Re: [PATCH 3/3 all repositories] autogen: Set a default subject prefix for patches
On Tue, Jan 24, 2017 at 02:35:34AM +, Emil Velikov wrote: > On 24 January 2017 at 00:40, Peter Hutterer wrote: > > From: Adam Jackson > > > > Per discussion at XDC2015, we want this so we can easily distinguish > > which module a patch is for. There's no way to set this in the > > server-side config, so setting a default at autogen time is about the > > best we can do. > > > > Reviewed-by: Eric Anholt > > Signed-off-by: Adam Jackson > > Signed-off-by: Peter Hutterer > > --- > > autogen.sh | 3 +++ > > 1 file changed, 3 insertions(+) > > > > diff --git a/autogen.sh b/autogen.sh > > index 0006de8..cda30cc 100755 > > --- a/autogen.sh > > +++ b/autogen.sh > > @@ -9,6 +9,9 @@ cd "$srcdir" > > autoreconf -v --install || exit 1 > > cd "$ORIGDIR" || exit $? > > > > +git config --local --get format.subjectPrefix || > > +git config --local format.subjectPrefix "PATCH bigreqsproto" > Seems off, plus a simple basename `pwd` might not always work. that's an artefact of the script I used here on the first repository, i.e. it shows that it's working (I took this patchset from the bigreqsproto repo). the prefix is hardcoded per repository. > Mihail Konev has a patch (message id > <20161014000652.GA12465@localhost>) which includes a tarball with all > the patches. > From a quick skim things seem ok - don't see anything wrong and/or > missing. Fwiw I'd go with that one. ok, I'll have a look at that one too then, thanks. Cheers, Peter > > The rest of the series is (fwiw) > Reviewed-by: Emil Velikov > > Thanks ! > Emil ___ xorg-devel@lists.x.org: X.Org development Archives: http://lists.x.org/archives/xorg-devel Info: https://lists.x.org/mailman/listinfo/xorg-devel
Re: [PATCH util-modular 4/4] release.sh: implement around git worktree
On 24 January 2017 at 01:14, Peter Hutterer wrote:
> On Mon, Jan 23, 2017 at 12:51:31PM +, Emil Velikov wrote:
>> On 23 January 2017 at 04:03, Peter Hutterer wrote:
>> > On Fri, Jan 20, 2017 at 02:19:06PM +, Emil Velikov wrote:
>> >> On 20 January 2017 at 02:49, Peter Hutterer
>> >> wrote:
>> >> > On Thu, Jan 19, 2017 at 07:30:10PM +, Emil Velikov wrote:
>> >> >> From: Emil Velikov
>> >> >>
>> >> >> Months ago, before my work in here there the script had a number of
>> >> >> assumptions:
>> >> >> - autoreconf or alike is ran by the user - slight nuisance esp. when
>> >> >> releasing multiple packages.
>> >> >> - the generated files were compatible with the ones on the server -
>> >> >> rarely and issue, but still
>> >> >> - config.status means 'all the autotools bits were generated
>> >> >> correctly'
>> >> >> - failed to consider if user has multiple config.status - using
>> >> >> ./build-64bit/ and ./build-32bit/ anyone ?
>> >> >> - ...
>> >> >>
>> >> >> With 663364cda5e316a0509ff5869293e3a815b9945f we mitigated a lot of
>> >> >> that
>> >> >> but did not consider
>> >> >> - forcing people to `git clean' is annoying and if not careful one can
>> >> >> purge local files that they want to keep
>> >> >> - the files generated by autotools were left behind
>> >> >>
>> >> >> In order to mitigate those use git worktree. This is more efficient
>> >> >> [both bandwidth and storage wise] than pulling/generating git tarballs,
>> >> >> git clone and friends.
>> >> >>
>> >> >> Cc: Peter Hutterer
>> >> >> Signed-off-by: Emil Velikov
>> >> >> ---
>> >> >> There's a git of noise produced by git worktree and autoreconf. I'm
>> >> >> fine
>> >> >> with 2>/dev/null but I'd leave that to others to decide
>> >> >> ---
>> >> >> release.sh | 34 +++---
>> >> >> 1 file changed, 19 insertions(+), 15 deletions(-)
>> >> >>
>> >> >> diff --git a/release.sh b/release.sh
>> >> >> index 8ed179d..0b3fcbf 100755
>> >> >> --- a/release.sh
>> >> >> +++ b/release.sh
>> >> >> @@ -357,25 +357,24 @@ process_module() {
>> >> >> return 1
>> >> >> fi
>> >> >>
>> >> >> -if [ -e configure ]; then
>> >> >> -echo "Error: the git repository contains configure"
>> >> >> -echo "Did you forget to run git clean -fXd (and git clean
>> >> >> -fxd) ?"
