From: Boqun Feng <boqun.feng@gmail.com>
To: "Paul E. McKenney" <paulmck@kernel.org>
Cc: linux-arch@vger.kernel.org, linux-kernel@vger.kernel.org,
elver@google.com, akpm@linux-foundation.org, tglx@linutronix.de,
peterz@infradead.org, dianders@chromium.org, pmladek@suse.com,
arnd@arndb.de, torvalds@linux-foundation.org,
kernel-team@meta.com, Mark Rutland <mark.rutland@arm.com>
Subject: Re: [PATCH v2 cmpxchg 09/13] lib: Add one-byte emulation function
Date: Tue, 14 May 2024 07:53:20 -0700 [thread overview]
Message-ID: <ZkN64LAeOfHAXyUM@boqun-archlinux> (raw)
In-Reply-To: <29f1d801-9fb4-4ecb-8d5e-cecb7d7a76e1@paulmck-laptop>
On Tue, May 14, 2024 at 07:22:47AM -0700, Paul E. McKenney wrote:
> On Mon, May 13, 2024 at 02:19:37PM -0700, Boqun Feng wrote:
> > On Mon, May 13, 2024 at 08:57:16AM -0700, Boqun Feng wrote:
> > > On Mon, May 13, 2024 at 08:41:27AM -0700, Paul E. McKenney wrote:
> > > [...]
> > > > > > +#include <linux/types.h>
> > > > > > +#include <linux/export.h>
> > > > > > +#include <linux/instrumented.h>
> > > > > > +#include <linux/atomic.h>
> > > > > > +#include <linux/panic.h>
> > > > > > +#include <linux/bug.h>
> > > > > > +#include <asm-generic/rwonce.h>
> > > > > > +#include <linux/cmpxchg-emu.h>
> > > > > > +
> > > > > > +union u8_32 {
> > > > > > + u8 b[4];
> > > > > > + u32 w;
> > > > > > +};
> > > > > > +
> > > > > > +/* Emulate one-byte cmpxchg() in terms of 4-byte cmpxchg. */
> > > > > > +uintptr_t cmpxchg_emu_u8(volatile u8 *p, uintptr_t old, uintptr_t new)
> > > > > > +{
> > > > > > + u32 *p32 = (u32 *)(((uintptr_t)p) & ~0x3);
> > > > > > + int i = ((uintptr_t)p) & 0x3;
> > > > > > + union u8_32 old32;
> > > > > > + union u8_32 new32;
> > > > > > + u32 ret;
> > > > > > +
> > > > > > + ret = READ_ONCE(*p32);
> > > > > > + do {
> > > > > > + old32.w = ret;
> > > > > > + if (old32.b[i] != old)
> > > > > > + return old32.b[i];
> > > > > > + new32.w = old32.w;
> > > > > > + new32.b[i] = new;
> > > > > > + instrument_atomic_read_write(p, 1);
> > > > > > + ret = data_race(cmpxchg(p32, old32.w, new32.w)); // Overridden above.
> > > > >
> > > > > Just out of curiosity, why is this `data_race` needed? cmpxchg is atomic
> > > > > so there should be no chance for a data race?
> > > >
> > > > That is what I thought, too. ;-)
> > > >
> > > > The problem is that the cmpxchg() covers 32 bits, and so without that
> > > > data_race(), KCSAN would complain about data races with perfectly
> > > > legitimate concurrent accesses to the other three bytes.
> > > >
> > > > The instrument_atomic_read_write(p, 1) beforehand tells KCSAN to complain
> > > > about concurrent accesses, but only to that one byte.
> > > >
> > >
> > > Oh, I see. For that purpose, maybe we can just use raw_cmpxchg() here,
> > > i.e. a cmpxchg() without any instrument in it. Cc Mark in case I'm
> > > missing something.
> > >
> >
> > I just realized that the KCSAN instrumentation is already done in
> > cmpxchg() layer:
> >
> > #define cmpxchg(ptr, ...) \
> > ({ \
> > typeof(ptr) __ai_ptr = (ptr); \
> > kcsan_mb(); \
> > instrument_atomic_read_write(__ai_ptr, sizeof(*__ai_ptr)); \
> > raw_cmpxchg(__ai_ptr, __VA_ARGS__); \
> > })
> >
> > and, this function is lower in the layer, so it shouldn't have the
> > instrumentation itself. How about the following (based on today's RCU
> > dev branch)?
