| Message ID | 20210913111928.98-1-xieyongji@bytedance.com (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
| Series | aio: Fix incorrect usage of eventfd_signal_allowed() | expand |
On Mon, Sep 13 2021 at 19:19, Xie Yongji wrote: > We should defer eventfd_signal() to the workqueue when > eventfd_signal_allowed() return false rather than return > true. > > Fixes: b542e383d8c0 ("eventfd: Make signal recursion protection a task bit") > Signed-off-by: Xie Yongji <xieyongji@bytedance.com> > --- > fs/aio.c | 2 +- > 1 file changed, 1 insertion(+), 1 deletion(-) > > diff --git a/fs/aio.c b/fs/aio.c > index 51b08ab01dff..8822e3ed4566 100644 > --- a/fs/aio.c > +++ b/fs/aio.c > @@ -1695,7 +1695,7 @@ static int aio_poll_wake(struct wait_queue_entry *wait, unsigned mode, int sync, > list_del(&iocb->ki_list); > iocb->ki_res.res = mangle_poll(mask); > req->done = true; > - if (iocb->ki_eventfd && eventfd_signal_allowed()) { > + if (iocb->ki_eventfd && !eventfd_signal_allowed()) { Thanks for catching that! > iocb = NULL; > INIT_WORK(&req->work, aio_poll_put_work); > schedule_work(&req->work);
Ping On Mon, Sep 13, 2021 at 7:20 PM Xie Yongji <xieyongji@bytedance.com> wrote: > > We should defer eventfd_signal() to the workqueue when > eventfd_signal_allowed() return false rather than return > true. > > Fixes: b542e383d8c0 ("eventfd: Make signal recursion protection a task bit") > Signed-off-by: Xie Yongji <xieyongji@bytedance.com> > --- > fs/aio.c | 2 +- > 1 file changed, 1 insertion(+), 1 deletion(-) > > diff --git a/fs/aio.c b/fs/aio.c > index 51b08ab01dff..8822e3ed4566 100644 > --- a/fs/aio.c > +++ b/fs/aio.c > @@ -1695,7 +1695,7 @@ static int aio_poll_wake(struct wait_queue_entry *wait, unsigned mode, int sync, > list_del(&iocb->ki_list); > iocb->ki_res.res = mangle_poll(mask); > req->done = true; > - if (iocb->ki_eventfd && eventfd_signal_allowed()) { > + if (iocb->ki_eventfd && !eventfd_signal_allowed()) { > iocb = NULL; > INIT_WORK(&req->work, aio_poll_put_work); > schedule_work(&req->work); > -- > 2.11.0 >
Ping again. On Sun, Nov 14, 2021 at 3:15 PM Yongji Xie <xieyongji@bytedance.com> wrote: > > Ping > > On Mon, Sep 13, 2021 at 7:20 PM Xie Yongji <xieyongji@bytedance.com> wrote: > > > > We should defer eventfd_signal() to the workqueue when > > eventfd_signal_allowed() return false rather than return > > true. > > > > Fixes: b542e383d8c0 ("eventfd: Make signal recursion protection a task bit") > > Signed-off-by: Xie Yongji <xieyongji@bytedance.com> > > --- > > fs/aio.c | 2 +- > > 1 file changed, 1 insertion(+), 1 deletion(-) > > > > diff --git a/fs/aio.c b/fs/aio.c > > index 51b08ab01dff..8822e3ed4566 100644 > > --- a/fs/aio.c > > +++ b/fs/aio.c > > @@ -1695,7 +1695,7 @@ static int aio_poll_wake(struct wait_queue_entry *wait, unsigned mode, int sync, > > list_del(&iocb->ki_list); > > iocb->ki_res.res = mangle_poll(mask); > > req->done = true; > > - if (iocb->ki_eventfd && eventfd_signal_allowed()) { > > + if (iocb->ki_eventfd && !eventfd_signal_allowed()) { > > iocb = NULL; > > INIT_WORK(&req->work, aio_poll_put_work); > > schedule_work(&req->work); > > -- > > 2.11.0 > >
On Mon, Sep 13, 2021 at 07:19:28PM +0800, Xie Yongji wrote: > We should defer eventfd_signal() to the workqueue when > eventfd_signal_allowed() return false rather than return > true. > > Fixes: b542e383d8c0 ("eventfd: Make signal recursion protection a task bit") > Signed-off-by: Xie Yongji <xieyongji@bytedance.com> > --- > fs/aio.c | 2 +- > 1 file changed, 1 insertion(+), 1 deletion(-) > > diff --git a/fs/aio.c b/fs/aio.c > index 51b08ab01dff..8822e3ed4566 100644 > --- a/fs/aio.c > +++ b/fs/aio.c > @@ -1695,7 +1695,7 @@ static int aio_poll_wake(struct wait_queue_entry *wait, unsigned mode, int sync, > list_del(&iocb->ki_list); > iocb->ki_res.res = mangle_poll(mask); > req->done = true; > - if (iocb->ki_eventfd && eventfd_signal_allowed()) { > + if (iocb->ki_eventfd && !eventfd_signal_allowed()) { > iocb = NULL; > INIT_WORK(&req->work, aio_poll_put_work); > schedule_work(&req->work); > -- > 2.11.0 > Since I was just working with this file...: Reviewed-by: Eric Biggers <ebiggers@google.com> I don't know who is taking aio fixes these days, but whoever does so probably should take this one at the same time as mine (https://lore.kernel.org/linux-fsdevel/20211207095726.169766-1-ebiggers@kernel.org). - Eric
