| Message ID | 20160822065524.8085-1-takahiro.akashi@linaro.org (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
On Mon, Aug 22, 2016 at 03:55:24PM +0900, AKASHI Takahiro wrote: > Kdump(kexec-tools) parses /proc/iomem to identify all the memory regions > on the system. Since the current kernel names "nomap" regions, like UEFI > runtime services code/data, as "System RAM," kexec-tools sets up elf core > header to include them in a crash dump file (/proc/vmcore). > > Then crash dump kernel parses UEFI memory map again, re-marks those regions > as "nomap" and does not create a memory mapping for them unlike the other > areas of System RAM. In this case, copying /proc/vmcore through > copy_oldmem_page() on crash dump kernel will end up with a kernel abort, > as reported in [1]. > > This patch names all the "nomap" regions explicitly as "reserved" so that > we can exclude them from a crash dump file. acpi_os_ioremap() must also > be modified because those regions have WB attributes [2]. > > Apart from kdump, this change also matches x86's use of acpi (and > /proc/iomem). > > [1] http://lists.infradead.org/pipermail/linux-arm-kernel/2016-August/448186.html > [2] http://lists.infradead.org/pipermail/linux-arm-kernel/2016-August/450089.html > > Signed-off-by: AKASHI Takahiro <takahiro.akashi@linaro.org> > --- > arch/arm64/include/asm/acpi.h | 8 ++++++-- > arch/arm64/kernel/setup.c | 9 +++++++-- > 2 files changed, 13 insertions(+), 4 deletions(-) The patch looks fine to me but it would be good if James gave it a try. FWIW: Reviewed-by: Catalin Marinas <catalin.marinas@arm.com>
Hi Akashi, On 22/08/16 07:55, AKASHI Takahiro wrote: > Kdump(kexec-tools) parses /proc/iomem to identify all the memory regions > on the system. Since the current kernel names "nomap" regions, like UEFI > runtime services code/data, as "System RAM," kexec-tools sets up elf core > header to include them in a crash dump file (/proc/vmcore). > > Then crash dump kernel parses UEFI memory map again, re-marks those regions > as "nomap" and does not create a memory mapping for them unlike the other > areas of System RAM. In this case, copying /proc/vmcore through > copy_oldmem_page() on crash dump kernel will end up with a kernel abort, > as reported in [1]. > > This patch names all the "nomap" regions explicitly as "reserved" so that > we can exclude them from a crash dump file. acpi_os_ioremap() must also > be modified because those regions have WB attributes [2]. > > Apart from kdump, this change also matches x86's use of acpi (and > /proc/iomem). > > [1] http://lists.infradead.org/pipermail/linux-arm-kernel/2016-August/448186.html > [2] http://lists.infradead.org/pipermail/linux-arm-kernel/2016-August/450089.html Tested on Seattle on v4.8-rc3, /proc/iomem Before: 8000000000-8001e7ffff : System RAM 8001e80000-83ff186fff : System RAM 8002080000-8002baffff : Kernel code 8002ca0000-8002db1fff : Kernel data 83ff187000-83ff1c4fff : System RAM 83ff1c5000-83ff224fff : System RAM 83ff225000-83ffe42fff : System RAM 83ffe43000-83ffffffff : System RAM /proc/iomem After: 8000000000-8001e7ffff : reserved 8001e80000-83ff186fff : System RAM 8002080000-8002baffff : Kernel code 8002ca0000-8002db1fff : Kernel data 83ff187000-83ff1c4fff : reserved 83ff1c5000-83ff224fff : System RAM 83ff225000-83ffe42fff : reserved 83ffe43000-83ffffffff : System RAM Which matches the output from 'efi=debug'. I can also kdump with v24 of your series, and copy the vmcore file with read()/mmap() and dig through it with crash. Tested-by: James Morse <james.morse@arm.com> Reviewed-by: James Morse <james.morse@arm.com> The only side-effect of this I could find is we now allow access to the UEFI areas when built with CONFIG_STRICT_DEVMEM, which is what the help text says should happen, so it was broken before. Thanks, James
diff --git a/arch/arm64/include/asm/acpi.h b/arch/arm64/include/asm/acpi.h index 5420cb0..e517088 100644 --- a/arch/arm64/include/asm/acpi.h +++ b/arch/arm64/include/asm/acpi.h @@ -12,7 +12,7 @@ #ifndef _ASM_ACPI_H #define _ASM_ACPI_H -#include <linux/mm.h> +#include <linux/memblock.h> #include <linux/psci.h> #include <asm/cputype.h> @@ -32,7 +32,11 @@ static inline void __iomem *acpi_os_ioremap(acpi_physical_address phys, acpi_size size) { - if (!page_is_ram(phys >> PAGE_SHIFT)) + /* + * EFI's reserve_regions() call adds memory with the WB attribute + * to memblock via early_init_dt_add_memory_arch(). + */ + if (!memblock_is_memory(phys)) return ioremap(phys, size); return ioremap_cache(phys, size); diff --git a/arch/arm64/kernel/setup.c b/arch/arm64/kernel/setup.c index 38eda13..38589b5 100644 --- a/arch/arm64/kernel/setup.c +++ b/arch/arm64/kernel/setup.c @@ -205,10 +205,15 @@ static void __init request_standard_resources(void) for_each_memblock(memory, region) { res = alloc_bootmem_low(sizeof(*res)); - res->name = "System RAM"; + if (memblock_is_nomap(region)) { + res->name = "reserved"; + res->flags = IORESOURCE_MEM | IORESOURCE_BUSY; + } else { + res->name = "System RAM"; + res->flags = IORESOURCE_SYSTEM_RAM | IORESOURCE_BUSY; + } res->start = __pfn_to_phys(memblock_region_memory_base_pfn(region)); res->end = __pfn_to_phys(memblock_region_memory_end_pfn(region)) - 1; - res->flags = IORESOURCE_SYSTEM_RAM | IORESOURCE_BUSY; request_resource(&iomem_resource, res);
Kdump(kexec-tools) parses /proc/iomem to identify all the memory regions on the system. Since the current kernel names "nomap" regions, like UEFI runtime services code/data, as "System RAM," kexec-tools sets up elf core header to include them in a crash dump file (/proc/vmcore). Then crash dump kernel parses UEFI memory map again, re-marks those regions as "nomap" and does not create a memory mapping for them unlike the other areas of System RAM. In this case, copying /proc/vmcore through copy_oldmem_page() on crash dump kernel will end up with a kernel abort, as reported in [1]. This patch names all the "nomap" regions explicitly as "reserved" so that we can exclude them from a crash dump file. acpi_os_ioremap() must also be modified because those regions have WB attributes [2]. Apart from kdump, this change also matches x86's use of acpi (and /proc/iomem). [1] http://lists.infradead.org/pipermail/linux-arm-kernel/2016-August/448186.html [2] http://lists.infradead.org/pipermail/linux-arm-kernel/2016-August/450089.html Signed-off-by: AKASHI Takahiro <takahiro.akashi@linaro.org> --- arch/arm64/include/asm/acpi.h | 8 ++++++-- arch/arm64/kernel/setup.c | 9 +++++++-- 2 files changed, 13 insertions(+), 4 deletions(-)