++++
> > 1 file changed, 73 insertions(+)
> > create mode 100644 Documentation/devicetree/bindings/csky/cpus.txt
> >
>
> Reviewed-by: Rob Herring <robh@kernel.org>
Thx Rob
Best Regards
Guo Ren
^ permalink raw reply [flat|nested] 68+ messages in thread
* [PATCH V9 19/21] dt-bindings: Add vendor prefix for csky
2018-10-16 2:58 [PATCH V9 00/21] C-SKY(csky) Linux Kernel Port Guo Ren
` (17 preceding siblings ...)
2018-10-16 2:58 ` [PATCH V9 18/21] dt-bindings: csky CPU Bindings Guo Ren
@ 2018-10-16 2:58 ` Guo Ren
2018-10-16 2:58 ` [PATCH V9 20/21] MAINTAINERS: Add csky Guo Ren
` (2 subsequent siblings)
21 siblings, 0 replies; 68+ messages in thread
From: Guo Ren @ 2018-10-16 2:58 UTC (permalink / raw)
To: akpm, arnd, daniel.lezcano, davem, gregkh, hch, marc.zyngier,
mark.rutland, peterz, robh, tglx
Cc: linux-kernel, linux-arch, devicetree, robh+dt,
c-sky_gcc_upstream, Guo Ren
Add csky vendor definition.
Signed-off-by: Guo Ren <ren_guo@c-sky.com>
Reviewed-by: Rob Herring <robh@kernel.org>
---
Documentation/devicetree/bindings/vendor-prefixes.txt | 1 +
1 file changed, 1 insertion(+)
diff --git a/Documentation/devicetree/bindings/vendor-prefixes.txt b/Documentation/devicetree/bindings/vendor-prefixes.txt
index 2c3fc51..a728ee3 100644
--- a/Documentation/devicetree/bindings/vendor-prefixes.txt
+++ b/Documentation/devicetree/bindings/vendor-prefixes.txt
@@ -84,6 +84,7 @@ cosmic Cosmic Circuits
crane Crane Connectivity Solutions
creative Creative Technology Ltd
crystalfontz Crystalfontz America, Inc.
+csky Hangzhou C-SKY Microsystems Co., Ltd
cubietech Cubietech, Ltd.
cypress Cypress Semiconductor Corporation
cznic CZ.NIC, z.s.p.o.
--
2.7.4
^ permalink raw reply related [flat|nested] 68+ messages in thread* [PATCH V9 20/21] MAINTAINERS: Add csky
2018-10-16 2:58 [PATCH V9 00/21] C-SKY(csky) Linux Kernel Port Guo Ren
` (18 preceding siblings ...)
2018-10-16 2:58 ` [PATCH V9 19/21] dt-bindings: Add vendor prefix for csky Guo Ren
@ 2018-10-16 2:58 ` Guo Ren
2018-10-17 15:51 ` Arnd Bergmann
2018-10-16 5:48 ` [PATCH V9 21/21] csky: support dword access for get_user_size() Guo Ren
2018-10-17 15:58 ` [PATCH V9 00/21] C-SKY(csky) Linux Kernel Port Arnd Bergmann
21 siblings, 1 reply; 68+ messages in thread
From: Guo Ren @ 2018-10-16 2:58 UTC (permalink / raw)
To: akpm, arnd, daniel.lezcano, davem, gregkh, hch, marc.zyngier,
mark.rutland, peterz, robh, tglx
Cc: linux-kernel, linux-arch, devicetree, robh+dt,
c-sky_gcc_upstream, Guo Ren
Add a maintainer information for the csky(C-SKY) architecture.
Signed-off-by: Guo Ren <ren_guo@c-sky.com>
---
MAINTAINERS | 9 +++++++++
1 file changed, 9 insertions(+)
diff --git a/MAINTAINERS b/MAINTAINERS
index d870cb5..9dac20e 100644
--- a/MAINTAINERS
+++ b/MAINTAINERS
@@ -3211,6 +3211,15 @@ T: git git://git.alsa-project.org/alsa-kernel.git
S: Maintained
F: sound/pci/oxygen/
+C-SKY ARCHITECTURE
+M: Guo Ren <ren_guo@c-sky.com>
+T: git https://github.com/c-sky/csky-linux.git
+S: Supported
+F: arch/csky/
+F: Documentation/devicetree/bindings/csky/
+K: csky
+N: csky
+
C6X ARCHITECTURE
M: Mark Salter <msalter@redhat.com>
M: Aurelien Jacquiot <jacquiot.aurelien@gmail.com>
--
2.7.4
^ permalink raw reply related [flat|nested] 68+ messages in thread* Re: [PATCH V9 20/21] MAINTAINERS: Add csky
2018-10-16 2:58 ` [PATCH V9 20/21] MAINTAINERS: Add csky Guo Ren
@ 2018-10-17 15:51 ` Arnd Bergmann
0 siblings, 0 replies; 68+ messages in thread
From: Arnd Bergmann @ 2018-10-17 15:51 UTC (permalink / raw)
To: Guo Ren
Cc: Andrew Morton, Daniel Lezcano, David Miller, gregkh,
Christoph Hellwig, Marc Zyngier, Mark Rutland, Peter Zijlstra,
Rob Herring, Thomas Gleixner, Linux Kernel Mailing List,
linux-arch, DTML, Rob Herring, c-sky_gcc_upstream
On Tue, Oct 16, 2018 at 5:31 AM Guo Ren <ren_guo@c-sky.com> wrote:
>
> Add a maintainer information for the csky(C-SKY) architecture.