>> >> >> -return 1
>> >> >> -fi
>> >> >> -
>> >> >> -# Create tmpdir for the build
>> >> >> +# Create tmpdir for the release
>> >> >> build_dir=`mktemp -d -p . build.XX`
>> >> >> if [ $? -ne 0 ]; then
>> >> >> -echo "Error: could not create a temporary directory for the
>> >> >> build."
>> >> >> +echo "Error: could not create a temporary directory for the
>> >> >> release"
>> >> >> echo "Do you have coreutils' mktemp ?"
>> >> >> return 1
>> >> >> fi
>> >> >>
>> >> >> -echo "Info: generating configure."
>> >> >> -autoreconf --force --install >/dev/null
>> >> >> +# Worktree removal is intentionally left to the user, due to:
>> >> >> +# - currently we cannot select only one worktree to prune
>> >> >> +# - requires to removal of $build_dir which might contradict
>> >> >> with the
>> >> >> +# user decision to keep some artefacts like tarballs or other
>> >> >> +echo "Info: creating new git worktree."
>> >> >> +git worktree add $build_dir
>> >> >> if [ $? -ne 0 ]; then
>> >> >> -echo "Error: failed to generate configure."
>> >> >> -return 1
>> >> >> + echo "Error: failed to create a git worktree."
>> >> >> + cd $top_src
>> >> >> + return 1
>> >> >
>> >> > there's an indentation error,
>> >> Funky indentation was there (it was a code movement) - we still have a
>> >> handful of tabs that are meant to be 8 spaces.
>> >>
>> >> > but otherwise the series looks good and is
>> >> > Reviewed-by: Peter Hutterer
>> >> >
>> >> Note: after another thought I've went with v2 for 3/4 (having a
>> >> version check and building from there).
>> >>
>> >> > I guess I'll test this on the next release ;)
>> >> >
>> >> Thank you very much. Please let me know if you spot anything odd.
>> >
>> > Ok, tested it on the libXi release and something is off, but haven't found
>> > out what it is yet. You can reproduce it by running a --dry-run on the
>> > libXi
>> > repo, for some reason make distcleancheck fails with
>> >
>> > ERROR: files left in build directory after distclean:
>> > ./config.sub
>> > ./ltmain.sh
>> > ./config.guess
>> > ./install-sh
>> > ./missing
>> > ./depcomp
>> > ./compile
>> > make[1]: *** [distcleancheck] Error 1
>> >
>> > since these are standard files, it's not a libXi issue, must be something
>> > to
>> > do with the worktrees or how we invoke configure. Reverting this patch
>> > makes it work again.
>> >
>> That's strange I haven't seen issues like the ones you mentioned.
>> Fetching a clean copy of libXi and doing "../path/to/release.sh .
>> --dry-run" brings up other issues :-]
>
Re: [PATCH v2 xserver 3/4] dri2: refine dri2_probe_driver_name (v2)
On 23 January 2017 at 06:30, Yu, Qiang wrote: > > Hi, > > If no further comments, can this patch be merged? > It's up-to our maintainers to pull the series. They might be a bit busy with other atm. -Emil ___ xorg-devel@lists.x.org: X.Org development Archives: http://lists.x.org/archives/xorg-devel Info: https://lists.x.org/mailman/listinfo/xorg-devel
Re: [PATCH 3/3 all repositories] autogen: Set a default subject prefix for patches
On 24 January 2017 at 00:40, Peter Hutterer wrote: > From: Adam Jackson > > Per discussion at XDC2015, we want this so we can easily distinguish > which module a patch is for. There's no way to set this in the > server-side config, so setting a default at autogen time is about the > best we can do. > > Reviewed-by: Eric Anholt > Signed-off-by: Adam Jackson > Signed-off-by: Peter Hutterer > --- > autogen.sh | 3 +++ > 1 file changed, 3 insertions(+) > > diff --git a/autogen.sh b/autogen.sh > index 0006de8..cda30cc 100755 > --- a/autogen.sh > +++ b/autogen.sh > @@ -9,6 +9,9 @@ cd "$srcdir" > autoreconf -v --install || exit 1 > cd "$ORIGDIR" || exit $? > > +git config --local --get format.subjectPrefix || > +git config --local format.subjectPrefix "PATCH bigreqsproto" Seems off, plus a simple basename `pwd` might not always work. Mihail Konev has a patch (message id <20161014000652.GA12465@localhost>) which includes a tarball with all the patches. From a quick skim things seem ok - don't see anything wrong and/or missing. Fwiw I'd go with that one. The rest of the series is (fwiw) Reviewed-by: Emil Velikov Thanks ! Emil ___ xorg-devel@lists.x.org: X.Org development Archives: http://lists.x.org/archives/xorg-devel Info: https://lists.x.org/mailman/listinfo/xorg-devel
[PATCH util-modular v3] release.sh: remove $MESA_VERSION from the destination location
From: Emil Velikov
Requested by a number of distribution maintainers. Used starting with
the 17.0 series.