>
> The raw_cmpxchg() looks nicer than the added data_race()!
>
> One question below, though.
>
> Thanx, Paul
>
> > Regards,
> > Boqun
> >
> > -------------------------------------------->8
> > Subject: [PATCH] lib: cmpxchg-emu: Make cmpxchg_emu_u8() noinstr
> >
> > Currently, cmpxchg_emu_u8() is called via cmpxchg() or raw_cmpxchg()
> > which already makes the instrumentation decision:
> >
> > * cmpxchg() case:
> >
> > cmpxchg():
> > kcsan_mb();
> > instrument_atomic_read_write(...);
> > raw_cmpxchg():
> > arch_cmpxchg():
> > cmpxchg_emu_u8();
> >
> > ... should have KCSAN instrumentation.
> >
> > * raw_cmpxchg() case:
> >
> > raw_cmpxchg():
> > arch_cmpxchg():
> > cmpxchg_emu_u8();
> >
> > ... shouldn't have KCSAN instrumentation.
> >
> > Therefore it's redundant to put KCSAN instrumentation in
> > cmpxchg_emu_u8() (along with the data_race() to get away the
> > instrumentation).
> >
> > So make cmpxchg_emu_u8() a noinstr function, and remove the KCSAN
> > instrumentation inside it.
> >
> > Signed-off-by: Boqun Feng <boqun.feng@gmail.com>
> > ---
> > include/linux/cmpxchg-emu.h | 4 +++-
> > lib/cmpxchg-emu.c | 14 ++++++++++----
> > 2 files changed, 13 insertions(+), 5 deletions(-)
> >
> > diff --git a/include/linux/cmpxchg-emu.h b/include/linux/cmpxchg-emu.h
> > index 998deec67740..c4c85f41d9f4 100644
> > --- a/include/linux/cmpxchg-emu.h
> > +++ b/include/linux/cmpxchg-emu.h
> > @@ -10,6 +10,8 @@
> > #ifndef __LINUX_CMPXCHG_EMU_H
> > #define __LINUX_CMPXCHG_EMU_H
> >
> > -uintptr_t cmpxchg_emu_u8(volatile u8 *p, uintptr_t old, uintptr_t new);
> > +#include <linux/compiler.h>
> > +
> > +noinstr uintptr_t cmpxchg_emu_u8(volatile u8 *p, uintptr_t old, uintptr_t new);
> >
> > #endif /* __LINUX_CMPXCHG_EMU_H */
> > diff --git a/lib/cmpxchg-emu.c b/lib/cmpxchg-emu.c
> > index 27f6f97cb60d..788c22cd4462 100644
> > --- a/lib/cmpxchg-emu.c
> > +++ b/lib/cmpxchg-emu.c
> > @@ -21,8 +21,13 @@ union u8_32 {
> > u32 w;
> > };
> >
> > -/* Emulate one-byte cmpxchg() in terms of 4-byte cmpxchg. */
> > -uintptr_t cmpxchg_emu_u8(volatile u8 *p, uintptr_t old, uintptr_t new)
> > +/*
> > + * Emulate one-byte cmpxchg() in terms of 4-byte cmpxchg.
> > + *
> > + * This function is marked as 'noinstr' as the instrumentation should be done at
> > + * outer layer.
> > + */
> > +noinstr uintptr_t cmpxchg_emu_u8(volatile u8 *p, uintptr_t old, uintptr_t new)
> > {
> > u32 *p32 = (u32 *)(((uintptr_t)p) & ~0x3);
> > int i = ((uintptr_t)p) & 0x3;
> > @@ -37,8 +42,9 @@ uintptr_t cmpxchg_emu_u8(volatile u8 *p, uintptr_t old, uintptr_t new)
> > return old32.b[i];
> > new32.w = old32.w;
> > new32.b[i] = new;
> > - instrument_atomic_read_write(p, 1);
>
> Don't we need to keep that instrument_atomic_read_write() in order
> to allow KCSAN to detect data races with plain C-language reads?
> Or is that being handled some other way?
