Re: [PATCH xserver 02/10] glamor: Require that 16bpp pixmap depths match for Render copies.

On 23/09/16 11:08 AM, Michel Dänzer wrote: > On 22/09/16 05:45 PM, Eric Anholt wrote: >> The copy optimization in d37329cba42fa8e72fe4be8a7be18e512268b5bd >> replicated a bug from last time we did a copy optimization, and didn't >> get rendercheck run on it. This is effectively a re-cherry-pick

Re: [PATCH xserver 02/10] glamor: Require that 16bpp pixmap depths match for Render copies.

On 22/09/16 05:45 PM, Eric Anholt wrote: > The copy optimization in d37329cba42fa8e72fe4be8a7be18e512268b5bd > replicated a bug from last time we did a copy optimization, and didn't > get rendercheck run on it. This is effectively a re-cherry-pick of > 510c8605641803f1f5b5d2de6d3bb422b148e0e7. >

[PATCH xbitmaps] Install pkgconfig file into arch-dependent location

Signed-off-by: Heiko Becker --- Makefile.am | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/Makefile.am b/Makefile.am index 060112d..9a89bbe 100644 --- a/Makefile.am +++ b/Makefile.am @@ -1,4 +1,4 @@ -pkgconfigdir = $(datadir)/pkgconfig +pkgconfigdir =

[PATCH 1/1] Simplify, statement is always true.

pVideo->bus is uint8_t, always less than 256. Signed-off-by: Maya Rashish --- hw/xfree86/common/xf86pciBus.c | 6 +- 1 file changed, 1 insertion(+), 5 deletions(-) diff --git a/hw/xfree86/common/xf86pciBus.c b/hw/xfree86/common/xf86pciBus.c index 8158c2b..bed4434 100644 ---

Re: [PATCH x11proto] Fix typo __has_extenstion -> __has_extension

Crap, my bad. Sorry. Reviewed-by: Jeremy Huddleston Sequoia Sent from my iPhone... > On Sep 22, 2016, at 16:38, Keith Packard wrote: > > Signed-off-by: Keith Packard > --- > Xfuncproto.h.in | 2 +- > 1 file changed, 1 insertion(+), 1

Re: [PULL] XQuartz cleanup related to inputthread changes

Jeremy Huddleston Sequoia writes: > [ Unknown signature status ] > Thanks for the reviews. Here is a pull request for the changes I've > got lined up for XQuartz. Note that I still have a few pending > patches out for review (which aren't part of this PR): There was a typo

[PATCH x11proto] Fix typo __has_extenstion -> __has_extension

Signed-off-by: Keith Packard --- Xfuncproto.h.in | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/Xfuncproto.h.in b/Xfuncproto.h.in index 8715c9d..b88493d 100644 --- a/Xfuncproto.h.in +++ b/Xfuncproto.h.in @@ -83,7 +83,7 @@ in this Software without prior

Re: [PATCH] remove dead code in dummy driver

On 09/21/2016 10:22 AM, Aaron Plattner wrote: On 09/20/2016 02:07 AM, Eric Engestrom wrote: On Tue, Sep 20, 2016 at 01:34:40PM +0700, Antoine Martin wrote: Signed-off-by: Antoine Martin Reviewed-by: Eric Engestrom Looks good to me too

[PATCH xserver 2/2] Eliminate coordinate mode in PolyPoint, PolyLines and FillPolygon ops

Doing this in DIX makes the drivers all simpler and fixes bugs in a couple of wrappers which got CoordModePrevious wrong. Signed-off-by: Keith Packard --- dix/dispatch.c | 36 ++-- exa/exa_accel.c| 26 +++--

[PATCH xserver 0/2] Simplify GC ops interface

Removing miTranslate from the GC and support for CoordModePrevious in points/lines/polygons makes drivers simpler and fixes bugs in CoordModePrevious support for dots in damage and rootless. My eventual goal is to make the primitive coordinates parameters const, which will eliminate the need to

[PATCH xserver 1/2] Eliminate miTranslate

miTranslate is expensive to provide for, and not that expensive to handle in the driver. Signed-off-by: Keith Packard --- fb/fbfillsp.c| 6 -- fb/fbgc.c| 2 -- glamor/glamor_glyphblt.c | 6 ++ glamor/glamor_spans.c| 4 ++--

[PULL] XQuartz cleanup related to inputthread changes

Thanks for the reviews. Here is a pull request for the changes I've got lined up for XQuartz. Note that I still have a few pending patches out for review (which aren't part of this PR): os/connection: Improve abstraction for launchd secure sockets randr: Initialize RandR even if