v2: Keep things conditional based on the version.
v3: Also update srv_path, to create the correct path
Signed-off-by: Emil Velikov
---
release.sh | 20
1 file changed, 12 insertions(+), 8 deletions(-)
diff --git a/release.sh b/release.sh
index f5ea397..3eca537 100755
--- a/release.sh
+++ b/release.sh
@@ -574,18 +574,22 @@ process_module() {
elif [ x"$section" = xmesa ]; then
host_current=$host_mesa
mesa_version=`echo $pkg_version | sed 's:-rc.*::'`
-section_path=archive/$mesa_version
+section_path=archive
srv_path="/srv/$host_current/www/$section_path"
list_to=$list_mesa_announce
list_cc=$list_mesa_devel
-# Mesa uses separate folder for each release
-echo "Info: creating mesa directory on web server:"
-ssh $USER_NAME$hostname mkdir -p $srv_path &>/dev/null
-if [ $? -ne 0 ]; then
-echo "Error: cannot create the path \"$srv_path\" on the web
server."
-cd $top_src
-return 1
+# Prior to 17.0.x Mesa uses separate folder for each release
+if test `echo $mesa_version | cut -d'.' -f1` -lt 17; then
+section_path=$section_path/$mesa_version
+srv_path="/srv/$host_current/www/$section_path"
+echo "Info: creating mesa directory on web server:"
+ssh $USER_NAME$hostname mkdir -p $srv_path &>/dev/null
+if [ $? -ne 0 ]; then
+echo "Error: cannot create the path \"$srv_path\" on the web
server."
+cd $top_src
+return 1
+fi
fi
fi
--
2.11.0
___
xorg-devel@lists.x.org: X.Org development
Archives: http://lists.x.org/archives/xorg-devel
Info: https://lists.x.org/mailman/listinfo/xorg-devel
Re: [PATCH util-modular 4/4] release.sh: implement around git worktree
On Mon, Jan 23, 2017 at 12:51:31PM +, Emil Velikov wrote:
> On 23 January 2017 at 04:03, Peter Hutterer wrote:
> > On Fri, Jan 20, 2017 at 02:19:06PM +, Emil Velikov wrote:
> >> On 20 January 2017 at 02:49, Peter Hutterer
> >> wrote:
> >> > On Thu, Jan 19, 2017 at 07:30:10PM +, Emil Velikov wrote:
> >> >> From: Emil Velikov
> >> >>
> >> >> Months ago, before my work in here there the script had a number of
> >> >> assumptions:
> >> >> - autoreconf or alike is ran by the user - slight nuisance esp. when
> >> >> releasing multiple packages.
> >> >> - the generated files were compatible with the ones on the server -
> >> >> rarely and issue, but still
> >> >> - config.status means 'all the autotools bits were generated correctly'
> >> >> - failed to consider if user has multiple config.status - using
> >> >> ./build-64bit/ and ./build-32bit/ anyone ?
> >> >> - ...
> >> >>
> >> >> With 663364cda5e316a0509ff5869293e3a815b9945f we mitigated a lot of that
> >> >> but did not consider
> >> >> - forcing people to `git clean' is annoying and if not careful one can
> >> >> purge local files that they want to keep
> >> >> - the files generated by autotools were left behind
> >> >>
> >> >> In order to mitigate those use git worktree. This is more efficient
> >> >> [both bandwidth and storage wise] than pulling/generating git tarballs,
> >> >> git clone and friends.
