* [PATCH v2 00/11] Raspberry Pi 4 DMA addressing support @ 2019-08-20 14:58 Nicolas Saenz Julienne 2019-08-20 14:58 ` [PATCH v2 09/11] dma-direct: turn ARCH_ZONE_DMA_BITS into a variable Nicolas Saenz Julienne 0 siblings, 1 reply; 4+ messages in thread From: Nicolas Saenz Julienne @ 2019-08-20 14:58 UTC (permalink / raw) To: catalin.marinas, hch, wahrenst, marc.zyngier, robh+dt, Robin Murphy, linux-arm-kernel, devicetree, linux-arch, iommu, linux-mm, linux-riscv Cc: phill, linux-s390, f.fainelli, frowand.list, linuxppc-dev, linux-kernel, eric, mbrugger, linux-rpi-kernel, akpm, will, nsaenzjulienne, m.szyprowski Hi all, this series attempts to address some issues we found while bringing up the new Raspberry Pi 4 in arm64 and it's intended to serve as a follow up of these discussions: v1: https://lkml.org/lkml/2019/7/31/922 RFC: https://lkml.org/lkml/2019/7/17/476 The new Raspberry Pi 4 has up to 4GB of memory but most peripherals can only address the first GB: their DMA address range is 0xc0000000-0xfc000000 which is aliased to the first GB of physical memory 0x00000000-0x3c000000. Note that only some peripherals have these limitations: the PCIe, V3D, GENET, and 40-bit DMA channels have a wider view of the address space by virtue of being hooked up trough a second interconnect. Part of this is solved in arm32 by setting up the machine specific '.dma_zone_size = SZ_1G', which takes care of reserving the coherent memory area at the right spot. That said no buffer bouncing (needed for dma streaming) is available at the moment, but that's a story for another series. Unfortunately there is no such thing as 'dma_zone_size' in arm64. Only ZONE_DMA32 is created which is interpreted by dma-direct and the arm64 arch code as if all peripherals where be able to address the first 4GB of memory. In the light of this, the series implements the following changes: - Create generic 'dma_zone_size' in order for hardware description code to set it up when needed. - Add a function in early_init_dt_scan() to setup 'dma_zone_size' for the RPi4. - Create both DMA zones in arm64, ZONE_DMA will contain the area addressable by all peripherals and ZONE_DMA32 the rest of the 32 bit addressable memory. ZONE_DMA32 might be left empty. - Reserve the CMA area in a place suitable for all peripherals. - Inform dma-direct of the new runtime calculated min_mask. This series has been tested on multiple devices both by checking the zones setup matches the expectations and by double-checking physical addresses on pages allocated on the three relevant areas GFP_DMA, GFP_DMA32, GFP_KERNEL: - On an RPi4 with variations on the ram memory size. But also forcing the situation where all three memory zones are nonempty by setting a 3G ZONE_DMA32 ceiling on a 4G setup. Both with and without NUMA support. - On a Synquacer box[1] with 32G of memory. - On an ACPI based Huawei TaiShan server[2] with 256G of memory. - On a QEMU virtual machine running arm64's OpenSUSE Tumbleweed. That's all. Regards, Nicolas [1] https://www.96boards.org/product/developerbox/ [2] https://e.huawei.com/en/products/cloud-computing-dc/servers/taishan-server/taishan-2280-v2 --- Changes in v2: - More in depth testing. - Create new global 'dma_zone_size'. - New approach to getting the dma_zone_size, instead of parsing the dts we hardcode it conditionally to the machine compatible name. - Fix ZONE_DMA and ZONE_DMA32 split, now ZONE_DMA32 remains empty if ZONE_DMA fits the whole 32 bit addressable space. - Take into account devices with DMA