>
> Signed-off-by: Guo Ren <ren_guo@c-sky.com>
Reviewed-by: Arnd Bergmann <arnd@arndb.de>
^ permalink raw reply [flat|nested] 68+ messages in thread
* [PATCH V9 21/21] csky: support dword access for get_user_size()
2018-10-16 2:58 [PATCH V9 00/21] C-SKY(csky) Linux Kernel Port Guo Ren
` (19 preceding siblings ...)
2018-10-16 2:58 ` [PATCH V9 20/21] MAINTAINERS: Add csky Guo Ren
@ 2018-10-16 5:48 ` Guo Ren
2018-10-17 15:44 ` Arnd Bergmann
2018-10-17 15:58 ` [PATCH V9 00/21] C-SKY(csky) Linux Kernel Port Arnd Bergmann
21 siblings, 1 reply; 68+ messages in thread
From: Guo Ren @ 2018-10-16 5:48 UTC (permalink / raw)
To: akpm, arnd, daniel.lezcano, davem, gregkh, hch, marc.zyngier,
mark.rutland, peterz, robh, tglx
Cc: linux-kernel, linux-arch, devicetree, robh+dt,
c-sky_gcc_upstream, Guo Ren
Support dword access for get_user_size and redesign put_user_size with
the same style of get_user_size. It's Ok to use xxx_user_asm_common for
all size of variable with ldb, ldh, ldw, ld.d
ld.d rx, (addr, 0) could "rx <= addr" "and r(x+1) <= addr+4" and this also
follow abiv2 gcc ABI for dword access.
abiv1 couldn't support __get_user_dword now.
Signed-off-by: Guo Ren <ren_guo@c-sky.com>
Cc: Arnd Bergmann <arnd@arndb.de>
---
arch/csky/include/asm/uaccess.h | 140 ++++++++++++++++------------------------
1 file changed, 56 insertions(+), 84 deletions(-)
diff --git a/arch/csky/include/asm/uaccess.h b/arch/csky/include/asm/uaccess.h
index acaf0e210..80233e4 100644
--- a/arch/csky/include/asm/uaccess.h
+++ b/arch/csky/include/asm/uaccess.h
@@ -96,90 +96,32 @@ extern int __put_user_bad(void);
__pu_err; \
})
-#define __put_user_size(x, ptr, size, retval) \
-do { \
- retval = 0; \
- switch (size) { \
- case 1: \
- __put_user_asm_b(x, ptr, retval); \
- break; \
- case 2: \
- __put_user_asm_h(x, ptr, retval); \
- break; \
- case 4: \
- __put_user_asm_w(x, ptr, retval); \
- break; \
- case 8: \
- __put_user_asm_64(x, ptr, retval); \
- break; \
- default: \
- __put_user_bad(); \
- } \
-} while (0)
-
-/*
- * We don't tell gcc that we are accessing memory, but this is OK
- * because we do not write to any memory gcc knows about, so there
- * are no aliasing issues.
- *
- * Note that PC at a fault is the address *after* the faulting
- * instruction.
- */
-#define __put_user_asm_b(x, ptr, err) \
-do { \
- int errcode; \
- asm volatile( \
- "1: stb %1, (%2,0) \n" \
- " br 3f \n" \
- "2: mov %0, %3 \n" \
- " br 3f \n" \
- ".section __ex_table, \"a\" \n" \
- ".align 2 \n" \
- ".long 1b,2b \n" \
- ".previous \n" \
- "3: \n" \
- : "=r"(err), "=r"(x), "=r"(ptr), "=r"(errcode) \
- : "0"(err), "1"(x), "2"(ptr), "3"(-EFAULT) \
- : "memory"); \
-} while (0)
-
-#define __put_user_asm_h(x, ptr, err) \
-do { \
- int errcode; \
- asm volatile( \
- "1: sth %1, (%2,0) \n" \
- " br 3f \n" \
- "2: mov %0, %3 \n" \
- " br 3f \n" \
- ".section __ex_table, \"a\" \n" \
- ".align 2 \n" \
- ".long 1b,2b \n" \
- ".previous \n" \
- "3: \n" \
- : "=r"(err), "=r"(x), "=r"(ptr), "=r"(errcode) \
- : "0"(err), "1"(x), "2"(ptr), "3"(-EFAULT) \
- : "memory"); \
-} while (0)
-
-#define __put_user_asm_w(x, ptr, err) \
-do { \
- int errcode; \
- asm volatile( \
- "1: stw %1, (%2,0) \n" \
- " br 3f \n" \
- "2: mov %0, %3 \n" \
- " br 3f \n" \
- ".section __ex_table,\"a\" \n" \
- ".align 2 \n" \
- ".long 1b, 2b \n" \
- ".previous \n" \
- "3: \n" \
- : "=r"(err), "=r"(x), "=r"(ptr), "=r"(errcode) \
- : "0"(err), "1"(x), "2"(ptr), "3"(-EFAULT) \