> >> >>
> >> >> Cc: Peter Hutterer
> >> >> Signed-off-by: Emil Velikov
> >> >> ---
> >> >> There's a git of noise produced by git worktree and autoreconf. I'm fine
> >> >> with 2>/dev/null but I'd leave that to others to decide
> >> >> ---
> >> >> release.sh | 34 +++---
> >> >> 1 file changed, 19 insertions(+), 15 deletions(-)
> >> >>
> >> >> diff --git a/release.sh b/release.sh
> >> >> index 8ed179d..0b3fcbf 100755
> >> >> --- a/release.sh
> >> >> +++ b/release.sh
> >> >> @@ -357,25 +357,24 @@ process_module() {
> >> >> return 1
> >> >> fi
> >> >>
> >> >> -if [ -e configure ]; then
> >> >> -echo "Error: the git repository contains configure"
> >> >> -echo "Did you forget to run git clean -fXd (and git clean
> >> >> -fxd) ?"
> >> >> -return 1
> >> >> -fi
> >> >> -
> >> >> -# Create tmpdir for the build
> >> >> +# Create tmpdir for the release
> >> >> build_dir=`mktemp -d -p . build.XX`
> >> >> if [ $? -ne 0 ]; then
> >> >> -echo "Error: could not create a temporary directory for the
> >> >> build."
> >> >> +echo "Error: could not create a temporary directory for the
> >> >> release"
> >> >> echo "Do you have coreutils' mktemp ?"
> >> >> return 1
> >> >> fi
> >> >>
> >> >> -echo "Info: generating configure."
> >> >> -autoreconf --force --install >/dev/null
> >> >> +# Worktree removal is intentionally left to the user, due to:
> >> >> +# - currently we cannot select only one worktree to prune
> >> >> +# - requires to removal of $build_dir which might contradict with
> >> >> the
> >> >> +# user decision to keep some artefacts like tarballs or other
> >> >> +echo "Info: creating new git worktree."
> >> >> +git worktree add $build_dir
> >> >> if [ $? -ne 0 ]; then
> >> >> -echo "Error: failed to generate configure."
> >> >> -return 1
> >> >> + echo "Error: failed to create a git worktree."
> >> >> + cd $top_src
> >> >> + return 1
> >> >
> >> > there's an indentation error,
> >> Funky indentation was there (it was a code movement) - we still have a
> >> handful of tabs that are meant to be 8 spaces.
> >>
> >> > but otherwise the series looks good and is
> >> > Reviewed-by: Peter Hutterer
> >> >
> >> Note: after another thought I've went with v2 for 3/4 (having a
> >> version check and building from there).
> >>
> >> > I guess I'll test this on the next release ;)
> >> >
> >> Thank you very much. Please let me know if you spot anything odd.
> >
> > Ok, tested it on the libXi release and something is off, but haven't found
> > out what it is yet. You can reproduce it by running a --dry-run on the libXi
> > repo, for some reason make distcleancheck fails with
> >
> > ERROR: files left in build directory after distclean:
> > ./config.sub
> > ./ltmain.sh
> > ./config.guess
> > ./install-sh
> > ./missing
> > ./depcomp
> > ./compile
> > make[1]: *** [distcleancheck] Error 1
> >
> > since these are standard files, it's not a libXi issue, must be something to
> > do with the worktrees or how we invoke configure. Reverting this patch
> > makes it work again.
> >
> That's strange I haven't seen issues like the ones you mentioned.
> Fetching a clean copy of libXi and doing "../path/to/release.sh .
> --dry-run" brings up other issues :-]
ok, verified, sorry. clean clone of the repo seems to succeed, at least with
--dry-run.
