diff options
author | Olof Johansson <olof@lixom.net> | 2014-09-24 11:20:48 -0700 |
---|---|---|
committer | Olof Johansson <olof@lixom.net> | 2014-09-24 11:21:01 -0700 |
commit | 28fd837204236cf5b5533525e5b53c5176fa97a3 (patch) | |
tree | ac7dcad925950bc0bb53c5a55cb199e31e7aaee0 /kernel/trace/ring_buffer.c | |
parent | c82eb464879dd0ecbe0c4cb1b80ac4e82b634872 (diff) | |
parent | 64d14a31d5410ea34641c41795e0ba222bda740c (diff) |
Merge tag 'imx-cleanup-3.18' of git://git.kernel.org/pub/scm/linux/kernel/git/shawnguo/linux into next/cleanup
Merge "ARM: imx: cleanup for 3.18" from Shawn Guo:
The i.MX cleanup for 3.18:
- Reomve a few i.MX27 and i.MX1 board files
- Remove imx_scu_standby_enable() since core code handles scu
standby now
- Remove unnecessary iomux declaration
- Remove useless sound card property from vf610-twr dts
* tag 'imx-cleanup-3.18' of git://git.kernel.org/pub/scm/linux/kernel/git/shawnguo/linux:
ARM: imx: Remove mach-mxt_td60 board file
ARM: i.MX: Remove i.MX1 ADS board support
ARM: dts: vf610-twr: remove useless property for sound card.
ARM: imx: remove imx_scu_standby_enable()
ARM: i.MX: Remove Phytec i.MX27 PCM038/PCM970 board files
ARM: i.MX: Remove mach-cpuimx27sd board file
ARM: imx: iomux: Do not export symbol without public declaration
Signed-off-by: Olof Johansson <olof@lixom.net>
Diffstat (limited to 'kernel/trace/ring_buffer.c')
-rw-r--r-- | kernel/trace/ring_buffer.c | 16 |
1 files changed, 15 insertions, 1 deletions
diff --git a/kernel/trace/ring_buffer.c b/kernel/trace/ring_buffer.c index afb04b9b818..b38fb2b9e23 100644 --- a/kernel/trace/ring_buffer.c +++ b/kernel/trace/ring_buffer.c @@ -626,8 +626,22 @@ int ring_buffer_poll_wait(struct ring_buffer *buffer, int cpu, work = &cpu_buffer->irq_work; } - work->waiters_pending = true; poll_wait(filp, &work->waiters, poll_table); + work->waiters_pending = true; + /* + * There's a tight race between setting the waiters_pending and + * checking if the ring buffer is empty. Once the waiters_pending bit + * is set, the next event will wake the task up, but we can get stuck + * if there's only a single event in. + * + * FIXME: Ideally, we need a memory barrier on the writer side as well, + * but adding a memory barrier to all events will cause too much of a + * performance hit in the fast path. We only need a memory barrier when + * the buffer goes from empty to having content. But as this race is + * extremely small, and it's not a problem if another event comes in, we + * will fix it later. + */ + smp_mb(); if ((cpu == RING_BUFFER_ALL_CPUS && !ring_buffer_empty(buffer)) || (cpu != RING_BUFFER_ALL_CPUS && !ring_buffer_empty_cpu(buffer, cpu))) |