[ANNOUNCE] xproto 7.0.30

-BEGIN PGP SIGNED MESSAGE- Hash: SHA1 Jeremy Huddleston Sequoia (5): Xfuncproto.h: Define __has_feature() if it isn't already Xfuncproto.h: Define __has_extension() if it isn't already Xfuncproto: Add support for _X_DEPRECATED_MSG() macro Xfuncproto: Add support

[ANNOUNCE] xproto 7.0.30

-BEGIN PGP SIGNED MESSAGE- Hash: SHA1 Jeremy Huddleston Sequoia (5): Xfuncproto.h: Define __has_feature() if it isn't already Xfuncproto.h: Define __has_extension() if it isn't already Xfuncproto: Add support for _X_DEPRECATED_MSG() macro Xfuncproto: Add support

Re: [PATCH 1/1] dix: Silence TSan warnings when checking for pending input

Jeremy Huddleston Sequoia writes: > Ok, sent update for a final r-b. Done. > This also requires an updated release of xproto for the macros. I > sent that series out a few days ago but haven't had a review yet. Done. > Are there any other changes to xproto that might be

Re: [PATCH xproto 1/4] Xfuncproto.h: Define __has_feature() if it isn't already

Jeremy Huddleston Sequoia writes: > Signed-off-by: Jeremy Huddleston Sequoia for the series: Reviewed-by: Keith Packard -- -keith signature.asc Description: PGP signature ___

Re: [PATCH 1/1] dix: Silence TSan warnings when checking for pending input

Jeremy Huddleston Sequoia writes: > V2: Moves InputCheckPending() into dix.h > > Bumps required version of xproto to 7.0.30 LGTM. Reviewed-by: Keith Packard -- -keith signature.asc Description: PGP signature

Re: [PATCH 1/1] dix: Silence TSan warnings when checking for pending input

> On Sep 22, 2016, at 11:26, Keith Packard wrote: > > Jeremy Huddleston Sequoia writes: > >> I would've but we're post API freeze at this point, yes? Would that be ok? > > Yeah, better than duplicating, and it doesn't affect the actual ABI/API > seen

[PATCH 1/1] dix: Silence TSan warnings when checking for pending input

V2: Moves InputCheckPending() into dix.h Bumps required version of xproto to 7.0.30 == WARNING: ThreadSanitizer: data race (pid=4943) Read of size 4 at 0x00010c4e3854 by thread T8: #0 WaitForSomething WaitFor.c:237 (X11.bin+0x00010049216c) #1 Dispatch dispatch.c:413

Re: modesetting: allow switching from software to hardware cursors (v4)

On 16.09.2016 17:49, Michael Thayer wrote: Split this up into the three following patches as suggested by Hans. Many thanks for your patience! Polite ping. Regards, Michael -- Michael Thayer | VirtualBox engineer ORACLE Deutschland B.V. & Co. KG | Werkstr. 24 | D-71384 Weinstadt ORACLE

Re: [PATCH 1/1] dix: Silence TSan warnings when checking for pending input

Jeremy Huddleston Sequoia writes: > I would've but we're post API freeze at this point, yes? Would that be ok? Yeah, better than duplicating, and it doesn't affect the actual ABI/API seen to drivers. -- -keith signature.asc Description: PGP signature

Re: [PATCH xserver 0/2] os: Fix write block handling in poll(2) path

Series is: Reviewed-by: Jeremy Huddleston Sequoia Tested-by: Jeremy Huddleston Sequoia I'd still like to get rid of the large NewOutputPending hammer, but this is good for now. Thanks, Jeremy > On Sep 21, 2016, at 16:53, Keith Packard

Re: [PATCH 1/1] dix: Silence TSan warnings when checking for pending input

> On Sep 22, 2016, at 00:22, Keith Packard wrote: > > Jeremy Huddleston Sequoia writes: > >> +static inline _X_NOTSAN Bool >> +InputCheckPending(void) >> +{ >> +return (*checkForInput[0] != *checkForInput[1]); >> +} > >> +static inline _X_NOTSAN

Re: [PATCH xserver] xwayland: Clear up x_cursor on UnrealizeCursor()

On Thu, Sep 22, 2016 at 08:34:59AM +0200, Olivier Fourdan wrote: > In Xwayland's xwl_unrealize_cursor(), the x_cursor is cleared up only > when a device value is provided to the UnrealizeCursor() routine, but > if the device is NULL as called from FreeCursor(), the corresponding > x_cursor for the

[PATCH xf86-video-dummy v2] Remove pointless empty functions

These functions might be useful in a real driver, but with no hardware, they're pointless. Get rid of them. v2: Rebase, get rid of pointless calls to DUMMYAdjustFrame, return TRUE from DUMMYSwitchMode. Signed-off-by: Aaron Plattner --- src/dummy_driver.c | 44

Re: [PATCH xserver 0/2] os: Fix write block handling in poll(2) path

On Thu, Sep 22, 2016 at 02:53:44AM +0300, Keith Packard wrote: > My previous attempt to fix this just hit the wrong variable to fix > edge triggering; it should have reset the POLLOUT bit in > ospoll->osfds[pos].revents instead of ospoll->fds[pos].revents. With > that fixed, it looks like the

Re: [PATCH xserver 06/10] test: Make the piglit-running script callable with an arbitrary server.

