search for: load_acquir

...mment. > - if ($line =~ /\b(mb|rmb|wmb|read_barrier_depends|smp_mb|smp_rmb|smp_wmb|smp_read_barrier_depends)\(/) { > + > + my $barriers = qr{ > + mb| > + rmb| > + wmb| > + read_barrier_depends > + }x; > + my $smp_barriers = qr{ > + store_release| > + load_acquire| > + store_mb| > + ($barriers) > + }x; If I use a variable called $smp_barriers, I'd expect it to actually be the smp_barriers, not to have to prefix it with smp_ before using it. my $smp_barriers = qr{ smp_store_release| smp_load_acquire| smp_store_mb| smp_read_b...

...mment. > - if ($line =~ /\b(mb|rmb|wmb|read_barrier_depends|smp_mb|smp_rmb|smp_wmb|smp_read_barrier_depends)\(/) { > + > + my $barriers = qr{ > + mb| > + rmb| > + wmb| > + read_barrier_depends > + }x; > + my $smp_barriers = qr{ > + store_release| > + load_acquire| > + store_mb| > + ($barriers) > + }x; If I use a variable called $smp_barriers, I'd expect it to actually be the smp_barriers, not to have to prefix it with smp_ before using it. my $smp_barriers = qr{ smp_store_release| smp_load_acquire| smp_store_mb| smp_read_b...

As part of memory barrier cleanup, this patchset extends checkpatch to make it easier to stop incorrect memory barrier usage. This applies on top of my series arch: barrier cleanup + barriers for virt and will be included in the next version of the series. Changes from v2: catch optional\s* before () in barriers rewrite using qr{} instead of map Michael S. Tsirkin (3): checkpatch.pl: add

As part of memory barrier cleanup, this patchset extends checkpatch to make it easier to stop incorrect memory barrier usage. This applies on top of my series arch: barrier cleanup + barriers for virt and will be included in the next version of the series. Changes from v2: catch optional\s* before () in barriers rewrite using qr{} instead of map Michael S. Tsirkin (3): checkpatch.pl: add

virtio ring entries have exactly the acquire/release semantics: - reading used index acquires a ring entry from host - updating the available index releases it to host Thus when using weak barriers and building for SMP (as most people do), smp_load_acquire and smp_store_release will do exactly the right thing to synchronize with the host. In fact, QEMU already uses __atomic_thread_fence(__ATOMIC_ACQUIRE) and __atomic_thread_fence(__ATOMIC_RELEASE); Documentation/circular-buffers.txt suggests smp_load_acquire and smp_store_release for head and tail...

virtio ring entries have exactly the acquire/release semantics: - reading used index acquires a ring entry from host - updating the available index releases it to host Thus when using weak barriers and building for SMP (as most people do), smp_load_acquire and smp_store_release will do exactly the right thing to synchronize with the host. In fact, QEMU already uses __atomic_thread_fence(__ATOMIC_ACQUIRE) and __atomic_thread_fence(__ATOMIC_RELEASE); Documentation/circular-buffers.txt suggests smp_load_acquire and smp_store_release for head and tail...

...--------------------------------------------------------------- Yes I know that the patch by Davidlohr Bueso has a typo in the subject :( Davidlohr Bueso (1): lcoking/barriers, arch: Use smp barriers in smp_store_release() Michael S. Tsirkin (40): asm-generic: guard smp_store_release/load_acquire ia64: rename nop->iosapic_nop ia64: reuse asm-generic/barrier.h powerpc: reuse asm-generic/barrier.h s390: reuse asm-generic/barrier.h sparc: reuse asm-generic/barrier.h arm: reuse asm-generic/barrier.h arm64: reuse asm-generic/barrier.h metag: re...

...--------------------------------------------------------------- Yes I know that the patch by Davidlohr Bueso has a typo in the subject :( Davidlohr Bueso (1): lcoking/barriers, arch: Use smp barriers in smp_store_release() Michael S. Tsirkin (40): asm-generic: guard smp_store_release/load_acquire ia64: rename nop->iosapic_nop ia64: reuse asm-generic/barrier.h powerpc: reuse asm-generic/barrier.h s390: reuse asm-generic/barrier.h sparc: reuse asm-generic/barrier.h arm: reuse asm-generic/barrier.h arm64: reuse asm-generic/barrier.h metag: re...

On Thu, Dec 17, 2015 at 12:29:03PM +0200, Michael S. Tsirkin wrote: > +static inline __virtio16 virtio_load_acquire(bool weak_barriers, __virtio16 *p) > +{ > + if (!weak_barriers) { > + rmb(); > + return READ_ONCE(*p); > + } > +#ifdef CONFIG_SMP > + return smp_load_acquire(p); > +#else > + dma_rmb(); > + return READ_ONCE(*p); > +#endif > +} This too is wrong. Look for exam...

