Re: [Mesa-dev] [PATCH 0/7] Fix GLX files generated by Mesa for the server

2011-12-13 Thread Adam Jackson
On Thu, 2011-12-08 at 12:47 -0800, Ian Romanick wrote: There has been quite a bit of skew between what's in Mesa and what's needed in the xserver. This patch series cleans that up. Most of the changes are quite mundane and just make the code compile inside the xserver. However, the changes

Re: [PATCH 3/5] dix: move storing last.valuators into a helper function

2011-12-13 Thread Chase Douglas
On 12/09/2011 11:54 PM, Peter Hutterer wrote: Signed-off-by: Peter Hutterer peter.hutte...@who-t.net --- dix/getevents.c | 38 ++ 1 files changed, 26 insertions(+), 12 deletions(-) diff --git a/dix/getevents.c b/dix/getevents.c index bc532ca..3b3194d

Re: [PATCH:xf86-input-synaptics] Revert: eventcomm: replace synaptics-custom TEST_BIT with server's BitIsOn.

2011-12-13 Thread Peter Hutterer
On Mon, Dec 12, 2011 at 01:49:23AM +0100, Cyril Brulebois wrote: This commit reverts 13543b156d78bc4d01a19844a5ee8f283269621b As seen in Debian's #648488, this switch causes a regression on PowerPC, especially seen on iBook G4 with appletouch. Take a defensive stance and revert back to a

[PATCH 0/4] Minor performance tuning

2011-12-13 Thread Adam Jackson
Found these in an old work branch. The first two are only relevant if you're using XACE or dtrace, the second two apply unconditionally. The total series is good for about 13% speedup on x11perf's 10x10 PutImage and ShmPutImage tests, and should be a win across the board. Alan, I assume the

[PATCH 1/4] xace: ricer tuning for XaceHook

2011-12-13 Thread Adam Jackson
gcc doesn't want to hoist the check for XaceHooks[hook] != NULL above the varargs code for some reason, so do it ourselves. Before: 4000 trep @ 0.0010 msec (1050420.2/sec): PutImage 10x10 square 6000 trep @ 0.0005 msec (1921147.6/sec): ShmPutImage 10x10 square After: 4000 trep @

[PATCH 2/4] dix: Tune dtrace hooks around Dispatch

2011-12-13 Thread Adam Jackson
Don't call LookupMajorName if the hooks aren't active, it's quite expensive. Before: 4000 trep @ 0.0009 msec (1087458.5/sec): PutImage 10x10 square 6000 trep @ 0.0005 msec (2012238.6/sec): ShmPutImage 10x10 square After: 4000 trep @ 0.0009 msec (1109091.3/sec): PutImage 10x10

[PATCH 3/4] dix: Tune dixLookupDrawable for success

2011-12-13 Thread Adam Jackson
The vast vast vast majority of resource lookups are successful. Move some work to the error paths so we don't punish success. Before: 4000 trep @ 0.0009 msec (1109091.3/sec): PutImage 10x10 square 6000 trep @ 0.0005 msec (2072652.2/sec): ShmPutImage 10x10 square After: 4000 trep

[PATCH 4/4] miext/damage: Only wrap into the GC ops chain if there's a listener

2011-12-13 Thread Adam Jackson
Before: 4000 trep @ 0.0009 msec (1148346.9/sec): PutImage 10x10 square 6000 trep @ 0.0005 msec (2091666.1/sec): ShmPutImage 10x10 square After: 4000 trep @ 0.0008 msec (1191807.5/sec): PutImage 10x10 square 6000 trep @ 0.0005 msec (2180983.0/sec): ShmPutImage 10x10 square

Re: [PATCH 1/4] xace: ricer tuning for XaceHook

2011-12-13 Thread Alan Coopersmith
On 12/13/11 14:31, Adam Jackson wrote: gcc doesn't want to hoist the check for XaceHooks[hook] != NULL above the varargs code for some reason, so do it ourselves. Perhaps because it would have to do some gymnastics to realize that in all cases we happen to either initialize prv to NULL or a

Re: [PATCH 2/4] dix: Tune dtrace hooks around Dispatch

2011-12-13 Thread Alan Coopersmith
On 12/13/11 14:31, Adam Jackson wrote: Don't call LookupMajorName if the hooks aren't active, it's quite expensive. It was a simple table lookup when I added it, before the conversion to the registry functions in commit 996b621bec. @@ -425,9 +425,11 @@ Dispatch(void)

