[PATCH AUTOSEL 4.14 24/33] ARCv2: lib: memcpy: fix doing prefetchw outside of buffer

[PATCH AUTOSEL 4.14 24/33] ARCv2: lib: memcpy: fix doing prefetchw outside of buffer

[Sasha Levin]

From: Eugeniy Paltsev <eugeniy.paltsev@synopsys.com>

[ Upstream commit f8a15f97664178f27dfbf86a38f780a532cb6df0 ]

ARCv2 optimized memcpy uses PREFETCHW instruction for prefetching the
next cache line but doesn't ensure that the line is not past the end of
the buffer. PRETECHW changes the line ownership and marks it dirty,
which can cause data corruption if this area is used for DMA IO.

Fix the issue by avoiding the PREFETCHW. This leads to performance
degradation but it is OK as we'll introduce new memcpy implementation
optimized for unaligned memory access using.

We also cut off all PREFETCH instructions at they are quite useless
here:
 * we call PREFETCH right before LOAD instruction call.
 * we copy 16 or 32 bytes of data (depending on CONFIG_ARC_HAS_LL64)
   in a main logical loop. so we call PREFETCH 4 times (or 2 times)
   for each L1 cache line (in case of 64B L1 cache Line which is
   default case). Obviously this is not optimal.

Signed-off-by: Eugeniy Paltsev <Eugeniy.Paltsev at synopsys.com>
Signed-off-by: Vineet Gupta <vgupta at synopsys.com>
Signed-off-by: Sasha Levin <sashal at kernel.org>
---
 arch/arc/lib/memcpy-archs.S | 14 --------------
 1 file changed, 14 deletions(-)

diff --git a/arch/arc/lib/memcpy-archs.S b/arch/arc/lib/memcpy-archs.S
index d61044dd8b58..ea14b0bf3116 100644
--- a/arch/arc/lib/memcpy-archs.S
+++ b/arch/arc/lib/memcpy-archs.S
@@ -25,15 +25,11 @@
 #endif
 
 #ifdef CONFIG_ARC_HAS_LL64
-# define PREFETCH_READ(RX)	prefetch    [RX, 56]
-# define PREFETCH_WRITE(RX)	prefetchw   [RX, 64]
 # define LOADX(DST,RX)		ldd.ab	DST, [RX, 8]
 # define STOREX(SRC,RX)		std.ab	SRC, [RX, 8]
 # define ZOLSHFT		5
 # define ZOLAND			0x1F
 #else
-# define PREFETCH_READ(RX)	prefetch    [RX, 28]
-# define PREFETCH_WRITE(RX)	prefetchw   [RX, 32]
 # define LOADX(DST,RX)		ld.ab	DST, [RX, 4]
 # define STOREX(SRC,RX)		st.ab	SRC, [RX, 4]
 # define ZOLSHFT		4
@@ -41,8 +37,6 @@
 #endif
 
 ENTRY_CFI(memcpy)
-	prefetch [r1]		; Prefetch the read location
-	prefetchw [r0]		; Prefetch the write location
 	mov.f	0, r2
 ;;; if size is zero
 	jz.d	[blink]
@@ -72,8 +66,6 @@ ENTRY_CFI(memcpy)
 	lpnz	@.Lcopy32_64bytes
 	;; LOOP START
 	LOADX (r6, r1)
-	PREFETCH_READ (r1)
-	PREFETCH_WRITE (r3)
 	LOADX (r8, r1)
 	LOADX (r10, r1)
 	LOADX (r4, r1)
@@ -117,9 +109,7 @@ ENTRY_CFI(memcpy)
 	lpnz	@.Lcopy8bytes_1
 	;; LOOP START
 	ld.ab	r6, [r1, 4]
-	prefetch [r1, 28]	;Prefetch the next read location
 	ld.ab	r8, [r1,4]
-	prefetchw [r3, 32]	;Prefetch the next write location
 
 	SHIFT_1	(r7, r6, 24)
 	or	r7, r7, r5
@@ -162,9 +152,7 @@ ENTRY_CFI(memcpy)
 	lpnz	@.Lcopy8bytes_2
 	;; LOOP START
 	ld.ab	r6, [r1, 4]
-	prefetch [r1, 28]	;Prefetch the next read location
 	ld.ab	r8, [r1,4]
-	prefetchw [r3, 32]	;Prefetch the next write location
 
 	SHIFT_1	(r7, r6, 16)
 	or	r7, r7, r5
@@ -204,9 +192,7 @@ ENTRY_CFI(memcpy)
 	lpnz	@.Lcopy8bytes_3
 	;; LOOP START
 	ld.ab	r6, [r1, 4]
-	prefetch [r1, 28]	;Prefetch the next read location
 	ld.ab	r8, [r1,4]
-	prefetchw [r3, 32]	;Prefetch the next write location
 
 	SHIFT_1	(r7, r6, 8)
 	or	r7, r7, r5
-- 
2.19.1

[PATCH AUTOSEL 4.14 25/33] ARC: uacces: remove lp_start, lp_end from clobber list

[Sasha Levin]

From: Vineet Gupta <vgupta@synopsys.com>

[ Upstream commit d5e3c55e01d8b1774b37b4647c30fb22f1d39077 ]

Newer ARC gcc handles lp_start, lp_end in a different way and doesn't
like them in the clobber list.