...{ } } # check for memory barriers without a comment. - if ($line =~ /\b(mb|rmb|wmb|read_barrier_depends|smp_mb|smp_rmb|smp_wmb|smp_read_barrier_depends)\(/) { + + my $barriers = qr{ + mb| + rmb| + wmb| + read_barrier_depends + }x; + my $smp_barriers = qr{ + store_release| + load_acquire| + store_mb| + ($barriers) + }x; + my $all_barriers = qr{ + $barriers| + smp_($smp_barriers) + }x; + + if ($line =~ /\b($all_barriers)\s*\(/) { if (!ctx_has_comment($first_line, $linenr)) { WARN("MEMORY_BARRIER", "memory barrier without comment\n"...

As part of memory barrier cleanup, this patchset extends checkpatch to make it easier to stop incorrect memory barrier usage. This replaces the checkpatch patches in my series arch: barrier cleanup + barriers for virt and will be included in the next version of the series. changes from v2: address comments by Joe Perches: use (?: ... ) to avoid unnecessary capture groups rename smp_barriers

As part of memory barrier cleanup, this patchset extends checkpatch to make it easier to stop incorrect memory barrier usage. This replaces the checkpatch patches in my series arch: barrier cleanup + barriers for virt and will be included in the next version of the series. changes from v2: address comments by Joe Perches: use (?: ... ) to avoid unnecessary capture groups rename smp_barriers

...----------------------- > > Yes I know that the patch by Davidlohr Bueso has a typo in the subject :( > > Davidlohr Bueso (1): > lcoking/barriers, arch: Use smp barriers in smp_store_release() > > Michael S. Tsirkin (40): > asm-generic: guard smp_store_release/load_acquire > ia64: rename nop->iosapic_nop > ia64: reuse asm-generic/barrier.h > powerpc: reuse asm-generic/barrier.h > s390: reuse asm-generic/barrier.h > sparc: reuse asm-generic/barrier.h > arm: reuse asm-generic/barrier.h > arm64: reuse as...

As part of memory barrier cleanup, this patchset extends checkpatch to make it easier to stop incorrect memory barrier usage. This replaces the checkpatch patches in my series arch: barrier cleanup + barriers for virt and will be included in the pull request including the series. changes from v3: rename smp_barrier_stems to barrier_stems as suggested by Julian Calaby. add (?: ... ) around a

As part of memory barrier cleanup, this patchset extends checkpatch to make it easier to stop incorrect memory barrier usage. This replaces the checkpatch patches in my series arch: barrier cleanup + barriers for virt and will be included in the pull request including the series. changes from v3: rename smp_barrier_stems to barrier_stems as suggested by Julian Calaby. add (?: ... ) around a

...en vq try to start or finish uses the map. This means, when it was even, we're sure there's no readers and MMU notifier is synchronized. When it was odd, it means there's a reader we need to wait it to be even again then we are synchronized. To avoid full memory barrier, store_release + load_acquire on the counter is used. Consider the read critical section is pretty small the synchronization should be done very fast. Note the patch lead about 3% PPS dropping. Reported-by: Michael S. Tsirkin <mst at redhat.com> Fixes: 7f466032dc9e ("vhost: access vq metadata through kernel virtu...

...ish > uses the map. This means, when it was even, we're sure there's no > readers and MMU notifier is synchronized. When it was odd, it means > there's a reader we need to wait it to be even again then we are > synchronized. To avoid full memory barrier, store_release + > load_acquire on the counter is used. Unfortunately this needs a lot of review and testing, so this can't make rc2, and I don't think this is the kind of patch I can merge after rc3. Subtle memory barrier tricks like this can introduce new bugs while they are fixing old ones. > > Consider t...

...ish > uses the map. This means, when it was even, we're sure there's no > readers and MMU notifier is synchronized. When it was odd, it means > there's a reader we need to wait it to be even again then we are > synchronized. To avoid full memory barrier, store_release + > load_acquire on the counter is used. Unfortunately this needs a lot of review and testing, so this can't make rc2, and I don't think this is the kind of patch I can merge after rc3. Subtle memory barrier tricks like this can introduce new bugs while they are fixing old ones. > > Consider t...

Hi all: This series try to fix several issues introduced by meta data accelreation series. Please review. Changes from V1: - Try not use RCU to syncrhonize MMU notifier with vhost worker - set dirty pages after no readers - return -EAGAIN only when we find the range is overlapped with metadata Jason Wang (9): vhost: don't set uaddr for invalid address vhost: validate MMU notifier

...hen it was even, we're sure there's no > >> readers and MMU notifier is synchronized. When it was odd, it means > >> there's a reader we need to wait it to be even again then we are > >> synchronized. To avoid full memory barrier, store_release + > >> load_acquire on the counter is used. > > > > Unfortunately this needs a lot of review and testing, so this can't make > > rc2, and I don't think this is the kind of patch I can merge after rc3. > > Subtle memory barrier tricks like this can introduce new bugs while they > &gt...