- : "memory"); \
+#define __put_user_size(x, ptr, size, retval) \
+do { \
+ retval = 0; \
+ switch (size) { \
+ case 1: \
+ __put_user_asm_common(x, ptr, "stb", retval); \
+ break; \
+ case 2: \
+ __put_user_asm_common(x, ptr, "sth", retval); \
+ break; \
+ case 4: \
+ __put_user_asm_common(x, ptr, "stw", retval); \
+ break; \
+ case 8: \
+ __put_user_asm_dword(x, ptr, retval); \
+ break; \
+ default: \
+ __put_user_bad(); \
+ } \
} while (0)
-#define __put_user_asm_64(x, ptr, err) \
+#if defined(__CSKYABIV2__)
+#define __put_user_asm_dword(x, ptr, err) \
+ __put_user_asm_common(x, ptr, "st.d", err)
+#else
+#define __put_user_asm_dword(x, ptr, err) \
do { \
int tmp; \
int errcode; \
@@ -205,6 +147,25 @@ do { \
: "0"(err), "1"(psrc), "2"(ptr), "3"(0), "4"(-EFAULT) \
: "memory"); \
} while (0)
+#endif
+
+#define __put_user_asm_common(x, ptr, ins, err) \
+do { \
+ int errcode; \
+ asm volatile( \
+ "1: " ins " %1, (%2,0) \n" \
+ " br 3f \n" \
+ "2: mov %0, %3 \n" \
+ " br 3f \n" \
+ ".section __ex_table, \"a\" \n" \
+ ".align 2 \n" \
+ ".long 1b,2b \n" \
+ ".previous \n" \
+ "3: \n" \
+ : "=r"(err), "=r"(x), "=r"(ptr), "=r"(errcode) \
+ : "0"(err), "1"(x), "2"(ptr), "3"(-EFAULT) \
+ : "memory"); \
+} while (0)
#define __get_user_nocheck(x, ptr, size) \
({ \
@@ -222,6 +183,7 @@ do { \
__gu_err; \
})
+extern int __get_user_bad(void);
#define __get_user_size(x, ptr, size, retval) \
do { \
switch (size) { \
@@ -234,12 +196,24 @@ do { \
case 4: \
__get_user_asm_common((x), ptr, "ldw", retval); \
break; \
+ case 8: \
+ __get_user_asm_common((x), ptr, "ld.d", retval);\
+ break; \
default: \
x = 0; \
(retval) = __get_user_bad(); \
} \
} while (0)
+#if defined(__CSKYABIV2__)
+#define __get_user_asm_dword(x, ptr, err) \
+ __get_user_asm_common(x, ptr, "ld.d", err)
+#else
+/* fixme: abiv1 not implement get_user_dword */
+#define __get_user_asm_dword(x, ptr, err) \
+ __get_user_bad()
+#endif
+
#define __get_user_asm_common(x, ptr, ins, err) \
do { \
int errcode; \
@@ -260,8 +234,6 @@ do { \
: "memory"); \
} while (0)
-extern int __get_user_bad(void);
-
#define __copy_user(to, from, n) \
do { \
int w0, w1, w2, w3; \
--
2.7.4
^ permalink raw reply related [flat|nested] 68+ messages in thread* Re: [PATCH V9 21/21] csky: support dword access for get_user_size()
2018-10-16 5:48 ` [PATCH V9 21/21] csky: support dword access for get_user_size() Guo Ren
@ 2018-10-17 15:44 ` Arnd Bergmann
2018-10-18 3:41 ` Guo Ren
0 siblings, 1 reply; 68+ messages in thread
From: Arnd Bergmann @ 2018-10-17 15:44 UTC (permalink / raw)
To: Guo Ren
Cc: Andrew Morton, Daniel Lezcano, David Miller, gregkh,
Christoph Hellwig, Marc Zyngier, Mark Rutland, Peter Zijlstra,
Rob Herring, Thomas Gleixner, Linux Kernel Mailing List,
linux-arch, DTML, Rob Herring, c-sky_gcc_upstream
On Tue, Oct 16, 2018 at 5:33 AM Guo Ren <ren_guo@c-sky.com> wrote:
>
> Support dword access for get_user_size and redesign put_user_size with
> the same style of get_user_size. It's Ok to use xxx_user_asm_common for
> all size of variable with ldb, ldh, ldw, ld.d
>
> ld.d rx, (addr, 0) could "rx <= addr" "and r(x+1) <= addr+4" and this also
> follow abiv2 gcc ABI for dword access.
Are you sure this is correct for this?
static inline u32 get_64_to_32(__u64 __user *p)
{
u32 ret;
get_user(ret, p);
return ret;
}
If I read __get_user_asm_common() right, the ld.d would overwrite
two registers, but the caller only expects one, so it clobbers one
that might be in use.
