[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <CAAhV-H59EzR7gN0UiJP5x2wbs2_TZCej-NE==OZxVcP0cvt7hg@mail.gmail.com>
Date: Wed, 20 Mar 2024 21:49:28 +0800
From: Huacai Chen <chenhuacai@...nel.org>
To: Xi Ruoyao <xry111@...111.site>
Cc: Huacai Chen <chenhuacai@...ngson.cn>, Arnd Bergmann <arnd@...db.de>, loongarch@...ts.linux.dev,
linux-arch@...r.kernel.org, Xuefeng Li <lixuefeng@...ngson.cn>,
Guo Ren <guoren@...nel.org>, Xuerui Wang <kernel@...0n.name>,
Jiaxun Yang <jiaxun.yang@...goat.com>, linux-kernel@...r.kernel.org,
loongson-kernel@...ts.loongnix.cn, stable@...r.kernel.org,
Xiaotian Wu <wuxiaotian@...ngson.cn>, Miao Wang <shankerwangmiao@...il.com>,
Xing Li <lixing@...ngson.cn>, Hongchen Zhang <zhanghongchen@...ngson.cn>,
Rui Wang <wangrui@...ngson.cn>
Subject: Re: [PATCH] LoongArch: Change __my_cpu_offset definition to avoid mis-optimization
Hi, Ruoyao,
On Wed, Mar 20, 2024 at 6:27 PM Xi Ruoyao <xry111@...111.site> wrote:
>
> On Fri, 2024-03-15 at 10:45 +0800, Huacai Chen wrote:
> > From GCC commit 3f13154553f8546a ("df-scan: remove ad-hoc handling of
> > global regs in asms"), global registers will no longer be forced to add
> > to the def-use chain. Then current_thread_info(), current_stack_pointer
> > and __my_cpu_offset may be lifted out of the loop because they are no
> > longer treated as "volatile variables".
>
> Ooops... I'm wondering why this issue has not blown up our systems
> before. The referred GCC commit is far before LoongArch CPUs are taped.
>
> > This optimization is still correct for the current_thread_info() and
> > current_stack_pointer usages because they are associated to a thread.
> > However it is wrong for __my_cpu_offset because it is associated to a
> > CPU rather than a thread: if the thread migrates to a different CPU in
> > the loop, __my_cpu_offset should be changed.
> >
> > Change __my_cpu_offset definition to treat it as a "volatile variable",
> > in order to avoid such a mis-optimization.
> >
> > Cc: stable@...r.kernel.org
>
> I suppose we should add Fixes: 5b0b14e550a0 ("LoongArch: Add
> atomic/locking header") here.
You are right, but since this patch is in loongarch-next for the pull
request, I don't want to change things.
Huacai
>
> > Reported-by: Xiaotian Wu <wuxiaotian@...ngson.cn>
> > Reported-by: Miao Wang <shankerwangmiao@...il.com>
> > Signed-off-by: Xing Li <lixing@...ngson.cn>
> > Signed-off-by: Hongchen Zhang <zhanghongchen@...ngson.cn>
> > Signed-off-by: Rui Wang <wangrui@...ngson.cn>
> > Signed-off-by: Huacai Chen <chenhuacai@...ngson.cn>
> > ---
> > arch/loongarch/include/asm/percpu.h | 6 +++++-
> > 1 file changed, 5 insertions(+), 1 deletion(-)
> >
> > diff --git a/arch/loongarch/include/asm/percpu.h b/arch/loongarch/include/asm/percpu.h
> > index 9b36ac003f89..03b98491d301 100644
> > --- a/arch/loongarch/include/asm/percpu.h
> > +++ b/arch/loongarch/include/asm/percpu.h
> > @@ -29,7 +29,12 @@ static inline void set_my_cpu_offset(unsigned long off)
> > __my_cpu_offset = off;
> > csr_write64(off, PERCPU_BASE_KS);
> > }
> > -#define __my_cpu_offset __my_cpu_offset
> > +
> > +#define __my_cpu_offset \
> > +({ \
> > + __asm__ __volatile__("":"+r"(__my_cpu_offset)); \
> > + __my_cpu_offset; \
> > +})
> >
> > #define PERCPU_OP(op, asm_op, c_op) \
> > static __always_inline unsigned long __percpu_##op(void *ptr, \
>
> --
> Xi Ruoyao <xry111@...111.site>
> School of Aerospace Science and Technology, Xidian University
Powered by blists - more mailing lists