>
I think that's already covered by the current code, cmpxchg_emu_u8() is
called from cmpxchg() macro, which has a:
instrument_atomic_read_write(p, sizeof(*p));
in it, and in the cmpxchg((u8*), ..) case, 'sizeof(*p)' is obviously 1
;-)
Regards,
Boqun
> > - ret = data_race(cmpxchg(p32, old32.w, new32.w)); // Overridden above.
> > +
> > + // raw_cmpxchg() is used here to avoid instrumentation.
> > + ret = raw_cmpxchg(p32, old32.w, new32.w); // Overridden above.
> > } while (ret != old32.w);
> > return old;
> > }
> > --
> > 2.44.0
> >
next prev parent reply other threads:[~2024-05-14 14:53 UTC|newest]
Thread overview: 71+ messages / expand[flat|nested] mbox.gz Atom feed top
2024-04-01 21:39 [PATCH RFC cmpxchg 0/8] Provide emulation for one- and two-byte cmpxchg() Paul E. McKenney
2024-04-01 21:39 ` [PATCH RFC cmpxchg 1/8] lib: Add one-byte and two-byte cmpxchg() emulation functions Paul E. McKenney
2024-04-02 13:07 ` Marco Elver
2024-04-02 17:15 ` Paul E. McKenney
2024-04-08 17:47 ` [PATCH RFC cmpxchg 0/8] Provide emulation for one- and two-byte cmpxchg() Paul E. McKenney
2024-04-08 17:49 ` [PATCH cmpxchg 01/14] sparc32: make __cmpxchg_u32() return u32 Paul E. McKenney
2024-04-08 17:49 ` [PATCH cmpxchg 02/14] sparc32: make the first argument of __cmpxchg_u64() volatile u64 * Paul E. McKenney
2024-04-08 17:49 ` [PATCH cmpxchg 03/14] sparc32: unify __cmpxchg_u{32,64} Paul E. McKenney
2024-04-08 17:49 ` [PATCH cmpxchg 04/14] sparc32: add __cmpxchg_u{8,16}() and teach __cmpxchg() to handle those sizes Paul E. McKenney
2024-04-08 17:49 ` [PATCH cmpxchg 05/14] parisc: __cmpxchg_u32(): lift conversion into the callers Paul E. McKenney
2024-04-08 17:49 ` [PATCH cmpxchg 06/14] parisc: unify implementations of __cmpxchg_u{8,32,64} Paul E. McKenney
2024-04-08 17:49 ` [PATCH cmpxchg 07/14] parisc: add missing export of __cmpxchg_u8() Paul E. McKenney
2024-04-08 17:49 ` [PATCH cmpxchg 08/14] parisc: add u16 support to cmpxchg() Paul E. McKenney
2024-04-08 20:10 ` Linus Torvalds
2024-04-08 20:53 ` Paul E. McKenney
2024-04-08 17:49 ` [PATCH cmpxchg 09/14] lib: Add one-byte emulation function Paul E. McKenney
2024-04-08 17:49 ` [PATCH cmpxchg 10/14] ARC: Emulate one-byte cmpxchg Paul E. McKenney
2024-04-08 17:49 ` [PATCH cmpxchg 11/14] csky: " Paul E. McKenney
2024-04-08 17:49 ` [PATCH cmpxchg 12/14] sh: " Paul E. McKenney
2024-04-18 8:04 ` Geert Uytterhoeven
2024-04-08 17:49 ` [PATCH cmpxchg 13/14] xtensa: " Paul E. McKenney
2024-04-18 8:06 ` Geert Uytterhoeven
2024-04-18 23:21 ` Paul E. McKenney
2024-04-19 5:07 ` Yujie Liu
2024-04-19 8:02 ` Geert Uytterhoeven
2024-04-20 14:03 ` Paul E. McKenney
2024-04-08 17:49 ` [PATCH cmpxchg 14/14] riscv: " Paul E. McKenney
2024-04-09 17:35 ` Andrea Parri
2024-04-09 18:08 ` Paul E. McKenney
2024-05-01 22:58 ` [PATCH v2 cmpxchg 0/8] Provide emulation for one--byte cmpxchg() Paul E. McKenney
2024-05-01 23:01 ` [PATCH v2 cmpxchg 01/13] sparc32: make __cmpxchg_u32() return u32 Paul E. McKenney