Arnd
^ permalink raw reply [flat|nested] 68+ messages in thread* Re: [PATCH V9 21/21] csky: support dword access for get_user_size()
2018-10-17 15:44 ` Arnd Bergmann
@ 2018-10-18 3:41 ` Guo Ren
2018-10-18 8:34 ` Arnd Bergmann
0 siblings, 1 reply; 68+ messages in thread
From: Guo Ren @ 2018-10-18 3:41 UTC (permalink / raw)
To: Arnd Bergmann
Cc: Andrew Morton, Daniel Lezcano, David Miller, gregkh,
Christoph Hellwig, Marc Zyngier, Mark Rutland, Peter Zijlstra,
Rob Herring, Thomas Gleixner, Linux Kernel Mailing List,
linux-arch, DTML, Rob Herring, c-sky_gcc_upstream
On Wed, Oct 17, 2018 at 05:44:17PM +0200, Arnd Bergmann wrote:
> On Tue, Oct 16, 2018 at 5:33 AM Guo Ren <ren_guo@c-sky.com> wrote:
> >
> > Support dword access for get_user_size and redesign put_user_size with
> > the same style of get_user_size. It's Ok to use xxx_user_asm_common for
> > all size of variable with ldb, ldh, ldw, ld.d
> >
> > ld.d rx, (addr, 0) could "rx <= addr" "and r(x+1) <= addr+4" and this also
> > follow abiv2 gcc ABI for dword access.
>
> Are you sure this is correct for this?
>
> static inline u32 get_64_to_32(__u64 __user *p)
> {
> u32 ret;
> get_user(ret, p);
> return ret;
> }
>
> If I read __get_user_asm_common() right, the ld.d would overwrite
> two registers, but the caller only expects one, so it clobbers one
> that might be in use.
Ah... BUG! I only consider the get_user(u64, u64 *) :P
Change to:
case 8: \
__get_user_asm_dword((x), ptr, "ld.d", retval); \
break;
#define __get_user_asm_dword(x, ptr, err) \
do { \
u64 tmp; \
__get_user_asm_common(tmp, ptr, "ld.d", err); \
x = typeof(x) tmp; \
} while(0)
#define __put_user_asm_dword(x, ptr, err) \
do { \
u64 tmp = (u64) x; \
__put_user_asm_common(tmp, ptr, "st.d", err); \
} while(0)
Hmm?
Best Regards
Guo Ren
^ permalink raw reply [flat|nested] 68+ messages in thread* Re: [PATCH V9 21/21] csky: support dword access for get_user_size()
2018-10-18 3:41 ` Guo Ren
@ 2018-10-18 8:34 ` Arnd Bergmann
2018-10-18 8:57 ` Guo Ren
0 siblings, 1 reply; 68+ messages in thread
From: Arnd Bergmann @ 2018-10-18 8:34 UTC (permalink / raw)
To: Guo Ren
Cc: Andrew Morton, Daniel Lezcano, David Miller, gregkh,
Christoph Hellwig, Marc Zyngier, Mark Rutland, Peter Zijlstra,
Rob Herring, Thomas Gleixner, Linux Kernel Mailing List,
linux-arch, DTML, Rob Herring, c-sky_gcc_upstream
On Thu, Oct 18, 2018 at 5:41 AM Guo Ren <ren_guo@c-sky.com> wrote:
>
> On Wed, Oct 17, 2018 at 05:44:17PM +0200, Arnd Bergmann wrote:
> > On Tue, Oct 16, 2018 at 5:33 AM Guo Ren <ren_guo@c-sky.com> wrote:
> > >
> > > Support dword access for get_user_size and redesign put_user_size with
> > > the same style of get_user_size. It's Ok to use xxx_user_asm_common for
> > > all size of variable with ldb, ldh, ldw, ld.d
> > >
> > > ld.d rx, (addr, 0) could "rx <= addr" "and r(x+1) <= addr+4" and this also
> > > follow abiv2 gcc ABI for dword access.
> >
> > Are you sure this is correct for this?
> >
> > static inline u32 get_64_to_32(__u64 __user *p)
> > {
> > u32 ret;
> > get_user(ret, p);
> > return ret;
> > }
> >
> > If I read __get_user_asm_common() right, the ld.d would overwrite
> > two registers, but the caller only expects one, so it clobbers one
> > that might be in use.
> Ah... BUG! I only consider the get_user(u64, u64 *) :P
>
> Change to:
> case 8: \
> __get_user_asm_dword((x), ptr, "ld.d", retval); \
> break;
>
> #define __get_user_asm_dword(x, ptr, err) \
> do { \
> u64 tmp; \
> __get_user_asm_common(tmp, ptr, "ld.d", err); \
> x = typeof(x) tmp; \
> } while(0)
>
> #define __put_user_asm_dword(x, ptr, err) \
> do { \
> u64 tmp = (u64) x; \
> __put_user_asm_common(tmp, ptr, "st.d", err); \
> } while(0)
>
I think this will cause warnings for code that passes a pointer.
The 64-bit __get_user() is really hard, and most 32-bit architectures don't
implement it at all. If you really want to add it, have a look at what
x86 and arm do. IIRC they both use __builtin_choose_expr(),
Arnd
^ permalink raw reply [flat|nested] 68+ messages in thread* Re: [PATCH V9 21/21] csky: support dword access for get_user_size()
2018-10-18 8:34 ` Arnd Bergmann
@ 2018-10-18 8:57 ` Guo Ren
2018-10-24 7:17 ` Arnd Bergmann
0 siblings, 1 reply; 68+ messages in thread
From: Guo Ren @ 2018-10-18 8:57 UTC (permalink / raw)
To: Arnd Bergmann
Cc: Andrew Morton, Daniel Lezcano, David Miller, gregkh,
Christoph Hellwig, Marc Zyngier, Mark Rutland, Peter Zijlstra,
Rob Herring, Thomas Gleixner, Linux Kernel Mailing List,
linux-arch, DTML, Rob Herring, c-sky_gcc_upstream
On Thu, Oct 18, 2018 at 10:34:00AM +0200, Arnd Bergmann wrote:
> On Thu, Oct 18, 2018 at 5:41 AM Guo Ren <ren_guo@c-sky.com> wrote:
> >
> > On Wed, Oct 17, 2018 at 05:44:17PM +0200, Arnd Bergmann wrote:
> > > On Tue, Oct 16, 2018 at 5:33 AM Guo Ren <ren_guo@c-sky.com> wrote:
> > > >
> > > > Support dword access for get_user_size and redesign put_user_size with
> > > > the same style of get_user_size. It's Ok to use xxx_user_asm_common for
> > > > all size of variable with ldb, ldh, ldw, ld.d
> > > >
> > > > ld.d rx, (addr, 0) could "rx <= addr" "and r(x+1) <= addr+4" and this also
> > > > follow abiv2 gcc ABI for dword access.
