On Wed, 2007-08-08 at 16:32 +0100, Alan Cox wrote:
> Ok try this for size folks. Use tty->flags bits for the flush status.
> Wait for the flag to clear again before returning
> Fix the doc error noted
> Fix flush of empty queue leaving stale flushpending
>
>
> Signed-off-by: Alan Cox <[EMAIL
On Wed, 8 Aug 2007 16:16:06 +0100
Alan Cox <[EMAIL PROTECTED]> wrote:
> > tty_buffer_flush() needs to wait for buf.flushpending to clear.
> Should probably make the flushpending an atomic bit op to avoid taking
> and retaking the lock.
Ok try this for size folks. Use tty->flags bits for the
> tty_buffer_flush() needs to wait for buf.flushpending to clear.
Not 100% sure it does as its no different to events really occuring with
a possible timing. Trivial to wait_event on the read queue for this
however and definitely worth doing to be sure.
Should probably make the flushpending an
On Wed, 2007-08-08 at 16:28 +0200, Laurent Pinchart wrote:
> The patch fixes the problem (at least under the test conditions that lead me
> to discover it in the first place). Thanks Alan.
It works here also, but I see a problem.
By deferring the flush, ioctl(TCFLSH) returns immediately
even
On Wed, Aug 08, 2007 at 02:45:32PM +0100, Alan Cox wrote:
> > I'm not familiar enough with the tty code to decide what the proper fix
> > should
> > be. I'll try to write a patch if someone could point me in the right
> > direction.
>
> Something like this perhaps ?
>
[]
> - * flush all
On Wednesday 08 August 2007 16:11, Paul Fulghum wrote:
> On Wed, 2007-08-08 at 14:45 +0100, Alan Cox wrote:
> > > I'm not familiar enough with the tty code to decide what the proper fix
> > > should be. I'll try to write a patch if someone could point me in the
> > > right direction.
> >
> >
On Wed, 2007-08-08 at 14:45 +0100, Alan Cox wrote:
> > I'm not familiar enough with the tty code to decide what the proper fix
> > should
> > be. I'll try to write a patch if someone could point me in the right
> > direction.
>
> Something like this perhaps ?
That looks good, a little better
> I'm not familiar enough with the tty code to decide what the proper fix
> should
> be. I'll try to write a patch if someone could point me in the right
> direction.
Something like this perhaps ?
diff -u --new-file --recursive --exclude-from /usr/src/exclude
Laurent Pinchart wrote:
Patch c5c34d4862e18ef07c1276d233507f540fb5a532 (tty: flush flip buffer on
ldisc input queue flush) introduces a race condition which can lead to memory
leaks.
The problem can be triggered when tcflush() is called when data are being
pushed to the line discipline
Hi everybody.
Patch c5c34d4862e18ef07c1276d233507f540fb5a532 (tty: flush flip buffer on
ldisc input queue flush) introduces a race condition which can lead to memory
leaks.
The problem can be triggered when tcflush() is called when data are being
pushed to the line discipline driver by
Hi everybody.
Patch c5c34d4862e18ef07c1276d233507f540fb5a532 (tty: flush flip buffer on
ldisc input queue flush) introduces a race condition which can lead to memory
leaks.
The problem can be triggered when tcflush() is called when data are being
pushed to the line discipline driver by
Laurent Pinchart wrote:
Patch c5c34d4862e18ef07c1276d233507f540fb5a532 (tty: flush flip buffer on
ldisc input queue flush) introduces a race condition which can lead to memory
leaks.
The problem can be triggered when tcflush() is called when data are being
pushed to the line discipline
I'm not familiar enough with the tty code to decide what the proper fix
should
be. I'll try to write a patch if someone could point me in the right
direction.
Something like this perhaps ?
diff -u --new-file --recursive --exclude-from /usr/src/exclude
On Wed, 2007-08-08 at 14:45 +0100, Alan Cox wrote:
I'm not familiar enough with the tty code to decide what the proper fix
should
be. I'll try to write a patch if someone could point me in the right
direction.
Something like this perhaps ?
That looks good, a little better than the
On Wednesday 08 August 2007 16:11, Paul Fulghum wrote:
On Wed, 2007-08-08 at 14:45 +0100, Alan Cox wrote:
I'm not familiar enough with the tty code to decide what the proper fix
should be. I'll try to write a patch if someone could point me in the
right direction.
Something like this
On Wed, Aug 08, 2007 at 02:45:32PM +0100, Alan Cox wrote:
I'm not familiar enough with the tty code to decide what the proper fix
should
be. I'll try to write a patch if someone could point me in the right
direction.
Something like this perhaps ?
[]
- * flush all the buffers
On Wed, 2007-08-08 at 16:28 +0200, Laurent Pinchart wrote:
The patch fixes the problem (at least under the test conditions that lead me
to discover it in the first place). Thanks Alan.
It works here also, but I see a problem.
By deferring the flush, ioctl(TCFLSH) returns immediately
even
tty_buffer_flush() needs to wait for buf.flushpending to clear.
Not 100% sure it does as its no different to events really occuring with
a possible timing. Trivial to wait_event on the read queue for this
however and definitely worth doing to be sure.
Should probably make the flushpending an
On Wed, 8 Aug 2007 16:16:06 +0100
Alan Cox [EMAIL PROTECTED] wrote:
tty_buffer_flush() needs to wait for buf.flushpending to clear.
Should probably make the flushpending an atomic bit op to avoid taking
and retaking the lock.
Ok try this for size folks. Use tty-flags bits for the flush
On Wed, 2007-08-08 at 16:32 +0100, Alan Cox wrote:
Ok try this for size folks. Use tty-flags bits for the flush status.
Wait for the flag to clear again before returning
Fix the doc error noted
Fix flush of empty queue leaving stale flushpending
Signed-off-by: Alan Cox [EMAIL PROTECTED]
20 matches
Mail list logo