| Message ID | 20200711035544.2832154-2-palmer@dabbelt.com (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
| Series | [1/3] lib: Add a generic copy_oldmem_page() | expand |
On Fri, Jul 10, 2020 at 08:55:42PM -0700, Palmer Dabbelt wrote: > +ssize_t copy_oldmem_page(unsigned long pfn, char *buf, > + size_t csize, unsigned long offset, > + int userbuf) > +{ > + void *vaddr; > + > + if (!csize) > + return 0; > + > + vaddr = memremap(__pfn_to_phys(pfn), PAGE_SIZE, MEMREMAP_WB); > + if (!vaddr) > + return -ENOMEM; Doing a memremap for every page is very inefficient. Also I don't see why you'd want to even do that. All memory is in the direct mapping for RISC-V. For other architecture that support highmem kmap_atomic_pfn would do the job, which is what I'd use in a generic version.
On Mon, Jul 13, 2020 at 3:07 PM Christoph Hellwig <hch@infradead.org> wrote: > On Fri, Jul 10, 2020 at 08:55:42PM -0700, Palmer Dabbelt wrote: > > +ssize_t copy_oldmem_page(unsigned long pfn, char *buf, > > + size_t csize, unsigned long offset, > > + int userbuf) > > +{ > > + void *vaddr; > > + > > + if (!csize) > > + return 0; > > + > > + vaddr = memremap(__pfn_to_phys(pfn), PAGE_SIZE, MEMREMAP_WB); > > + if (!vaddr) > > + return -ENOMEM; > > Doing a memremap for every page is very inefficient. Also I don't see > why you'd want to even do that. All memory is in the direct mapping > for RISC-V. For other architecture that support highmem kmap_atomic_pfn > would do the job, which is what I'd use in a generic version. I would expect the 'oldmem' data to not have a 'struct page', which would be a problem at least for the generic implementation of kmap_atomic_pfn() include/linux/highmem.h:#define kmap_atomic_pfn(pfn) kmap_atomic(pfn_to_page(pfn)) kmap_atomic() might still work with a bogus page pointer if it only transforms it back into a pfn, but some implementations just have this one that cannot work: static inline void *page_address(const struct page *page) { return page->virtual; } I have not checked how the crash dump code works though, maybe it does work after all. Arnd
On Mon, Jul 13, 2020 at 03:39:17PM +0200, Arnd Bergmann wrote: > On Mon, Jul 13, 2020 at 3:07 PM Christoph Hellwig <hch@infradead.org> wrote: > > On Fri, Jul 10, 2020 at 08:55:42PM -0700, Palmer Dabbelt wrote: > > > +ssize_t copy_oldmem_page(unsigned long pfn, char *buf, > > > + size_t csize, unsigned long offset, > > > + int userbuf) > > > +{ > > > + void *vaddr; > > > + > > > + if (!csize) > > > + return 0; > > > + > > > + vaddr = memremap(__pfn_to_phys(pfn), PAGE_SIZE, MEMREMAP_WB); > > > + if (!vaddr) > > > + return -ENOMEM; > > > > Doing a memremap for every page is very inefficient. Also I don't see > > why you'd want to even do that. All memory is in the direct mapping > > for RISC-V. For other architecture that support highmem kmap_atomic_pfn > > would do the job, which is what I'd use in a generic version. > > I would expect the 'oldmem' data to not have a 'struct page', which would > be a problem at least for the generic implementation of kmap_atomic_pfn() Why do you expect it to not have a struct page?
