[Qemu-devel] [PATCH 4/6] sheepdog: don't leak socket file descriptor upon connection failure
From: Jim Meyering Signed-off-by: Jim Meyering --- block/sheepdog.c | 1 + 1 file changed, 1 insertion(+) diff --git a/block/sheepdog.c b/block/sheepdog.c index e01d371..a5c834f 100644 --- a/block/sheepdog.c +++ b/block/sheepdog.c @@ -489,6 +489,7 @@ static int connect_to_sdog(const char *addr, const char *port) if (errno == EINTR) { goto reconnect; } +close(fd); break; } -- 1.7.10.2.520.g6a4a482
Re: [Qemu-devel] [PATCH 4/6] sheepdog: don't leak socket file descriptor upon connection failure
Kevin Wolf wrote: > Am 16.05.2012 15:07, schrieb Jim Meyering: >> From: Jim Meyering >> >> Signed-off-by: Jim Meyering > > Acked-by: Kevin Wolf Hi Kevin, AFAICS, only one of these 6 patches has been applied. >From what I recall (it's been nearly 3mo), there was good feedback and I posted at least one V2 patch. For reference, here's the start of the series: http://marc.info/?l=qemu-devel&m=133717388221635&w=2 Let me know if there's anything I can do to help. Jim Meyering (5): qemu-ga: don't leak a file descriptor upon failed lockf linux-user: do_msgrcv: don't leak host_mb upon TARGET_EFAULT failure sheepdog: don't leak socket file descriptor upon connection failure arm-semi: don't leak 1KB user string lock buffer upon TARGET_SYS_OPEN softmmu-semi: fix lock_user* functions not to deref NULL upon OOM block/sheepdog.c |1 + linux-user/syscall.c |4 ++-- qemu-ga.c |3 +++ softmmu-semi.h|5 - target-arm/arm-semi.c | 13 +++-- 5 files changed, 17 insertions(+), 9 deletions(-)
Re: [Qemu-devel] [PATCH 4/6] sheepdog: don't leak socket file descriptor upon connection failure
Am 17.08.2012 15:30, schrieb Jim Meyering: > Kevin Wolf wrote: >> Am 16.05.2012 15:07, schrieb Jim Meyering: >>> From: Jim Meyering >>> >>> Signed-off-by: Jim Meyering >> >> Acked-by: Kevin Wolf > > Hi Kevin, > > AFAICS, only one of these 6 patches has been applied. > From what I recall (it's been nearly 3mo), there was good > feedback and I posted at least one V2 patch. > For reference, here's the start of the series: > > http://marc.info/?l=qemu-devel&m=133717388221635&w=2 > > Let me know if there's anything I can do to help. Oh, that's bad. This series is spreads across several subsystems, so by acking the sheepdog patch (the only block layer one) I was intending to signal that I'm okay with merging it, but that I expect a "global maintainer" to actually commit it. Did all your other series get merged? There were a lot more patches with small fixes and I can't see them in git master at all. I seem to remember that they got delayed because you posted them late during the last freeze, but obviously they should have been long committed now. Anthony, what happened with these series? I think it makes sense to pull them into -rc1 because all of them were bug fixes, even though mostly minor ones. Kevin > Jim Meyering (5): > qemu-ga: don't leak a file descriptor upon failed lockf > linux-user: do_msgrcv: don't leak host_mb upon TARGET_EFAULT failure > sheepdog: don't leak socket file descriptor upon connection failure > arm-semi: don't leak 1KB user string lock buffer upon TARGET_SYS_OPEN > softmmu-semi: fix lock_user* functions not to deref NULL upon OOM > > block/sheepdog.c |1 + > linux-user/syscall.c |4 ++-- > qemu-ga.c |3 +++ > softmmu-semi.h|5 - > target-arm/arm-semi.c | 13 +++-- > 5 files changed, 17 insertions(+), 9 deletions(-) >
Re: [Qemu-devel] [PATCH 4/6] sheepdog: don't leak socket file descriptor upon connection failure
Kevin Wolf wrote: > Am 17.08.2012 15:30, schrieb Jim Meyering: >> Kevin Wolf wrote: >>> Am 16.05.2012 15:07, schrieb Jim Meyering: From: Jim Meyering Signed-off-by: Jim Meyering >>> >>> Acked-by: Kevin Wolf >> >> Hi Kevin, >> >> AFAICS, only one of these 6 patches has been applied. >> From what I recall (it's been nearly 3mo), there was good >> feedback and I posted at least one V2 patch. >> For reference, here's the start of the series: >> >> http://marc.info/?l=qemu-devel&m=133717388221635&w=2 >> >> Let me know if there's anything I can do to help. > > Oh, that's bad. This series is spreads across several subsystems, so by > acking the sheepdog patch (the only block layer one) I was intending to > signal that I'm okay with merging it, but that I expect a "global > maintainer" to actually commit it. > > Did all your other series get merged? There were a lot more patches with > small fixes and I can't see them in git master at all. I seem to > remember that they got delayed because you posted them late during the > last freeze, but obviously they should have been long committed now. I'm going through them now. So far, it looks like most have been deferred.
Re: [Qemu-devel] [PATCH 4/6] sheepdog: don't leak socket file descriptor upon connection failure
Am 16.05.2012 15:07, schrieb Jim Meyering: > From: Jim Meyering > > > Signed-off-by: Jim Meyering Acked-by: Kevin Wolf