Date: Wed, 12 Aug 2020 17:35:17 +0300 From: Konstantin Belousov <kostikbel@gmail.com> To: =?utf-8?Q?T=C4=B3l?= Coosemans <tijl@freebsd.org> Cc: Gleb Popov <arrowd@freebsd.org>, toolchain@freebsd.org Subject: Re: Undefined reference to __atomic_store_8 Message-ID: <20200812143517.GZ2551@kib.kiev.ua> In-Reply-To: <20200812134158.249dc5ca@FreeBSD.org> References: <CALH631=FxoEdp_akV-szxTuLVrZ4oFgHWKcWJNOGEuLwqJ3Hyw@mail.gmail.com> <20200808133000.GC2551@kib.kiev.ua> <CALH631nqdHq02xYM4cnOaPhmB5L3O5TJyZiKVNrUJu7kHg0uFQ@mail.gmail.com> <20200809143742.430764e7@FreeBSD.org> <20200809154312.GH2551@kib.kiev.ua> <CALH631=7M6J3L7AzgHqEwtyBo6f-9SPM4w3PDAULUACVwJMVyg@mail.gmail.com> <20200811235210.41049ad1@FreeBSD.org> <CALH631kJhO3LTYqHsd%2Bu0c_KNegQVaphbyhrbtTxGLkLHQ9s3w@mail.gmail.com> <CALH631mVop8dKbe775VdyjSeLA1sKEO-yDdufWa%2BiT2Vi8TEFg@mail.gmail.com> <20200812134158.249dc5ca@FreeBSD.org>
next in thread | previous in thread | raw e-mail | index | archive | help
On Wed, Aug 12, 2020 at 01:41:58PM +0200, Tijl Coosemans wrote:
> On Wed, 12 Aug 2020 09:44:25 +0400 Gleb Popov <arrowd@freebsd.org> wrote:
> > On Wed, Aug 12, 2020 at 9:21 AM Gleb Popov <arrowd@freebsd.org> wrote:
> >> Indeed, this looks like a culprit! When compiling using first command line
> >> (the long one) I get following warnings:
> >>
> >> /wrkdirs/usr/ports/lang/ghc/work/ghc-8.10.1/libraries/ghc-prim/cbits/atomic.c:369:10:
> >> warning: misaligned atomic operation may incur significant performance
> >> penalty [-Watomic-alignment]
> >> return __atomic_load_n((StgWord64 *) x, __ATOMIC_SEQ_CST);
> >> ^
> >> /wrkdirs/usr/ports/lang/ghc/work/ghc-8.10.1/libraries/ghc-prim/cbits/atomic.c:417:3:
> >> warning: misaligned atomic operation may incur significant performance
> >> penalty [-Watomic-alignment]
> >> __atomic_store_n((StgWord64 *) x, (StgWord64) val, __ATOMIC_SEQ_CST);
> >> ^
> >> 2 warnings generated.
> >>
> >> I guess this basically means "I'm emitting a call there". So, what's the
> >> correct fix in this case?
> >
> > I just noticed that Clang emits these warnings (and the call instruction)
> > only for functions handling StgWord64 type. For the same code with
> > StgWord32, like
> >
> > StgWord
> > hs_atomicread32(StgWord x)
> > {
> > #if HAVE_C11_ATOMICS
> > return __atomic_load_n((StgWord32 *) x, __ATOMIC_SEQ_CST);
> > #else
> > return __sync_add_and_fetch((StgWord32 *) x, 0);
> > #endif
> > }
> >
> > no warning is emitted as well as no call.
> >
> > How does clang infer alignment in these cases? What's so special about
> > StgWord64?
>
> StgWord64 is uint64_t which is unsigned long long which is 4 byte
> aligned on i386. Clang wants 8 byte alignment to use the fildll
> instruction.
This all is very strange.
How could code use fildll to load 8 bytes as bit-value ? FILDLL converts
single and double precision fp into long-double fp, so it would change
the bit-value.
Also, both ISA and x86 psABI only require 4-byte alignment for the
double precision fp variables.
If the variable memory spans over two cache lines, then SDM states that
the access can be not atomic, but I believe it cannot happen for any
existing CPU. It might be slow. For some future CPUs, Intel provides
control which would cause such accesses to trap.
>
> You could change the definition of the StgWord64 type to look like:
>
> typedef uint64_t StgWord64 __attribute__((aligned(8)));
>
> But this only works if all calls to hs_atomicread64 pass a StgWord64
> as argument and not some other 64 bit value.
>
>
> Another solution I already mentioned in a previous message: replace
> HAVE_C11_ATOMICS with 0 in hs_atomicread64 so it uses
> __sync_add_and_fetch instead of __atomic_load_n. That uses the
> cmpxchg8b instruction which doesn't care about alignment. It's much
> slower but I guess 64 bit atomic loads are rare enough that this
> doesn't matter much.
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20200812143517.GZ2551>