Signed-off-by: Vineet Gupta <vgupta at synopsys.com>
Signed-off-by: Sasha Levin <sashal at kernel.org>
---
 arch/arc/include/asm/uaccess.h | 8 ++++----
 1 file changed, 4 insertions(+), 4 deletions(-)

diff --git a/arch/arc/include/asm/uaccess.h b/arch/arc/include/asm/uaccess.h
index c9173c02081c..eabc3efa6c6d 100644
--- a/arch/arc/include/asm/uaccess.h
+++ b/arch/arc/include/asm/uaccess.h
@@ -207,7 +207,7 @@ raw_copy_from_user(void *to, const void __user *from, unsigned long n)
 		*/
 		  "=&r" (tmp), "+r" (to), "+r" (from)
 		:
-		: "lp_count", "lp_start", "lp_end", "memory");
+		: "lp_count", "memory");
 
 		return n;
 	}
@@ -433,7 +433,7 @@ raw_copy_to_user(void __user *to, const void *from, unsigned long n)
 		 */
 		  "=&r" (tmp), "+r" (to), "+r" (from)
 		:
-		: "lp_count", "lp_start", "lp_end", "memory");
+		: "lp_count", "memory");
 
 		return n;
 	}
@@ -653,7 +653,7 @@ static inline unsigned long __arc_clear_user(void __user *to, unsigned long n)
 	"	.previous			\n"
 	: "+r"(d_char), "+r"(res)
 	: "i"(0)
-	: "lp_count", "lp_start", "lp_end", "memory");
+	: "lp_count", "memory");
 
 	return res;
 }
@@ -686,7 +686,7 @@ __arc_strncpy_from_user(char *dst, const char __user *src, long count)
 	"	.previous			\n"
 	: "+r"(res), "+r"(dst), "+r"(src), "=r"(val)
 	: "g"(-EFAULT), "r"(count)
-	: "lp_count", "lp_start", "lp_end", "memory");
+	: "lp_count", "memory");
 
 	return res;
 }
-- 
2.19.1

[PATCH AUTOSEL 4.14 26/33] ARCv2: support manual regfile save on interrupts

[Sasha Levin]

From: Vineet Gupta <vgupta@synopsys.com>

[ Upstream commit e494239a007e601448110ac304fe055951f9de3b ]

There's a hardware bug which affects the HSDK platform, triggered by
micro-ops for auto-saving regfile on taken interrupt. The workaround is
to inhibit autosave.

Signed-off-by: Vineet Gupta <vgupta at synopsys.com>
Signed-off-by: Sasha Levin <sashal at kernel.org>
---
 arch/arc/Kconfig                   |  8 +++++
 arch/arc/include/asm/entry-arcv2.h | 54 ++++++++++++++++++++++++++++++
 arch/arc/kernel/entry-arcv2.S      |  4 ++-
 arch/arc/kernel/intc-arcv2.c       |  2 ++
 arch/arc/plat-hsdk/Kconfig         |  1 +
 5 files changed, 68 insertions(+), 1 deletion(-)

diff --git a/arch/arc/Kconfig b/arch/arc/Kconfig
index 9d06c9478a0d..82050893d0b3 100644
--- a/arch/arc/Kconfig
+++ b/arch/arc/Kconfig
@@ -417,6 +417,14 @@ config ARC_HAS_ACCL_REGS
 	  (also referred to as r58:r59). These can also be used by gcc as GPR so
 	  kernel needs to save/restore per process
 
+config ARC_IRQ_NO_AUTOSAVE
+	bool "Disable hardware autosave regfile on interrupts"
+	default n
+	help
+	  On HS cores, taken interrupt auto saves the regfile on stack.
+	  This is programmable and can be optionally disabled in which case
+	  software INTERRUPT_PROLOGUE/EPILGUE do the needed work
+
 endif	# ISA_ARCV2
 
 endmenu   # "ARC CPU Configuration"
diff --git a/arch/arc/include/asm/entry-arcv2.h b/arch/arc/include/asm/entry-arcv2.h
index 257a68f3c2fe..9f581553dcc3 100644
--- a/arch/arc/include/asm/entry-arcv2.h
+++ b/arch/arc/include/asm/entry-arcv2.h
@@ -17,6 +17,33 @@
 	;
 	; Now manually save: r12, sp, fp, gp, r25
 