> > >
> > > Are you sure this is correct for this?
> > >
> > > static inline u32 get_64_to_32(__u64 __user *p)
> > > {
> > > u32 ret;
> > > get_user(ret, p);
> > > return ret;
> > > }
> > >
> > > If I read __get_user_asm_common() right, the ld.d would overwrite
> > > two registers, but the caller only expects one, so it clobbers one
> > > that might be in use.
> > Ah... BUG! I only consider the get_user(u64, u64 *) :P
> >
> > Change to:
> > case 8: \
> > __get_user_asm_dword((x), ptr, "ld.d", retval); \
> > break;
> >
> > #define __get_user_asm_dword(x, ptr, err) \
> > do { \
> > u64 tmp; \
> > __get_user_asm_common(tmp, ptr, "ld.d", err); \
> > x = typeof(x) tmp; \
> > } while(0)
> >
> > #define __put_user_asm_dword(x, ptr, err) \
> > do { \
> > u64 tmp = (u64) x; \
> > __put_user_asm_common(tmp, ptr, "st.d", err); \
> > } while(0)
> >
>
> I think this will cause warnings for code that passes a pointer.
>
> The 64-bit __get_user() is really hard, and most 32-bit architectures don't
> implement it at all. If you really want to add it, have a look at what
> x86 and arm do. IIRC they both use __builtin_choose_expr(),
Thx for the tips and I'll drop the patch first for the upstream.
I want to implement it because of make allmodconfig and
drivers/android/binder.c need it. I'll learn __builtin_choose_expr()
and prepare patch next.
Best Regards
Guo Ren
^ permalink raw reply [flat|nested] 68+ messages in thread* Re: [PATCH V9 21/21] csky: support dword access for get_user_size()
2018-10-18 8:57 ` Guo Ren
@ 2018-10-24 7:17 ` Arnd Bergmann
2018-10-25 17:08 ` Guo Ren
0 siblings, 1 reply; 68+ messages in thread
From: Arnd Bergmann @ 2018-10-24 7:17 UTC (permalink / raw)
To: Guo Ren
Cc: Andrew Morton, Daniel Lezcano, David Miller, gregkh,
Christoph Hellwig, Marc Zyngier, Mark Rutland, Peter Zijlstra,
Rob Herring, Thomas Gleixner, Linux Kernel Mailing List,
linux-arch, DTML, Rob Herring, c-sky_gcc_upstream
On Thu, Oct 18, 2018 at 9:57 AM Guo Ren <ren_guo@c-sky.com> wrote:
>
> On Thu, Oct 18, 2018 at 10:34:00AM +0200, Arnd Bergmann wrote:
> > On Thu, Oct 18, 2018 at 5:41 AM Guo Ren <ren_guo@c-sky.com> wrote:
> > >
> > > On Wed, Oct 17, 2018 at 05:44:17PM +0200, Arnd Bergmann wrote:
> > > > On Tue, Oct 16, 2018 at 5:33 AM Guo Ren <ren_guo@c-sky.com> wrote:
> > > > >
> > > > > Support dword access for get_user_size and redesign put_user_size with
> > > > > the same style of get_user_size. It's Ok to use xxx_user_asm_common for
> > > > > all size of variable with ldb, ldh, ldw, ld.d
> > > > >
> > > > > ld.d rx, (addr, 0) could "rx <= addr" "and r(x+1) <= addr+4" and this also
> > > > > follow abiv2 gcc ABI for dword access.
> > > >
> > > > Are you sure this is correct for this?
> > > >
> > > > static inline u32 get_64_to_32(__u64 __user *p)
> > > > {
> > > > u32 ret;
> > > > get_user(ret, p);
> > > > return ret;
> > > > }
> > > >
> > > > If I read __get_user_asm_common() right, the ld.d would overwrite
> > > > two registers, but the caller only expects one, so it clobbers one
> > > > that might be in use.
> > > Ah... BUG! I only consider the get_user(u64, u64 *) :P
> > >
> > > Change to:
> > > case 8: \
> > > __get_user_asm_dword((x), ptr, "ld.d", retval); \
> > > break;
> > >
> > > #define __get_user_asm_dword(x, ptr, err) \
> > > do { \
> > > u64 tmp; \
> > > __get_user_asm_common(tmp, ptr, "ld.d", err); \
> > > x = typeof(x) tmp; \
> > > } while(0)
> > >
> > > #define __put_user_asm_dword(x, ptr, err) \
> > > do { \
> > > u64 tmp = (u64) x; \
> > > __put_user_asm_common(tmp, ptr, "st.d", err); \
> > > } while(0)
> > >
> >
> > I think this will cause warnings for code that passes a pointer.