On Tue, Dec 07, 2021 at 03:33:47PM -0800, Eric Biggers wrote: > On Mon, Sep 13, 2021 at 07:19:28PM +0800, Xie Yongji wrote: > > We should defer eventfd_signal() to the workqueue when > > eventfd_signal_allowed() return false rather than return > > true. > > > > Fixes: b542e383d8c0 ("eventfd: Make signal recursion protection a task bit") > > Signed-off-by: Xie Yongji <xieyongji@bytedance.com> > > --- > > fs/aio.c | 2 +- > > 1 file changed, 1 insertion(+), 1 deletion(-) > > > > diff --git a/fs/aio.c b/fs/aio.c > > index 51b08ab01dff..8822e3ed4566 100644 > > --- a/fs/aio.c > > +++ b/fs/aio.c > > @@ -1695,7 +1695,7 @@ static int aio_poll_wake(struct wait_queue_entry *wait, unsigned mode, int sync, > > list_del(&iocb->ki_list); > > iocb->ki_res.res = mangle_poll(mask); > > req->done = true; > > - if (iocb->ki_eventfd && eventfd_signal_allowed()) { > > + if (iocb->ki_eventfd && !eventfd_signal_allowed()) { > > iocb = NULL; > > INIT_WORK(&req->work, aio_poll_put_work); > > schedule_work(&req->work); > > -- > > 2.11.0 > > > > Since I was just working with this file...: > > Reviewed-by: Eric Biggers <ebiggers@google.com> > > I don't know who is taking aio fixes these days, but whoever does so probably > should take this one at the same time as mine > (https://lore.kernel.org/linux-fsdevel/20211207095726.169766-1-ebiggers@kernel.org). Apparently no one is, so I've included this patch in the pull request I've sent (https://lore.kernel.org/r/YbOdV8CPbyPAF234@sol.localdomain). - Eric
On Sat, Dec 11, 2021 at 2:53 AM Eric Biggers <ebiggers@kernel.org> wrote: > > On Tue, Dec 07, 2021 at 03:33:47PM -0800, Eric Biggers wrote: > > On Mon, Sep 13, 2021 at 07:19:28PM +0800, Xie Yongji wrote: > > > We should defer eventfd_signal() to the workqueue when > > > eventfd_signal_allowed() return false rather than return > > > true. > > > > > > Fixes: b542e383d8c0 ("eventfd: Make signal recursion protection a task bit") > > > Signed-off-by: Xie Yongji <xieyongji@bytedance.com> > > > --- > > > fs/aio.c | 2 +- > > > 1 file changed, 1 insertion(+), 1 deletion(-) > > > > > > diff --git a/fs/aio.c b/fs/aio.c > > > index 51b08ab01dff..8822e3ed4566 100644 > > > --- a/fs/aio.c > > > +++ b/fs/aio.c > > > @@ -1695,7 +1695,7 @@ static int aio_poll_wake(struct wait_queue_entry *wait, unsigned mode, int sync, > > > list_del(&iocb->ki_list); > > > iocb->ki_res.res = mangle_poll(mask); > > > req->done = true; > > > - if (iocb->ki_eventfd && eventfd_signal_allowed()) { > > > + if (iocb->ki_eventfd && !eventfd_signal_allowed()) { > > > iocb = NULL; > > > INIT_WORK(&req->work, aio_poll_put_work); > > > schedule_work(&req->work); > > > -- > > > 2.11.0 > > > > > > > Since I was just working with this file...: > > > > Reviewed-by: Eric Biggers <ebiggers@google.com> > > > > I don't know who is taking aio fixes these days, but whoever does so probably > > should take this one at the same time as mine > > (https://lore.kernel.org/linux-fsdevel/20211207095726.169766-1-ebiggers@kernel.org). > > Apparently no one is, so I've included this patch in the pull request I've sent > (https://lore.kernel.org/r/YbOdV8CPbyPAF234@sol.localdomain). > Thank you! Thanks， Yongji
diff --git a/fs/aio.c b/fs/aio.c index 51b08ab01dff..8822e3ed4566 100644 --- a/fs/aio.c +++ b/fs/aio.c @@ -1695,7 +1695,7 @@ static int aio_poll_wake(struct wait_queue_entry *wait, unsigned mode, int sync, list_del(&iocb->ki_list); iocb->ki_res.res = mangle_poll(mask); req->done = true; - if (iocb->ki_eventfd && eventfd_signal_allowed()) { + if (iocb->ki_eventfd && !eventfd_signal_allowed()) { iocb = NULL; INIT_WORK(&req->work, aio_poll_put_work); schedule_work(&req->work);
We should defer eventfd_signal() to the workqueue when eventfd_signal_allowed() return false rather than return true. Fixes: b542e383d8c0 ("eventfd: Make signal recursion protection a task bit") Signed-off-by: Xie Yongji <xieyongji@bytedance.com> --- fs/aio.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)