+#ifdef CONFIG_ARC_IRQ_NO_AUTOSAVE
+.ifnc \called_from, exception
+	st.as	r9, [sp, -10]	; save r9 in it's final stack slot
+	sub	sp, sp, 12	; skip JLI, LDI, EI
+
+	PUSH	lp_count
+	PUSHAX	lp_start
+	PUSHAX	lp_end
+	PUSH	blink
+
+	PUSH	r11
+	PUSH	r10
+
+	sub	sp, sp, 4	; skip r9
+
+	PUSH	r8
+	PUSH	r7
+	PUSH	r6
+	PUSH	r5
+	PUSH	r4
+	PUSH	r3
+	PUSH	r2
+	PUSH	r1
+	PUSH	r0
+.endif
+#endif
+
 #ifdef CONFIG_ARC_HAS_ACCL_REGS
 	PUSH	r59
 	PUSH	r58
@@ -86,6 +113,33 @@
 	POP	r59
 #endif
 
+#ifdef CONFIG_ARC_IRQ_NO_AUTOSAVE
+.ifnc \called_from, exception
+	POP	r0
+	POP	r1
+	POP	r2
+	POP	r3
+	POP	r4
+	POP	r5
+	POP	r6
+	POP	r7
+	POP	r8
+	POP	r9
+	POP	r10
+	POP	r11
+
+	POP	blink
+	POPAX	lp_end
+	POPAX	lp_start
+
+	POP	r9
+	mov	lp_count, r9
+
+	add	sp, sp, 12	; skip JLI, LDI, EI
+	ld.as	r9, [sp, -10]	; reload r9 which got clobbered
+.endif
+#endif
+
 .endm
 
 /*------------------------------------------------------------------------*/
diff --git a/arch/arc/kernel/entry-arcv2.S b/arch/arc/kernel/entry-arcv2.S
index cc558a25b8fa..562089d62d9d 100644
--- a/arch/arc/kernel/entry-arcv2.S
+++ b/arch/arc/kernel/entry-arcv2.S
@@ -209,7 +209,9 @@ restore_regs:
 ;####### Return from Intr #######
 
 debug_marker_l1:
-	bbit1.nt r0, STATUS_DE_BIT, .Lintr_ret_to_delay_slot
+	; bbit1.nt r0, STATUS_DE_BIT, .Lintr_ret_to_delay_slot
+	btst	r0, STATUS_DE_BIT		; Z flag set if bit clear
+	bnz	.Lintr_ret_to_delay_slot	; branch if STATUS_DE_BIT set
 
 .Lisr_ret_fast_path:
 	; Handle special case #1: (Entry via Exception, Return via IRQ)
diff --git a/arch/arc/kernel/intc-arcv2.c b/arch/arc/kernel/intc-arcv2.c
index 067ea362fb3e..cf18b3e5a934 100644
--- a/arch/arc/kernel/intc-arcv2.c
+++ b/arch/arc/kernel/intc-arcv2.c
@@ -49,11 +49,13 @@ void arc_init_IRQ(void)
 
 	*(unsigned int *)&ictrl = 0;
 
+#ifndef CONFIG_ARC_IRQ_NO_AUTOSAVE
 	ictrl.save_nr_gpr_pairs = 6;	/* r0 to r11 (r12 saved manually) */
 	ictrl.save_blink = 1;
 	ictrl.save_lp_regs = 1;		/* LP_COUNT, LP_START, LP_END */
 	ictrl.save_u_to_u = 0;		/* user ctxt saved on kernel stack */
 	ictrl.save_idx_regs = 1;	/* JLI, LDI, EI */
+#endif
 
 	WRITE_AUX(AUX_IRQ_CTRL, ictrl);
 
diff --git a/arch/arc/plat-hsdk/Kconfig b/arch/arc/plat-hsdk/Kconfig
index fcc9a9e27e9c..8fb1600b29b7 100644
--- a/arch/arc/plat-hsdk/Kconfig
+++ b/arch/arc/plat-hsdk/Kconfig
@@ -9,5 +9,6 @@ menuconfig ARC_SOC_HSDK
 	bool "ARC HS Development Kit SOC"
 	depends on ISA_ARCV2
 	select ARC_HAS_ACCL_REGS
+	select ARC_IRQ_NO_AUTOSAVE
 	select CLK_HSDK
 	select RESET_HSDK
-- 
2.19.1