| Message ID | 20200124161810.24322-1-guoren@linux.alibaba.com (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
| Series | riscv: Use flush_icache_mm for flush_icache_user_range | expand |
On Fri, 24 Jan 2020 16:18:10 GMT (+0000), guoren@linux.alibaba.com wrote: > The only call path is: > > __access_remote_vm -> copy_to_user_page -> flush_icache_user_range > > Seems it's ok to use flush_icache_mm instead of flush_icache_all and > it could reduce flush_icache_all called on other harts. > > I think the patch is the fixup for the commit 08f051eda33b. > > Signed-off-by: Guo Ren <guoren@linux.alibaba.com> > Cc: Andrew Waterman <andrew@sifive.com> > Cc: Palmer Dabbelt <palmer@sifive.com> > --- > arch/riscv/include/asm/cacheflush.h | 2 +- > 1 file changed, 1 insertion(+), 1 deletion(-) > > diff --git a/arch/riscv/include/asm/cacheflush.h b/arch/riscv/include/asm/cacheflush.h > index b69aecbb36d3..26589623fd57 100644 > --- a/arch/riscv/include/asm/cacheflush.h > +++ b/arch/riscv/include/asm/cacheflush.h > @@ -85,7 +85,7 @@ static inline void flush_dcache_page(struct page *page) > * so instead we just flush the whole thing. > */ > #define flush_icache_range(start, end) flush_icache_all() > -#define flush_icache_user_range(vma, pg, addr, len) flush_icache_all() > +#define flush_icache_user_range(vma, pg, addr, len) flush_icache_mm(vma->vm_mm, 0) > > void dma_wbinv_range(unsigned long start, unsigned long end); > void dma_wb_range(unsigned long start, unsigned long end); Reviewed-by: Palmer Dabbelt <palmerdabbelt@google.com> I've added this to for-next with some minor modifications as 4d99abce8ce80e866020ffa5b2bd790269235f37. It missed the PR I'm sending now-ish, but I'll include it as part of the next one even if it's during an early RC. Thanks!
diff --git a/arch/riscv/include/asm/cacheflush.h b/arch/riscv/include/asm/cacheflush.h index b69aecbb36d3..26589623fd57 100644 --- a/arch/riscv/include/asm/cacheflush.h +++ b/arch/riscv/include/asm/cacheflush.h @@ -85,7 +85,7 @@ static inline void flush_dcache_page(struct page *page) * so instead we just flush the whole thing. */ #define flush_icache_range(start, end) flush_icache_all() -#define flush_icache_user_range(vma, pg, addr, len) flush_icache_all() +#define flush_icache_user_range(vma, pg, addr, len) flush_icache_mm(vma->vm_mm, 0) void dma_wbinv_range(unsigned long start, unsigned long end); void dma_wb_range(unsigned long start, unsigned long end);
The only call path is: __access_remote_vm -> copy_to_user_page -> flush_icache_user_range Seems it's ok to use flush_icache_mm instead of flush_icache_all and it could reduce flush_icache_all called on other harts. I think the patch is the fixup for the commit 08f051eda33b. Signed-off-by: Guo Ren <guoren@linux.alibaba.com> Cc: Andrew Waterman <andrew@sifive.com> Cc: Palmer Dabbelt <palmer@sifive.com> --- arch/riscv/include/asm/cacheflush.h | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)