On Tue, Jul 14, 2020 at 1:06 PM Christoph Hellwig <hch@infradead.org> wrote: > On Mon, Jul 13, 2020 at 03:39:17PM +0200, Arnd Bergmann wrote: > > On Mon, Jul 13, 2020 at 3:07 PM Christoph Hellwig <hch@infradead.org> wrote: > > > On Fri, Jul 10, 2020 at 08:55:42PM -0700, Palmer Dabbelt wrote: > > > > +ssize_t copy_oldmem_page(unsigned long pfn, char *buf, > > > > + size_t csize, unsigned long offset, > > > > + int userbuf) > > > > +{ > > > > + void *vaddr; > > > > + > > > > + if (!csize) > > > > + return 0; > > > > + > > > > + vaddr = memremap(__pfn_to_phys(pfn), PAGE_SIZE, MEMREMAP_WB); > > > > + if (!vaddr) > > > > + return -ENOMEM; > > > > > > Doing a memremap for every page is very inefficient. Also I don't see > > > why you'd want to even do that. All memory is in the direct mapping > > > for RISC-V. For other architecture that support highmem kmap_atomic_pfn > > > would do the job, which is what I'd use in a generic version. > > > > I would expect the 'oldmem' data to not have a 'struct page', which would > > be a problem at least for the generic implementation of kmap_atomic_pfn() > > Why do you expect it to not have a struct page? I was under the impression that the kdump kernel only accesses a small amount of memory itself and would allocate its mem_map to fit that but not pages of that were used by the kernel it is dumping. Arnd
On Mon, 13 Jul 2020 at 16:06, Christoph Hellwig <hch@infradead.org> wrote: > > On Fri, Jul 10, 2020 at 08:55:42PM -0700, Palmer Dabbelt wrote: > > +ssize_t copy_oldmem_page(unsigned long pfn, char *buf, > > + size_t csize, unsigned long offset, > > + int userbuf) > > +{ > > + void *vaddr; > > + > > + if (!csize) > > + return 0; > > + > > + vaddr = memremap(__pfn_to_phys(pfn), PAGE_SIZE, MEMREMAP_WB); > > + if (!vaddr) > > + return -ENOMEM; > > Doing a memremap for every page is very inefficient. memremap(MEMREMAP_WB) will reuse the existing linear address if the PA is covered by the linear mapping. > Also I don't see > why you'd want to even do that. All memory is in the direct mapping > for RISC-V. For other architecture that support highmem kmap_atomic_pfn > would do the job, which is what I'd use in a generic version.
diff --git a/lib/Kconfig b/lib/Kconfig index df3f3da95990..25544abc9547 100644 --- a/lib/Kconfig +++ b/lib/Kconfig @@ -676,3 +676,6 @@ config GENERIC_LIB_CMPDI2 config GENERIC_LIB_UCMPDI2 bool + +config GENERIC_LIB_COPY_OLDMEM_PAGE + bool diff --git a/lib/Makefile b/lib/Makefile index b1c42c10073b..30d57d8b32b1 100644 --- a/lib/Makefile +++ b/lib/Makefile @@ -318,3 +318,5 @@ obj-$(CONFIG_OBJAGG) += objagg.o # KUnit tests obj-$(CONFIG_LIST_KUNIT_TEST) += list-test.o obj-$(CONFIG_LINEAR_RANGES_TEST) += test_linear_ranges.o + +obj-$(CONFIG_GENERIC_LIB_COPY_OLDMEM_PAGE) += copy_oldmem_page.o diff --git a/lib/copy_oldmem_page.c b/lib/copy_oldmem_page.c new file mode 100644 index 000000000000..f0090027218a --- /dev/null +++ b/lib/copy_oldmem_page.c @@ -0,0 +1,51 @@ +// SPDX-License-Identifier: GPL-2.0-only +/* + * Routines for doing kexec-based kdump + * + * Originally part of arch/arm64/kernel/crash_dump.c + * Copyright (C) 2017 Linaro Limited + * Author: AKASHI Takahiro <takahiro.akashi@linaro.org> + */ + +#include <linux/io.h> +#include <linux/types.h> +#include <linux/uaccess.h> + +/** + * copy_oldmem_page() - copy one page from old kernel memory + * @pfn: page frame number to be copied + * @buf: buffer where the copied page is placed + * @csize: number of bytes to copy + * @offset: offset in bytes into the page + * @userbuf: if set, @buf is in a user address space + * + * This function copies one page from old kernel memory into buffer pointed by + * @buf. If @buf is in userspace, set @userbuf to %1. Returns number of bytes + * copied or negative error in case of failure. + */ +ssize_t copy_oldmem_page(unsigned long pfn, char *buf, + size_t csize, unsigned long offset, + int userbuf) +{ + void *vaddr; + + if (!csize) + return 0; + + vaddr = memremap(__pfn_to_phys(pfn), PAGE_SIZE, MEMREMAP_WB); + if (!vaddr) + return -ENOMEM; + + if (userbuf) { + if (copy_to_user((char __user *)buf, vaddr + offset, csize)) { + memunmap(vaddr); + return -EFAULT; + } + } else { + memcpy(buf, vaddr + offset, csize); + } + + memunmap(vaddr); + + return csize; +}