[PATCH inputproto multitouch] specs: drop leftover from active_touches removal

2011-12-13 Thread Peter Hutterer
Signed-off-by: Peter Hutterer peter.hutte...@who-t.net --- specs/XI2proto.txt |5 - 1 files changed, 0 insertions(+), 5 deletions(-) diff --git a/specs/XI2proto.txt b/specs/XI2proto.txt index 6082166..dcc2bc9 100644 --- a/specs/XI2proto.txt +++ b/specs/XI2proto.txt @@ -2112,11 +2112,6 @@

[PATCH inputproto multitouch] specs: clarify button state in touch events

2011-12-13 Thread Peter Hutterer
Emulated pointer events will have button 1 logically down, but touch events only represent the actual button state, irrespective of the touches. Signed-off-by: Peter Hutterer peter.hutte...@who-t.net --- specs/XI2proto.txt |9 +++-- 1 files changed, 7 insertions(+), 2 deletions(-) diff

Re: [PATCH 3/4] dix: Tune dixLookupDrawable for success

2011-12-13 Thread Alan Coopersmith
On 12/13/11 14:31, Adam Jackson wrote: The vast vast vast majority of resource lookups are successful. Move some work to the error paths so we don't punish success. Reviewed-by: Alan Coopersmith alan.coopersm...@oracle.com -- -Alan Coopersmith-alan.coopersm...@oracle.com

[PATCH] configure: split the required modules up

2011-12-13 Thread Peter Hutterer
We do the same thing for libraries and optional modules already, and it's much easier to read when one of them changes Signed-off-by: Peter Hutterer peter.hutte...@who-t.net --- configure.ac | 18 -- 1 files changed, 16 insertions(+), 2 deletions(-) diff --git a/configure.ac

Re: [PATCH] configure: split the required modules up

2011-12-13 Thread Alan Coopersmith
On 12/13/11 16:34, Peter Hutterer wrote: We do the same thing for libraries and optional modules already, and it's much easier to read when one of them changes Signed-off-by: Peter Huttererpeter.hutte...@who-t.net Reviewed-by: Alan Coopersmith alan.coopersm...@oracle.com -- -Alan

Re: [PATCH 4/4] miext/damage: Only wrap into the GC ops chain if there's a listener

2011-12-13 Thread Eric Anholt
On Tue, 13 Dec 2011 17:31:53 -0500, Adam Jackson a...@redhat.com wrote: Before: 4000 trep @ 0.0009 msec (1148346.9/sec): PutImage 10x10 square 6000 trep @ 0.0005 msec (2091666.1/sec): ShmPutImage 10x10 square After: 4000 trep @ 0.0008 msec (1191807.5/sec): PutImage 10x10

Re: [PATCH] configure: split the required modules up

2011-12-13 Thread Gaetan Nadon
On Wed, 2011-12-14 at 10:34 +1000, Peter Hutterer wrote: We do the same thing for libraries and optional modules already, and it's much easier to read when one of them changes Signed-off-by: Peter Hutterer peter.hutte...@who-t.net --- configure.ac | 18 -- 1 files

Re: [PATCH 11/31] mi: _X_UNEXPORT

2011-12-13 Thread Adam Tkac
On Wed, Dec 07, 2011 at 08:40:41AM +1000, Peter Hutterer wrote: On Fri, Dec 02, 2011 at 11:31:17AM +, Daniel Stone wrote: Remove unnecessary _X_EXPORT tags from mi headers. Signed-off-by: Daniel Stone dan...@fooishbar.org some of these are still used by input drivers, most likely in

[PATCH libX11] XQueryColors: Split a request into multiple requests if necessary

2011-12-13 Thread Kusanagi Kouichi
https://bugs.freedesktop.org/show_bug.cgi?id=9160 Signed-off-by: Kusanagi Kouichi sl...@ac.auone-net.jp --- src/QuColors.c | 32 +++- 1 files changed, 27 insertions(+), 5 deletions(-) diff --git a/src/QuColors.c b/src/QuColors.c index e51375f..237b8bf 100644 ---

Re: [PATCH 03/31] Input: _X_UNEXPORT

2011-12-13 Thread Adam Tkac
On Wed, Dec 07, 2011 at 08:45:03AM +1000, Peter Hutterer wrote: On Fri, Dec 02, 2011 at 11:31:09AM +, Daniel Stone wrote: Remove _X_EXPORT tag from needlessly exported symbols. Signed-off-by: Daniel Stone dan...@fooishbar.org PickPointer(), IsMaster() and quite a few others are used

Re: [PATCH] xf86-video-dummy allow up to 32767x32767 (re-send)

2011-12-13 Thread Antoine Martin
On 05/25/2011 04:57 PM, Adam Jackson wrote: On 5/19/11 2:17 AM, Antoine Martin wrote: Hi Adam, Can you apply this patch please? I have tested this up to 16384x8192, higher resolutions make xtiming spew out garbage (integer overflow?) so I couldn't generate modelines. Apologies for letting