> >
> > The 64-bit __get_user() is really hard, and most 32-bit architectures don't
> > implement it at all. If you really want to add it, have a look at what
> > x86 and arm do. IIRC they both use __builtin_choose_expr(),
> Thx for the tips and I'll drop the patch first for the upstream.
>
> I want to implement it because of make allmodconfig and
> drivers/android/binder.c need it. I'll learn __builtin_choose_expr()
> and prepare patch next.
I think we should fix binder.c instead. Many other architectures
have the same problem.
Arnd
^ permalink raw reply [flat|nested] 68+ messages in thread* Re: [PATCH V9 21/21] csky: support dword access for get_user_size()
2018-10-24 7:17 ` Arnd Bergmann
@ 2018-10-25 17:08 ` Guo Ren
0 siblings, 0 replies; 68+ messages in thread
From: Guo Ren @ 2018-10-25 17:08 UTC (permalink / raw)
To: Arnd Bergmann
Cc: Andrew Morton, Daniel Lezcano, David Miller, gregkh,
Christoph Hellwig, Marc Zyngier, Mark Rutland, Peter Zijlstra,
Rob Herring, Thomas Gleixner, Linux Kernel Mailing List,
linux-arch, DTML, Rob Herring, c-sky_gcc_upstream
On Wed, Oct 24, 2018 at 08:17:47AM +0100, Arnd Bergmann wrote:
> On Thu, Oct 18, 2018 at 9:57 AM Guo Ren <ren_guo@c-sky.com> wrote:
> >
> > On Thu, Oct 18, 2018 at 10:34:00AM +0200, Arnd Bergmann wrote:
> > > On Thu, Oct 18, 2018 at 5:41 AM Guo Ren <ren_guo@c-sky.com> wrote:
> > > >
> > > > On Wed, Oct 17, 2018 at 05:44:17PM +0200, Arnd Bergmann wrote:
> > > > > On Tue, Oct 16, 2018 at 5:33 AM Guo Ren <ren_guo@c-sky.com> wrote:
> > > > > >
> > > > > > Support dword access for get_user_size and redesign put_user_size with
> > > > > > the same style of get_user_size. It's Ok to use xxx_user_asm_common for
> > > > > > all size of variable with ldb, ldh, ldw, ld.d
> > > > > >
> > > > > > ld.d rx, (addr, 0) could "rx <= addr" "and r(x+1) <= addr+4" and this also
> > > > > > follow abiv2 gcc ABI for dword access.
> > > > >
> > > > > Are you sure this is correct for this?
> > > > >
> > > > > static inline u32 get_64_to_32(__u64 __user *p)
> > > > > {
> > > > > u32 ret;
> > > > > get_user(ret, p);
> > > > > return ret;
> > > > > }
> > > > >
> > > > > If I read __get_user_asm_common() right, the ld.d would overwrite
> > > > > two registers, but the caller only expects one, so it clobbers one
> > > > > that might be in use.
> > > > Ah... BUG! I only consider the get_user(u64, u64 *) :P
> > > >
> > > > Change to:
> > > > case 8: \
> > > > __get_user_asm_dword((x), ptr, "ld.d", retval); \
> > > > break;
> > > >
> > > > #define __get_user_asm_dword(x, ptr, err) \
> > > > do { \
> > > > u64 tmp; \
> > > > __get_user_asm_common(tmp, ptr, "ld.d", err); \
> > > > x = typeof(x) tmp; \
> > > > } while(0)
> > > >
> > > > #define __put_user_asm_dword(x, ptr, err) \
> > > > do { \
> > > > u64 tmp = (u64) x; \
> > > > __put_user_asm_common(tmp, ptr, "st.d", err); \
> > > > } while(0)
> > > >
> > >
> > > I think this will cause warnings for code that passes a pointer.
> > >
> > > The 64-bit __get_user() is really hard, and most 32-bit architectures don't
> > > implement it at all. If you really want to add it, have a look at what
> > > x86 and arm do. IIRC they both use __builtin_choose_expr(),
> > Thx for the tips and I'll drop the patch first for the upstream.
> >
> > I want to implement it because of make allmodconfig and
> > drivers/android/binder.c need it. I'll learn __builtin_choose_expr()
> > and prepare patch next.
>
> I think we should fix binder.c instead. Many other architectures
> have the same problem.
Yes, and I'll study the file.
Best Regards
Guo Ren
^ permalink raw reply [flat|nested] 68+ messages in thread
* Re: [PATCH V9 00/21] C-SKY(csky) Linux Kernel Port
2018-10-16 2:58 [PATCH V9 00/21] C-SKY(csky) Linux Kernel Port Guo Ren
` (20 preceding siblings ...)