2024-05-01 23:01 ` [PATCH v2 cmpxchg 02/13] sparc32: make the first argument of __cmpxchg_u64() volatile u64 * Paul E. McKenney
2024-05-01 23:01 ` [PATCH v2 cmpxchg 03/13] sparc32: unify __cmpxchg_u{32,64} Paul E. McKenney
2024-05-01 23:01 ` [PATCH v2 cmpxchg 04/13] sparc32: add __cmpxchg_u{8,16}() and teach __cmpxchg() to handle those sizes Paul E. McKenney
2024-05-01 23:01 ` [PATCH v2 cmpxchg 05/13] parisc: __cmpxchg_u32(): lift conversion into the callers Paul E. McKenney
2024-05-01 23:01 ` [PATCH v2 cmpxchg 06/13] parisc: unify implementations of __cmpxchg_u{8,32,64} Paul E. McKenney
2024-05-01 23:01 ` [PATCH v2 cmpxchg 07/13] parisc: add missing export of __cmpxchg_u8() Paul E. McKenney
2024-05-01 23:01 ` [PATCH v2 cmpxchg 08/13] parisc: add u16 support to cmpxchg() Paul E. McKenney
2024-05-01 23:01 ` [PATCH v2 cmpxchg 09/13] lib: Add one-byte emulation function Paul E. McKenney
2024-05-13 14:44 ` Boqun Feng
2024-05-13 15:41 ` Paul E. McKenney
2024-05-13 15:57 ` Boqun Feng
2024-05-13 21:19 ` Boqun Feng
2024-05-14 14:22 ` Paul E. McKenney
2024-05-14 14:53 ` Boqun Feng [this message]
2024-05-14 15:02 ` Paul E. McKenney
2024-05-01 23:01 ` [PATCH v2 cmpxchg 10/13] ARC: Emulate one-byte cmpxchg Paul E. McKenney
2024-05-01 23:01 ` [PATCH v2 cmpxchg 11/13] csky: " Paul E. McKenney
2024-05-11 6:42 ` Guo Ren
2024-05-11 14:49 ` Paul E. McKenney
2024-05-01 23:01 ` [PATCH v2 cmpxchg 12/13] sh: " Paul E. McKenney
2024-05-02 4:52 ` John Paul Adrian Glaubitz
2024-05-02 5:06 ` Paul E. McKenney
2024-05-02 5:11 ` John Paul Adrian Glaubitz
2024-05-02 13:33 ` Paul E. McKenney
2024-05-02 20:53 ` Al Viro
2024-05-02 21:01 ` alpha cmpxchg.h (was Re: [PATCH v2 cmpxchg 12/13] sh: Emulate one-byte cmpxchg) Al Viro
2024-05-02 22:16 ` Linus Torvalds
2024-05-02 21:18 ` [PATCH v2 cmpxchg 12/13] sh: Emulate one-byte cmpxchg Paul E. McKenney
2024-05-02 22:07 ` Al Viro
2024-05-02 23:12 ` Paul E. McKenney
2024-05-02 23:24 ` Al Viro
2024-05-02 23:45 ` Paul E. McKenney
2024-05-02 23:32 ` Linus Torvalds
2024-05-03 0:16 ` Paul E. McKenney
2024-05-02 21:50 ` Arnd Bergmann
2024-05-02 5:42 ` D. Jeff Dionne
2024-05-02 11:30 ` Arnd Bergmann
2024-05-01 23:01 ` [PATCH v2 cmpxchg 13/13] xtensa: " Paul E. McKenney
2024-05-02 20:01 ` [PATCH v2 cmpxchg 0/8] Provide emulation for one--byte cmpxchg() Al Viro
2024-05-02 21:20 ` Paul E. McKenney
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=ZkN64LAeOfHAXyUM@boqun-archlinux \
--to=boqun.feng@gmail.com \
--cc=akpm@linux-foundation.org \
--cc=arnd@arndb.de \
--cc=dianders@chromium.org \
--cc=elver@google.com \
--cc=kernel-team@meta.com \
--cc=linux-arch@vger.kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=mark.rutland@arm.com \
--cc=paulmck@kernel.org \
--cc=peterz@infradead.org \
--cc=pmladek@suse.com \
--cc=tglx@linutronix.de \
--cc=torvalds@linux-foundation.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).