offset. - Rename new dma-direct variable to zone_dma_bits. - Try new approach by merging both ZONE_DMA and ZONE_DMA32 comments in mmzone.h, add new up to date examples. Nicolas Saenz Julienne (11): asm-generic: add dma_zone_size arm: use generic dma_zone_size of/fdt: add of_fdt_machine_is_compatible function of/fdt: add early_init_dt_get_dma_zone_size() arm64: mm: use arm64_dma_phys_limit instead of calling max_zone_dma_phys() arm64: rename variables used to calculate ZONE_DMA32's size arm64: re-introduce max_zone_dma_phys() arm64: use both ZONE_DMA and ZONE_DMA32 dma-direct: turn ARCH_ZONE_DMA_BITS into a variable arm64: edit zone_dma_bits to fine tune dma-direct min mask mm: refresh ZONE_DMA and ZONE_DMA32 comments in 'enum zone_type' arch/arm/include/asm/dma.h | 8 ++-- arch/arm/mm/init.c | 12 ++---- arch/arm64/Kconfig | 4 ++ arch/arm64/mm/init.c | 73 +++++++++++++++++++++++++-------- arch/powerpc/include/asm/page.h | 9 ---- arch/powerpc/mm/mem.c | 16 +++++--- arch/s390/include/asm/page.h | 2 - arch/s390/mm/init.c | 1 + drivers/of/fdt.c | 15 +++++++ include/asm-generic/dma.h | 8 +++- include/linux/dma-direct.h | 2 + include/linux/mmzone.h | 46 ++++++++++++--------- kernel/dma/direct.c | 13 +++--- mm/page_alloc.c | 3 ++ 14 files changed, 140 insertions(+), 72 deletions(-) -- 2.22.0 ^ permalink raw reply [flat|nested] 4+ messages in thread
* [PATCH v2 09/11] dma-direct: turn ARCH_ZONE_DMA_BITS into a variable 2019-08-20 14:58 [PATCH v2 00/11] Raspberry Pi 4 DMA addressing support Nicolas Saenz Julienne @ 2019-08-20 14:58 ` Nicolas Saenz Julienne 2019-08-26 7:05 ` Christoph Hellwig 2019-08-26 11:33 ` Michael Ellerman 0 siblings, 2 replies; 4+ messages in thread From: Nicolas Saenz Julienne @ 2019-08-20 14:58 UTC (permalink / raw) To: catalin.marinas, hch, wahrenst, marc.zyngier, robh+dt, Robin Murphy, linux-arm-kernel, devicetree, linux-arch, iommu, linux-mm, linux-riscv, Marek Szyprowski Cc: phill, linux-s390, f.fainelli, Vasily Gorbik, Christian Borntraeger, frowand.list, linuxppc-dev, Heiko Carstens, linux-kernel, eric, mbrugger, Paul Mackerras, linux-rpi-kernel, akpm, will, nsaenzjulienne Some architectures, notably arm64, are interested in tweaking this depending on their runtime dma addressing limitations. Signed-off-by: Nicolas Saenz Julienne <nsaenzjulienne@suse.de> --- Changes in v2: - Rename new variable to zone_dma_bits - Update comment with Christoph's suggestion - Remove old powerpc comment arch/powerpc/include/asm/page.h | 9 --------- arch/powerpc/mm/mem.c | 16 +++++++++++----- arch/s390/include/asm/page.h | 2 -- arch/s390/mm/init.c | 1 + include/linux/dma-direct.h | 2 ++ kernel/dma/direct.c | 13 ++++++------- 6 files changed, 20 insertions(+), 23 deletions(-) diff --git a/arch/powerpc/include/asm/page.h b/arch/powerpc/include/asm/page.h index 0d52f57fca04..73668a21ae78 100644 --- a/arch/powerpc/include/asm/page.h +++ b/arch/powerpc/include/asm/page.h @@ -319,13 +319,4 @@ struct vm_area_struct; #endif /* __ASSEMBLY__ */ #include <asm/slice.h> -/* - * Allow 30-bit DMA for very limited Broadcom wifi chips on many powerbooks. - */ -#ifdef CONFIG_PPC32 -#define ARCH_ZONE_DMA_BITS 30 -#else -#define ARCH_ZONE_DMA_BITS 31 -#endif - #endif /* _ASM_POWERPC_PAGE_H */ diff --git a/arch/powerpc/mm/mem.c b/arch/powerpc/mm/mem.c index 9191a66b3bc5..2a69f87585df 100644 --- a/arch/powerpc/mm/mem.c +++ b/arch/powerpc/mm/mem.c @@ -31,6 +31,7 @@ #include <linux/slab.h> #include <linux/vmalloc.h> #include <linux/memremap.h> +#include <linux/dma-direct.h> #include <asm/pgalloc.h> #include <asm/prom.h> @@ -201,10 +202,10 @@ static int __init mark_nonram_nosave(void) * everything else. GFP_DMA32 page allocations automatically fall back to * ZONE_DMA. * - * By using 31-bit unconditionally, we can exploit ARCH_ZONE_DMA_BITS to - * inform the generic DMA mapping code. 32-bit only devices (if not handled - * by an IOMMU anyway) will take a first dip into ZONE_NORMAL and get - * otherwise served by ZONE_DMA. + * By using 31-bit unconditionally, we can exploit zone_dma_bits to inform the + * generic DMA mapping code. 32-bit only devices (if not handled by an IOMMU + * anyway) will take a first dip into ZONE_NORMAL and get otherwise served by + * ZONE_DMA. */ static unsigned long max_zone_pfns[MAX_NR_ZONES]; @@ -237,9 +238,14 @@ void __init paging_init(void) printk(KERN_DEBUG "Memory hole size: %ldMB\n", (long int)((top_of_ram - total_ram) >> 20)); + if (IS_ENABLED(CONFIG_PPC32)) + zone_dma_bits = 30; + else + zone_dma_bits = 31; + #ifdef CONFIG_ZONE_DMA max_zone_pfns[ZONE_DMA] = min(max_low_pfn, - 1UL << (ARCH_ZONE_DMA_BITS - PAGE_SHIFT)); + 1UL << (zone_dma_bits - PAGE_SHIFT)); #endif max_zone_pfns[ZONE_NORMAL] = max_low_pfn; #ifdef CONFIG_HIGHMEM diff --git a/arch/s390/include/asm/page.h b/arch/s390/include/asm/page.h index 823578c6b9e2..a4d38092530a 100644 --- a/arch/s390/include/asm/page.h +++ b/arch/s390/include/asm/page.h @@ -177,8 +177,6 @@ static inline int devmem_is_allowed(unsigned long pfn) #define VM_DATA_DEFAULT_FLAGS (VM_READ | VM_WRITE | \ VM_MAYREAD | VM_MAYWRITE | VM_MAYEXEC) -#define ARCH_ZONE_DMA_BITS 31 - #include <asm-generic/memory_model.h> #include <asm-generic/getorder.h> diff --git a/arch/s390/mm/init.c b/arch/s390/mm/init.c index 20340a03ad90..bd98465b8b9f 100644 --- a/arch/s390/mm/init.c +++ b/arch/s390/mm/init.c @@ -118,6 +118,7 @@ void __init paging_init(void) sparse_memory_present_with_active_regions(MAX_NUMNODES); sparse_init(); + zone_dma_bits = 31; memset(max_zone_pfns, 0, sizeof(max_zone_pfns)); max_zone_pfns[ZONE_DMA] = PFN_DOWN(MAX_DMA_ADDRESS); max_zone_pfns[ZONE_NORMAL] = max_low_pfn; diff --git a/include/linux/dma-direct.h b/include/linux/dma-direct.h index adf993a3bd58..d03af3605460 100644 --- a/include/linux/dma-direct.h +++ b/include/linux/dma-direct.h @@ -5,6 +5,8 @@ #include <linux/dma-mapping.h> #include <linux/mem_encrypt.h> +extern unsigned int zone_dma_bits; + #ifdef CONFIG_ARCH_HAS_PHYS_TO_DMA #include <asm/dma-direct.h> #else diff --git a/kernel/dma/direct.c b/kernel/dma/direct.c index 795c9b095d75..b23cd65f26e0 100644 --- a/kernel/dma/direct.c +++ b/kernel/dma/direct.c @@ -16,12 +16,11 @@ #include <linux/swiotlb.h> /* - * Most architectures use ZONE_DMA for the first 16 Megabytes, but - * some use it for entirely different regions: + * Most architectures use ZONE_DMA for the first 16 Megabytes, but some use it + * it for entirely different regions. In that case the arch code needs to + * override the variable below for dma-direct to work properly. */ -#ifndef ARCH_ZONE_DMA_BITS -#define ARCH_ZONE_DMA_BITS 24 -#endif +unsigned int zone_dma_bits __ro_after_init = 24; static void report_addr(struct device *dev, dma_addr_t dma_addr, size_t size) { @@ -69,7 +68,7 @@ static gfp_t __dma_direct_optimal_gfp_mask(struct device *dev, u64 dma_mask, * Note that GFP_DMA32 and GFP_DMA are no ops without the corresponding * zones. */ - if (*phys_mask <= DMA_BIT_MASK(ARCH_ZONE_DMA_BITS)) + if (*phys_mask <= DMA_BIT_MASK(zone_dma_bits)) return GFP_DMA; if (*phys_mask <= DMA_BIT_MASK(32)) return GFP_DMA32; @@ -387,7 +386,7 @@ int dma_direct_supported(struct device *dev, u64 mask) u64 min_mask; if (IS_ENABLED(CONFIG_ZONE_DMA)) - min_mask = DMA_BIT_MASK(ARCH_ZONE_DMA_BITS); + min_mask = DMA_BIT_MASK(zone_dma_bits); else min_mask = DMA_BIT_MASK(32); -- 2.22.0 ^ permalink raw reply related [flat|nested] 4+ messages in thread