2018-10-16 5:48 ` [PATCH V9 21/21] csky: support dword access for get_user_size() Guo Ren
@ 2018-10-17 15:58 ` Arnd Bergmann
2018-10-18 4:10 ` Guo Ren
21 siblings, 1 reply; 68+ messages in thread
From: Arnd Bergmann @ 2018-10-17 15:58 UTC (permalink / raw)
To: Guo Ren
Cc: Andrew Morton, Daniel Lezcano, David Miller, gregkh,
Christoph Hellwig, Marc Zyngier, Mark Rutland, Peter Zijlstra,
Rob Herring, Thomas Gleixner, Linux Kernel Mailing List,
linux-arch, DTML, Rob Herring, c-sky_gcc_upstream
On Tue, Oct 16, 2018 at 4:58 AM Guo Ren <ren_guo@c-sky.com> wrote:
>
> This is the 9th version patchset to add the Linux kernel port for
> C-SKY(csky) based on linux-4.19-rc3.
>
> There are only a few changes between V8 patchset. Hope it could be
> merged into linux-4.20 and I'm very grateful for any help.
I've gone through the entire series once more and saw no show-stoppers.
The last patch looked like it introduced a bug, but with that one dropped,
I'm happy for the architecture to get merged, unless anyone else
has any last-minute concerns. (Alternatively, explain why I'm wrong
and the code works correctly, of course).
I'd appreciate having someone else take another look at the signal
handling code, the atomics, and the DT bindings and provide another
Ack for those.
The remaining open question is about the 32-bit time_t interfaces.
With 4.20, I did not manage to get the required system calls in place
for using 64-bit time_t in a new architecture, so you will at least
start out using 32-bit time_t and likely have to keep supporting
that going forward, unless we decide to break the ABI here later
on .This is something we normally don't do, but we might make
an exception here, under the assumption that there are no
existing users with the ABI. We can debate that once we get there.
Arnd
^ permalink raw reply [flat|nested] 68+ messages in thread* Re: [PATCH V9 00/21] C-SKY(csky) Linux Kernel Port
2018-10-17 15:58 ` [PATCH V9 00/21] C-SKY(csky) Linux Kernel Port Arnd Bergmann
@ 2018-10-18 4:10 ` Guo Ren
2018-10-18 8:36 ` Arnd Bergmann
0 siblings, 1 reply; 68+ messages in thread
From: Guo Ren @ 2018-10-18 4:10 UTC (permalink / raw)
To: Arnd Bergmann
Cc: Andrew Morton, Daniel Lezcano, David Miller, gregkh,
Christoph Hellwig, Marc Zyngier, Mark Rutland, Peter Zijlstra,
Rob Herring, Thomas Gleixner, Linux Kernel Mailing List,
linux-arch, DTML, Rob Herring, c-sky_gcc_upstream
On Wed, Oct 17, 2018 at 05:58:46PM +0200, Arnd Bergmann wrote:
> On Tue, Oct 16, 2018 at 4:58 AM Guo Ren <ren_guo@c-sky.com> wrote:
> >
> > This is the 9th version patchset to add the Linux kernel port for
> > C-SKY(csky) based on linux-4.19-rc3.
> >
> > There are only a few changes between V8 patchset. Hope it could be
> > merged into linux-4.20 and I'm very grateful for any help.
>
> I've gone through the entire series once more and saw no show-stoppers.
> The last patch looked like it introduced a bug, but with that one dropped,
> I'm happy for the architecture to get merged, unless anyone else
> has any last-minute concerns. (Alternatively, explain why I'm wrong
> and the code works correctly, of course).
Ok and thx for the job of csky subsystem.
>
> I'd appreciate having someone else take another look at the signal
> handling code, the atomics, and the DT bindings and provide another
> Ack for those.
>
> The remaining open question is about the 32-bit time_t interfaces.
> With 4.20, I did not manage to get the required system calls in place
> for using 64-bit time_t in a new architecture, so you will at least
> start out using 32-bit time_t and likely have to keep supporting
> that going forward, unless we decide to break the ABI here later
> on .This is something we normally don't do, but we might make
> an exception here, under the assumption that there are no
> existing users with the ABI. We can debate that once we get there.
We support uclibc-ng and glibc.
1. For uclibc-ng, linux-4.20 could run with it.
2. For glibc, Maybe we could support 32-bit + 64-bit time_t with
KERNEL_VERSION, or just only 64-bit then linux-4.20 couldn't work with
the csky first glibc release.
Best Regards
Guo Ren
^ permalink raw reply [flat|nested] 68+ messages in thread
* Re: [PATCH V9 00/21] C-SKY(csky) Linux Kernel Port
2018-10-18 4:10 ` Guo Ren
@ 2018-10-18 8:36 ` Arnd Bergmann
2018-10-18 9:03 ` Guo Ren
0 siblings, 1 reply; 68+ messages in thread
From: Arnd Bergmann @ 2018-10-18 8:36 UTC (permalink / raw)
To: Guo Ren
Cc: Andrew Morton, Daniel Lezcano, David Miller, gregkh,
Christoph Hellwig, Marc Zyngier, Mark Rutland, Peter Zijlstra,
Rob Herring, Thomas Gleixner, Linux Kernel Mailing List,
linux-arch, DTML, Rob Herring, c-sky_gcc_upstream
On Thu, Oct 18, 2018 at 6:11 AM Guo Ren <ren_guo@c-sky.com> wrote:
>
> On Wed, Oct 17, 2018 at 05:58:46PM +0200, Arnd Bergmann wrote:
> > On Tue, Oct 16, 2018 at 4:58 AM Guo Ren <ren_guo@c-sky.com> wrote:
> > >
> > > This is the 9th version patchset to add the Linux kernel port for
> > > C-SKY(csky) based on linux-4.19-rc3.
