[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <CAGsJ_4y-bnuS_EkHuvm5BpmWb2BrTvYBU97ObTo3bwmHBNiqfA@mail.gmail.com>
Date: Mon, 18 Mar 2024 08:37:44 +0800
From: Barry Song <21cnbao@...il.com>
To: Guenter Roeck <linux@...ck-us.net>
Cc: akpm@...ux-foundation.org, linux-kernel@...r.kernel.org, chris@...kel.net,
jcmvbkbc@...il.com, willy@...radead.org, alexghiti@...osinc.com,
rppt@...nel.org, dennis@...nel.org, Barry Song <v-songbaohua@...o.com>,
Huacai Chen <chenhuacai@...ngson.cn>, Herbert Xu <herbert@...dor.apana.org.au>,
kernel test robot <lkp@...el.com>
Subject: Re: [PATCH] xtensa: remove redundant flush_dcache_page and
ARCH_IMPLEMENTS_FLUSH_DCACHE_PAGE macros
On Mon, Mar 18, 2024 at 8:34 AM Guenter Roeck <linux@...ck-us.net> wrote:
>
> On 3/17/24 17:27, Barry Song wrote:
> > On Mon, Mar 18, 2024 at 8:16 AM Guenter Roeck <linux@...ck-us.net> wrote:
> >>
> >> On Wed, Mar 13, 2024 at 05:50:36PM +1300, Barry Song wrote:
> >>> From: Barry Song <v-songbaohua@...o.com>
> >>>
> >>> xtensa's flush_dcache_page() can be a no-op sometimes. There is a
> >>> generic implementation for this case in include/asm-generic/
> >>> cacheflush.h.
> >>> #ifndef ARCH_IMPLEMENTS_FLUSH_DCACHE_PAGE
> >>> static inline void flush_dcache_page(struct page *page)
> >>> {
> >>> }
> >>>
> >>> #define ARCH_IMPLEMENTS_FLUSH_DCACHE_PAGE 0
> >>> #endif
> >>>
> >>> So remove the superfluous flush_dcache_page() definition, which also
> >>> helps silence potential build warnings complaining the page variable
> >>> passed to flush_dcache_page() is not used.
> >>>
> >>> In file included from crypto/scompress.c:12:
> >>> include/crypto/scatterwalk.h: In function 'scatterwalk_pagedone':
> >>> include/crypto/scatterwalk.h:76:30: warning: variable 'page' set but not used [-Wunused-but-set-variable]
> >>> 76 | struct page *page;
> >>> | ^~~~
> >>> crypto/scompress.c: In function 'scomp_acomp_comp_decomp':
> >>>>> crypto/scompress.c:174:38: warning: unused variable 'dst_page' [-Wunused-variable]
> >>> 174 | struct page *dst_page = sg_page(req->dst);
> >>> |
> >>>
> >>> The issue was originally reported on LoongArch by kernel test
> >>> robot. And Huacai fixed it on LoongArch, but I found xtensa
> >>> obviously has the same issue.
> >>>
> >>
> >> Maybe I am doing something wrong, but this patch doesn't build
> >> for me.
> >>
> >> CC arch/xtensa/kernel/asm-offsets.s
> >> In file included from ./include/linux/highmem.h:8,
> >> from ./include/linux/bvec.h:10,
> >> from ./include/linux/blk_types.h:10,
> >> from ./include/linux/writeback.h:13,
> >> from ./include/linux/memcontrol.h:23,
> >> from ./include/linux/swap.h:9,
> >> from ./include/linux/suspend.h:5,
> >> from arch/xtensa/kernel/asm-offsets.c:24:
> >> ./include/linux/cacheflush.h:9:5: error: "ARCH_IMPLEMENTS_FLUSH_DCACHE_PAGE" is not defined, evaluates to 0 [-Werror=undef]
> >> 9 | #if ARCH_IMPLEMENTS_FLUSH_DCACHE_PAGE
> >
> > is it because xtensa doesn't include this at the end of
> > arch/xtensa/include/asm/cacheflush.h
> > while other architectures do?
> >
> > #include <asm-generic/cacheflush.h>
> >
>
> Looks like it (see my other e-mail).
i will send v2 and it seems a lot of other code can also be removed in
arch/xtensa/include/asm/cacheflush.h,
for example
#define flush_cache_all() do { } while (0)
#define flush_cache_mm(mm) do { } while (0)
#define flush_cache_dup_mm(mm) do { } while (0)
#define flush_cache_vmap(start,end) do { } while (0)
#define flush_cache_vmap_early(start,end) do { } while (0)
#define flush_cache_vunmap(start,end) do { } while (0)
#define ARCH_IMPLEMENTS_FLUSH_DCACHE_PAGE 0
#define flush_dcache_page(page) do { } while (0)
#define flush_cache_page(vma, addr, pfn) do { } while (0)
#define flush_cache_range(vma, start, end) do { } while (0)
#define flush_dcache_mmap_lock(mapping) do { } while (0)
#define flush_dcache_mmap_unlock(mapping) do { } while (0)
>
> Guenter
>
Powered by blists - more mailing lists