* Re: [PATCH v2 09/11] dma-direct: turn ARCH_ZONE_DMA_BITS into a variable 2019-08-20 14:58 ` [PATCH v2 09/11] dma-direct: turn ARCH_ZONE_DMA_BITS into a variable Nicolas Saenz Julienne @ 2019-08-26 7:05 ` Christoph Hellwig 2019-08-26 11:33 ` Michael Ellerman 1 sibling, 0 replies; 4+ messages in thread From: Christoph Hellwig @ 2019-08-26 7:05 UTC (permalink / raw) To: Nicolas Saenz Julienne Cc: catalin.marinas, Heiko Carstens, eric, Paul Mackerras, linux-riscv, will, hch, Marek Szyprowski, linux-arch, linux-s390, f.fainelli, frowand.list, Christian Borntraeger, devicetree, Vasily Gorbik, marc.zyngier, robh+dt, linux-rpi-kernel, linux-arm-kernel, phill, mbrugger, linux-mm, linuxppc-dev, linux-kernel, iommu, wahrenst, akpm, Robin Murphy Looks good, Reviewed-by: Christoph Hellwig <hch@lst.de> ^ permalink raw reply [flat|nested] 4+ messages in thread
* Re: [PATCH v2 09/11] dma-direct: turn ARCH_ZONE_DMA_BITS into a variable 2019-08-20 14:58 ` [PATCH v2 09/11] dma-direct: turn ARCH_ZONE_DMA_BITS into a variable Nicolas Saenz Julienne 2019-08-26 7:05 ` Christoph Hellwig @ 2019-08-26 11:33 ` Michael Ellerman 1 sibling, 0 replies; 4+ messages in thread From: Michael Ellerman @ 2019-08-26 11:33 UTC (permalink / raw) To: Nicolas Saenz Julienne, catalin.marinas, hch, wahrenst, marc.zyngier, robh+dt, Robin Murphy, linux-arm-kernel, devicetree, linux-arch, iommu, linux-mm, linux-riscv, Marek Szyprowski Cc: linux-s390, f.fainelli, Vasily Gorbik, Christian Borntraeger, frowand.list, linuxppc-dev, Heiko Carstens, linux-kernel, eric, mbrugger, Paul Mackerras, linux-rpi-kernel, akpm, will, nsaenzjulienne Nicolas Saenz Julienne <nsaenzjulienne@suse.de> writes: > diff --git a/arch/powerpc/include/asm/page.h b/arch/powerpc/include/asm/page.h > index 0d52f57fca04..73668a21ae78 100644 > --- a/arch/powerpc/include/asm/page.h > +++ b/arch/powerpc/include/asm/page.h > @@ -319,13 +319,4 @@ struct vm_area_struct; > #endif /* __ASSEMBLY__ */ > #include <asm/slice.h> > > -/* > - * Allow 30-bit DMA for very limited Broadcom wifi chips on many powerbooks. This comment got lost. > diff --git a/arch/powerpc/mm/mem.c b/arch/powerpc/mm/mem.c > index 9191a66b3bc5..2a69f87585df 100644 > --- a/arch/powerpc/mm/mem.c > +++ b/arch/powerpc/mm/mem.c > @@ -237,9 +238,14 @@ void __init paging_init(void) > printk(KERN_DEBUG "Memory hole size: %ldMB\n", > (long int)((top_of_ram - total_ram) >> 20)); > > + if (IS_ENABLED(CONFIG_PPC32)) Can you please propagate it here? > + zone_dma_bits = 30; > + else > + zone_dma_bits = 31; > + cheers ^ permalink raw reply [flat|nested] 4+ messages in thread
end of thread, other threads:[~2019-08-26 11:35 UTC | newest] Thread overview: 4+ messages (download: mbox.gz / follow: Atom feed) -- links below jump to the message on this page -- 2019-08-20 14:58 [PATCH v2 00/11] Raspberry Pi 4 DMA addressing support Nicolas Saenz Julienne 2019-08-20 14:58 ` [PATCH v2 09/11] dma-direct: turn ARCH_ZONE_DMA_BITS into a variable Nicolas Saenz Julienne 2019-08-26 7:05 ` Christoph Hellwig 2019-08-26 11:33 ` Michael Ellerman
This is a public inbox, see mirroring instructions for how to clone and mirror all data and code used for this inbox; as well as URLs for NNTP newsgroup(s).