| Message ID | 20220607125027.44946-1-wangkefeng.wang@huawei.com (mailing list archive) |
|---|---|
| Headers | show |
| Series | arm64: Cleanup ioremap() and support ioremap_prot() | expand |
Hi Catalin, could you help to pick up it, thanks. On 2022/6/7 20:50, Kefeng Wang wrote: > 1. Enhance generic ioremap to make it more useful. > 2. Let's arm64 use GENERIC_IOREMAP to cleanup code. > 3. Support HAVE_IOREMAP_PROT on arm64, which enable generic_access_phys(), > it is useful when debug(eg, gdb) via access_process_vm device memory > infrastructure. > > v5: > - break long lines(> 80 cols), per Christoph Hellwig > - move is_vmalloc_addr() check from arm64 into generic ioremap, per > Christoph Hellwig > - make arm64's ioremap_cache as an inline function, per Christoph > - keep changes simple, make ioremap/iounmap_allowed return bool, per > Baoquan He > - simplify use 'void *' instead of 'void __iomem *' in iounmap, then > drop __force annotation > > v4: > - update based on v5.19-rc1 > - add generic arch_ioremap/arch_iounmap define, per Andrew Monrton > - simply return an int for arch_ioremap and rename arch_ioremap/arch_iounmap > to a better name, ioremap_allowed/iounmap_allowed, per Arnd Bergmann > - add __force annotation to slince sparse warning in vunmap() > > Note, > 1) after the renaming, the arm's change(patch1) is not the necessary > dependence for the following changes, but as a cleanup, still post > it here, hope it go in via the arm64 tree with reset of the series > directly if no object. > 2) the changes in this version only influence on patch4/5, so retain > the ack/review. > > v3: > - add cleanup patch to kill ARM's unused arch_iounmap(the naming will be > used in GENERIC_IOREMAP) and add comments for arch_ioremap/arch_iounmap > hooks, per Anshuman Khandual > - collect ack/review > > v2: > - s/addr/phys_addr in ioremap_prot, suggested by Andrew Morton > - rename arch_ioremap/iounmap_check to arch_ioremap/iounmap > and change return value, per Christoph Hellwig and Andrew Morton > - and use 'ifndef arch_ioremap' instead of weak function, per Arnd Bergmann > - collect ack/review > > Kefeng Wang (6): > ARM: mm: kill unused runtime hook arch_iounmap() > mm: ioremap: Use more sensibly name in ioremap_prot() > mm: ioremap: Setup phys_addr of struct vm_struct > mm: ioremap: Add ioremap/iounmap_allowed() > arm64: mm: Convert to GENERIC_IOREMAP > arm64: Add HAVE_IOREMAP_PROT support > > .../features/vm/ioremap_prot/arch-support.txt | 2 +- > arch/arm/include/asm/io.h | 4 +- > arch/arm/mm/ioremap.c | 9 +- > arch/arm/mm/nommu.c | 9 +- > arch/arm64/Kconfig | 2 + > arch/arm64/include/asm/io.h | 24 +++-- > arch/arm64/include/asm/pgtable.h | 10 +++ > arch/arm64/kernel/acpi.c | 2 +- > arch/arm64/mm/hugetlbpage.c | 10 --- > arch/arm64/mm/ioremap.c | 90 ++----------------- > include/asm-generic/io.h | 29 +++++- > mm/ioremap.c | 26 ++++-- > 12 files changed, 90 insertions(+), 127 deletions(-) >
On 2022/6/14 11:21, Kefeng Wang wrote: > Hi Catalin, could you help to pick up it, thanks. Kindly ping... > > On 2022/6/7 20:50, Kefeng Wang wrote: >> 1. Enhance generic ioremap to make it more useful. >> 2. Let's arm64 use GENERIC_IOREMAP to cleanup code. >> 3. Support HAVE_IOREMAP_PROT on arm64, which enable >> generic_access_phys(), >> it is useful when debug(eg, gdb) via access_process_vm device memory >> infrastructure.
On Mon, Jun 27, 2022 at 07:06:41PM +0800, Kefeng Wang wrote: > > On 2022/6/14 11:21, Kefeng Wang wrote: > > Hi Catalin, could you help to pick up it, thanks. > Kindly ping... Sorry, this is on my plate for 5.20. I'll try to get to it today. Will
On 2022/6/27 19:14, Will Deacon wrote: > On Mon, Jun 27, 2022 at 07:06:41PM +0800, Kefeng Wang wrote: >> On 2022/6/14 11:21, Kefeng Wang wrote: >>> Hi Catalin, could you help to pick up it, thanks. >> Kindly ping... > Sorry, this is on my plate for 5.20. I'll try to get to it today. Good to know it, thank you. > > Will > .
On Tue, 7 Jun 2022 20:50:21 +0800, Kefeng Wang wrote: > 1. Enhance generic ioremap to make it more useful. > 2. Let's arm64 use GENERIC_IOREMAP to cleanup code. > 3. Support HAVE_IOREMAP_PROT on arm64, which enable generic_access_phys(), > it is useful when debug(eg, gdb) via access_process_vm device memory > infrastructure. > > v5: > - break long lines(> 80 cols), per Christoph Hellwig > - move is_vmalloc_addr() check from arm64 into generic ioremap, per > Christoph Hellwig > - make arm64's ioremap_cache as an inline function, per Christoph > - keep changes simple, make ioremap/iounmap_allowed return bool, per > Baoquan He > - simplify use 'void *' instead of 'void __iomem *' in iounmap, then > drop __force annotation > > [...] Applied to arm64 (for-next/ioremap), thanks! [1/6] ARM: mm: kill unused runtime hook arch_iounmap() https://git.kernel.org/arm64/c/d803336abdbc [2/6] mm: ioremap: Use more sensibly name in ioremap_prot() https://git.kernel.org/arm64/c/abc5992b9dd0 [3/6] mm: ioremap: Setup phys_addr of struct vm_struct https://git.kernel.org/arm64/c/a14fff1c0379 [4/6] mm: ioremap: Add ioremap/iounmap_allowed() https://git.kernel.org/arm64/c/18e780b4e6ab [5/6] arm64: mm: Convert to GENERIC_IOREMAP https://git.kernel.org/arm64/c/f23eab0bfaef [6/6] arm64: Add HAVE_IOREMAP_PROT support https://git.kernel.org/arm64/c/893dea9ccd08 Cheers,