On Thu, Sep 22, 2016 at 11:45:07 +0300, Eric Anholt wrote: > diff --git a/test/scripts/xvfb-piglit.sh b/test/scripts/run-piglit.sh > similarity index 90% > copy from test/scripts/xvfb-piglit.sh > copy to test/scripts/run-piglit.sh > index b775239e34f5..ab37ee3b92ad 100755 > ---

Re: [PATCH xserver 05/10] test: Add a little xinit-like program for starting servers for testing.

On Thu, Sep 22, 2016 at 11:45:06 +0300, Eric Anholt wrote: > +asprintf(_string, "%d", displayfd); > +if (!displayfd_string) > +exit(1); I think you need to check the return value from asprintf, not displayfd_string: When successful, these functions return the number of

Re: [PATCH] remove dead code in dummy driver

On 22/09/16 03:22, Aaron Plattner wrote: > On 09/20/2016 02:07 AM, Eric Engestrom wrote: >> On Tue, Sep 20, 2016 at 01:34:40PM +0700, Antoine Martin wrote: >>> Signed-off-by: Antoine Martin >> >> Reviewed-by: Eric Engestrom > > Looks good to me

[solved] Re: Huge mouse-click delay

also sprach martin f krafft [2016-09-21 13:07 +0200]: > The more I am fiddling with this, the more I am suspecting the > Awesome window manager to be at fault. At least my switch to Git > master could well coincide with the start of these problems. > Moreover, if I kill

[PATCH xserver 2/2] os: Clear saved poll events in listen so that edge triggering works

When a client is marked as write blocked, clear any old 'write ready' bit in the osfds structure so that a new indication of write ready (which is marked as edge trigggered) will trigger the callback. Signed-off-by: Keith Packard --- os/ospoll.c | 8 ++-- 1 file changed,

[PATCH xserver 0/2] os: Fix write block handling in poll(2) path

My previous attempt to fix this just hit the wrong variable to fix edge triggering; it should have reset the POLLOUT bit in ospoll->osfds[pos].revents instead of ospoll->fds[pos].revents. With that fixed, it looks like the poll(2) path is working. ___

[PATCH xserver 1/2] os: Ready clients with pending output aren't flushed, so set NewOutputPending

When a client with pending output is ready (has request data pending), FlushAllOutput will skip it to get all of the requests processed before sending any queued output. That means FlushAllOutput is going to return with some output pending to a client which isn't known to be write blocked. And

Re: [PATCH xserver v2] xwayland: Clear up x_cursor on UnrealizeCursor()

Hi, On 09/22/2016 10:38 AM, Olivier Fourdan wrote: In Xwayland's xwl_unrealize_cursor(), the x_cursor is cleared up only when a device value is provided to the UnrealizeCursor() routine, but if the device is NULL as called from FreeCursor(), the corresponding x_cursor for the xwl_seat is left

Re: [PATCH xserver] os: Eliminate NewOutputPending

Hi, On 09/22/2016 11:34 AM, Keith Packard wrote: By removing clients from output_pending_clients list when flushing is attempted, that list now becomes the list of clients that need flushing, which is equivalent to what NewOutputPending was used for. Clients are added to output_pending_clients

[Bug 97827] (EE) [mi] EQ overflowing -> freeze

https://bugs.freedesktop.org/show_bug.cgi?id=97827 --- Comment #8 from Jérôme --- Done. Thanks for the command line. aptitude -t jessie-backports upgrade '?source-package("^mesa$")~i' Resolving dependencies... The following NEW packages will be installed:

[PATCH xserver] xwayland: Clear up x_cursor on UnrealizeCursor()

In Xwayland's xwl_unrealize_cursor(), the x_cursor is cleared up only when a device value is provided to the UnrealizeCursor() routine, but if the device is NULL as called from FreeCursor(), the corresponding x_cursor for the xwl_seat is left untouched. This might cause a segfault when trying to

[PATCH xserver v2] xwayland: Clear up x_cursor on UnrealizeCursor()

In Xwayland's xwl_unrealize_cursor(), the x_cursor is cleared up only when a device value is provided to the UnrealizeCursor() routine, but if the device is NULL as called from FreeCursor(), the corresponding x_cursor for the xwl_seat is left untouched. This might cause a segfault when trying to

[PATCH xserver 08/10] test: Update piglit HTML even when tests all pass.