> > >
> > > There are only a few changes between V8 patchset. Hope it could be
> > > merged into linux-4.20 and I'm very grateful for any help.
> >
> > I've gone through the entire series once more and saw no show-stoppers.
> > The last patch looked like it introduced a bug, but with that one dropped,
> > I'm happy for the architecture to get merged, unless anyone else
> > has any last-minute concerns. (Alternatively, explain why I'm wrong
> > and the code works correctly, of course).
> Ok and thx for the job of csky subsystem.
>
> >
> > I'd appreciate having someone else take another look at the signal
> > handling code, the atomics, and the DT bindings and provide another
> > Ack for those.
> >
> > The remaining open question is about the 32-bit time_t interfaces.
> > With 4.20, I did not manage to get the required system calls in place
> > for using 64-bit time_t in a new architecture, so you will at least
> > start out using 32-bit time_t and likely have to keep supporting
> > that going forward, unless we decide to break the ABI here later
> > on .This is something we normally don't do, but we might make
> > an exception here, under the assumption that there are no
> > existing users with the ABI. We can debate that once we get there.
> We support uclibc-ng and glibc.
>
> 1. For uclibc-ng, linux-4.20 could run with it.
>
> 2. For glibc, Maybe we could support 32-bit + 64-bit time_t with
> KERNEL_VERSION, or just only 64-bit then linux-4.20 couldn't work with
> the csky first glibc release.
Yes, it is always an option to make glibc more restrictive than the kernel.
We could also just make it a configuration option in the kernel whether
the system calls are provided, so they don't use memory for the
implementation.
You will probably want musl support at some point. musl-1.x always
uses 32-bit time_t today, but musl-2.x will use the 64-bit interfaces,
so just waiting a bit will probably make it work out for you.
Arnd
^ permalink raw reply [flat|nested] 68+ messages in thread* Re: [PATCH V9 00/21] C-SKY(csky) Linux Kernel Port
2018-10-18 8:36 ` Arnd Bergmann
@ 2018-10-18 9:03 ` Guo Ren
0 siblings, 0 replies; 68+ messages in thread
From: Guo Ren @ 2018-10-18 9:03 UTC (permalink / raw)
To: Arnd Bergmann
Cc: Andrew Morton, Daniel Lezcano, David Miller, gregkh,
Christoph Hellwig, Marc Zyngier, Mark Rutland, Peter Zijlstra,
Rob Herring, Thomas Gleixner, Linux Kernel Mailing List,
linux-arch, DTML, Rob Herring, c-sky_gcc_upstream
On Thu, Oct 18, 2018 at 10:36:45AM +0200, Arnd Bergmann wrote:
> On Thu, Oct 18, 2018 at 6:11 AM Guo Ren <ren_guo@c-sky.com> wrote:
> >
> > On Wed, Oct 17, 2018 at 05:58:46PM +0200, Arnd Bergmann wrote:
> > > On Tue, Oct 16, 2018 at 4:58 AM Guo Ren <ren_guo@c-sky.com> wrote:
> > > >
> > > > This is the 9th version patchset to add the Linux kernel port for
> > > > C-SKY(csky) based on linux-4.19-rc3.
> > > >
> > > > There are only a few changes between V8 patchset. Hope it could be
> > > > merged into linux-4.20 and I'm very grateful for any help.
> > >
> > > I've gone through the entire series once more and saw no show-stoppers.
> > > The last patch looked like it introduced a bug, but with that one dropped,
> > > I'm happy for the architecture to get merged, unless anyone else
> > > has any last-minute concerns. (Alternatively, explain why I'm wrong
> > > and the code works correctly, of course).
> > Ok and thx for the job of csky subsystem.
> >
> > >
> > > I'd appreciate having someone else take another look at the signal
> > > handling code, the atomics, and the DT bindings and provide another
> > > Ack for those.
> > >
> > > The remaining open question is about the 32-bit time_t interfaces.
> > > With 4.20, I did not manage to get the required system calls in place
> > > for using 64-bit time_t in a new architecture, so you will at least
> > > start out using 32-bit time_t and likely have to keep supporting
> > > that going forward, unless we decide to break the ABI here later
> > > on .This is something we normally don't do, but we might make
> > > an exception here, under the assumption that there are no
> > > existing users with the ABI. We can debate that once we get there.
> > We support uclibc-ng and glibc.
> >
> > 1. For uclibc-ng, linux-4.20 could run with it.
> >
> > 2. For glibc, Maybe we could support 32-bit + 64-bit time_t with
> > KERNEL_VERSION, or just only 64-bit then linux-4.20 couldn't work with
> > the csky first glibc release.
>
> Yes, it is always an option to make glibc more restrictive than the kernel.
> We could also just make it a configuration option in the kernel whether
> the system calls are provided, so they don't use memory for the
> implementation.
Ok.
> You will probably want musl support at some point. musl-1.x always
> uses 32-bit time_t today, but musl-2.x will use the 64-bit interfaces,
> so just waiting a bit will probably make it work out for you.
Thx for the tips, we'll consider musl in the future.
Best Regards
Guo Ren
^ permalink raw reply [flat|nested] 68+ messages in thread