linux-kernel - Re: [PATCH V2] riscv: Use flush_icache_mm for flush_icache_user

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [day] [month] [year] [list]

Message-ID: <20200127152557.GA8980@parallels-Parallels-Virtual-Platform>
Date:   Mon, 27 Jan 2020 23:25:57 +0800
From:   Guo Ren <guoren@...ux.alibaba.com>
To:     paul.walmsley@...ive.com, andrew@...ive.com, palmer@...belt.com
Cc:     linux-arch@...r.kernel.org, linux-kernel@...r.kernel.org,
        linux-riscv@...ts.infradead.org, linux-csky@...r.kernel.org,
        Albert Ou <aou@...s.berkeley.edu>
Subject: Re: [PATCH V2] riscv: Use flush_icache_mm for flush_icache_user_range

Hi all,

No ARCH_IMPLEMENTS_FLUSH_DCACHE_PAGE 1 in this patch, I'll remove it
in V3.

The update_mmu_cache() is wrong with sfence.vma and I'll give another
patch to fixup it with ARCH_IMPLEMENTS_FLUSH_DCACHE_PAGE 1.

Best Regards
 Guo Ren

On Mon, Jan 27, 2020 at 10:50:08PM +0800, Guo Ren wrote:
> The patch is the fixup for the commit 08f051eda33b by Andrew.
> 
> For copy_to_user_page, 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.
> 
> Add (vma->vm_flags & VM_EXEC) condition to flush icache only for
> executable vma area.
> 
> ARCH_IMPLEMENTS_FLUSH_DCACHE_PAGE is used in a lot of fs/block codes.
> We need it to make their pages dirty and defer sync i/dcache in
> update_mmu_cache().
> 
> Signed-off-by: Guo Ren <guoren@...ux.alibaba.com>
> Cc: Paul Walmsley <paul.walmsley@...ive.com>
> Cc: Andrew Waterman <andrew@...ive.com>
> Cc: Palmer Dabbelt <palmer@...belt.com>
> Cc: Albert Ou <aou@...s.berkeley.edu>
> 
> ---
> Changelog V2:
>  - Add VM_EXEC condition.
>  - Remove flush_icache_user_range definition.
>  - define ARCH_IMPLEMENTS_FLUSH_DCACHE_PAGE 1
> ---
>  arch/riscv/include/asm/cacheflush.h | 6 +++---
>  1 file changed, 3 insertions(+), 3 deletions(-)
> 
> diff --git a/arch/riscv/include/asm/cacheflush.h b/arch/riscv/include/asm/cacheflush.h
> index b69aecbb36d3..ae57d6ce63a9 100644
> --- a/arch/riscv/include/asm/cacheflush.h
> +++ b/arch/riscv/include/asm/cacheflush.h
> @@ -8,7 +8,7 @@
>  
>  #include <linux/mm.h>
>  
> -#define ARCH_IMPLEMENTS_FLUSH_DCACHE_PAGE 0
> +#define ARCH_IMPLEMENTS_FLUSH_DCACHE_PAGE 1
>  
>  /*
>   * The cache doesn't need to be flushed when TLB entries change when
> @@ -62,7 +62,8 @@ static inline void flush_cache_vunmap(unsigned long start, unsigned long end)
>  #define copy_to_user_page(vma, page, vaddr, dst, src, len) \
>  	do { \
>  		memcpy(dst, src, len); \
> -		flush_icache_user_range(vma, page, vaddr, len); \
> +		if (vma->vm_flags & VM_EXEC) \
> +			flush_icache_mm(vma->vm_mm, 0); \
>  	} while (0)
>  #define copy_from_user_page(vma, page, vaddr, dst, src, len) \
>  	memcpy(dst, src, len)
> @@ -85,7 +86,6 @@ 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()
>  
>  void dma_wbinv_range(unsigned long start, unsigned long end);
>  void dma_wb_range(unsigned long start, unsigned long end);
> -- 
> 2.17.0