I was confused by the behavior I'd written before. keithp and mattst88 responded with shock that I would have made it so surprising, as well. Signed-off-by: Eric Anholt --- test/scripts/run-piglit.sh | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) diff --git

[PATCH xserver 09/10] test: Switch our testing X server to being spawned with simple-xinit.

Once I introduced a second X server being tested, I found that startx hit races in choosing a display. Signed-off-by: Eric Anholt --- test/scripts/run-piglit.sh | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/test/scripts/run-piglit.sh

[PATCH xserver 04/10] test: Handle srcdir != builddir in Xvfb testing.

Signed-off-by: Eric Anholt --- test/Makefile.am| 3 ++- test/scripts/xvfb-piglit.sh | 8 +++- 2 files changed, 9 insertions(+), 2 deletions(-) diff --git a/test/Makefile.am b/test/Makefile.am index 5a35e2bb198f..4ccadf5b0005 100644 --- a/test/Makefile.am +++

[PATCH xserver 05/10] test: Add a little xinit-like program for starting servers for testing.

The normal xinit is racy because it doesn't use -displayfd. This implements the bare minimum for testing purposes, using -displayfd to sequence starting the client, and avoids adding yet another dependency to the server. Signed-off-by: Eric Anholt --- test/.gitignore | 1

[PATCH xserver 01/10] glamor: Fix some awful formatting of some fallback debug code.

This was clearly x-indent.sh damage. Signed-off-by: Eric Anholt --- glamor/glamor_render.c | 12 ++-- 1 file changed, 6 insertions(+), 6 deletions(-) diff --git a/glamor/glamor_render.c b/glamor/glamor_render.c index 64141ac0f324..c590580412c9 100644 ---

[PATCH xserver 10/10] test: Run xts against Xephyr -glamor when present.

Signed-off-by: Eric Anholt --- test/Makefile.am | 6 ++ test/scripts/xephyr-glamor-piglit.sh | 29 + 2 files changed, 35 insertions(+) create mode 100755 test/scripts/xephyr-glamor-piglit.sh diff --git a/test/Makefile.am

[PATCH xserver 06/10] test: Make the piglit-running script callable with an arbitrary server.

Signed-off-by: Eric Anholt --- test/.gitignore| 1 + test/Makefile.am | 1 + test/scripts/{xvfb-piglit.sh => run-piglit.sh} | 12 +++-- test/scripts/xvfb-piglit.sh| 73 ++

[PATCH xserver 02/10] glamor: Require that 16bpp pixmap depths match for Render copies.

The copy optimization in d37329cba42fa8e72fe4be8a7be18e512268b5bd replicated a bug from last time we did a copy optimization, and didn't get rendercheck run on it. This is effectively a re-cherry-pick of 510c8605641803f1f5b5d2de6d3bb422b148e0e7. Fixes rendercheck -t blend -o src -f

[PATCH xserver 00/10] Automated glamor testing

Here's a series that gets us automated headless make check of glamor. The run takes about 10 minutes with llvmpipe on new hardware. Along the way, I had to fix glamor so that more of its tests pass, and fix test result reporting (oops), and hack Xephyr to make the testing fast enough (I do really

[PATCH xserver 07/10] test: Fix parsing of piglit results.

The "dmesg-fail" line was matching our "fail" regex, so if you didn't have those we would ignore fails. Signed-off-by: Eric Anholt --- test/scripts/run-piglit.sh | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) diff --git a/test/scripts/run-piglit.sh

[PATCH xserver 03/10] ephyr: Add a mode for skipping redisplay in glamor.

This speeds up headless testing of Xephyr -glamor with softpipe from "a test per minute or so" to "a test every few seconds". Signed-off-by: Eric Anholt --- hw/kdrive/ephyr/ephyr_glamor_glx.c | 8 hw/kdrive/ephyr/ephyrinit.c| 7 ++- 2 files changed, 14

[PATCH xserver] os: Eliminate NewOutputPending

By removing clients from output_pending_clients list when flushing is attempted, that list now becomes the list of clients that need flushing, which is equivalent to what NewOutputPending was used for. Clients are added to output_pending_clients when data are added to their output buffer and when

Re: [PATCH 1/1] dix: Silence TSan warnings when checking for pending input

Jeremy Huddleston Sequoia writes: > +static inline _X_NOTSAN Bool > +InputCheckPending(void) > +{ > +return (*checkForInput[0] != *checkForInput[1]); > +} > +static inline _X_NOTSAN Bool > +InputCheckPending(void) > +{ > +return (*checkForInput[0] !=