> Which version of autotools are you using ? See libtool --help
>
> Using Arch here, packages patched as mentioned:
>
libx11: Issues with Data32/_XData32
Hi, I've been debugging an issue in gtk2-perl causing it to SIGBUS on sparc64, and traced it back to what seems to be dodgy code inside libx11. One of the tests calls gdk_window_set_opacity, which calls XChangeProperty with a pointer to a guint32, cast to char*, with the length set to 32 bits as expected. However, this data pointer then gets cast to a (long *) on line 83 of ChProp.c when calling Data32. Indeed, the definition of Data32 specifies that data is a (long *) on sparc64, since LONG64 is defined. This feels very wrong, but in and of itself is not too bad (I believe it violates strict aliasing). The problem really comes inside the implementation of _XData32. The destination buffer, buf, is an (int *), but data is still a (long *). These are not the same size. The issues with this are as follows: 1. Incrementing data increases the pointer by 8, so it skips over every other 4-byte int, and reads twice as many bytes as it should. 2. On big-endian systems, (int)*(long *)an_int_pointer will end up getting the 4 bytes *after* an_int_pointer, which means it gets completely the wrong data even in the case of len being 4 (bytes). 3. On sparc64, pointers have strict alignment requirements - they must be size aligned (ie (int *) must be 4-byte aligned, (long *) must be 8-byte aligned). In this case, the data pointer (which came all the way from gdk_window_set_opacity) is only 4-byte aligned (which is fine, since it's a pointer to a guint32), but it has been cast to a (long *) and dereferenced, so it is required to be 8-byte aligned. This is the cause of the observed crash. However, I wonder how 1. is not seen currently given the wide use of X11 on amd64. I can only assume 2. not being seen is because the only 64-bit big-endian architectures are generally used for servers, and there aren't many of them. I don't know what the solution to the problem is. There are various places in the call stack where this could be fixed up. The "correct" fix seems to be to change Data32 to take an (int *) (or some other data type of your choosing if you're worried about int's size not being portable) and fix up the casts at all the call sites, but this is an intrusive change to a public header, and I worry that there are things out there relying on the current behaviour. Alternatively, Data32 could be altered to still take a (long *), but cast it back to an (int *). This has the advantage of not touching public headers, but the prototype is then lying about what it's actually doing, and this still has the problem of breaking behaviour. The final, ugly alternative, is to alter XChangeProperty so it makes a copy of data as an array of long, padding or sign-extending each element, before passing it to Data32. I can't claim to have spent much time looking through the code, so it's highly likely I've missed something. Could those with more knowledge please comment on the above? Thanks, James ___ xorg-devel@lists.x.org: X.Org development Archives: http://lists.x.org/archives/xorg-devel Info: https://lists.x.org/mailman/listinfo/xorg-devel
Re: [PATCH] libXinerama: Set number to 0 on error.
On Mon, Jan 23, 2017 at 11:52:13AM -0500, Adam Jackson wrote: > Not that any caller has likely made this mistake, but you want an if > (number) before this, otherwise you turn a protocol error into a > segfault. If a caller supplies NULL, a segfault would always occur because the pointer is never checked for NULL. Tobias ___ xorg-devel@lists.x.org: X.Org development Archives: http://lists.x.org/archives/xorg-devel Info: https://lists.x.org/mailman/listinfo/xorg-devel
Re: [Mesa-dev] Time to update GSoC/EVoC ideas page
On Fri, Jan 20, 2017 at 08:44:19AM -0800, Jason Ekstrand wrote: > On Fri, Jan 20, 2017 at 2:15 AM, Nicolai Hähnle wrote: > > > Hi Rob, > > > > On 19.01.2017 23:32, Rob Clark wrote: > > > >> Just a friendly reminder that now would be a good time to update the > >> wiki page for GSoC/EVoC ideas: > >> > >> https://www.x.org/wiki/SummerOfCodeIdeas/ > >> > >> There are currently still some stale ideas there (and probably plenty > >> of missing good ideas). > >> > >> Also, I've added a "Potential Mentors" section. Please add yourself > >> if you are willing to be a mentor (and what sorts of projects you > >> would be willing to mentor) > >> > > > > I'd be happy to be listed as a possible mentor on the DriConf replacement > > project (nha on IRC/freenode), but I either don't have a Wiki account or > > forgot it a long time ago. Could you put my name down on the page? > > > > Pro tip: You can just git clone the wiki, change a markdown file, and git > push it back up. That's the way I always make edits. Way nicer than a web > GUI. :-) And if you have a shell account, it's easy to resurrect your web wiki account: https://wiki.freedesktop.org/sitewranglers/wiki/401/ Cheers, Daniel -- Daniel Vetter Software Engineer, Intel Corporation http://blog.ffwll.ch ___ xorg-devel@lists.x.org: X.Org development Archives: http://lists.x.org/archives/xorg-devel Info: https://lists.x.org/mailman/listinfo/xorg-devel
[PATCH 2/3 all repositories] autogen.sh: use quoted string variables
From: Emil Velikov Place quotes around the $srcdir, $ORIGDIR and $0 variables to prevent fall-outs, when they contain space. Signed-off-by: Emil Velikov Reviewed-by: Peter Hutterer Signed-off-by: Peter Hutterer --- autogen.sh | 8 1 file changed, 4 insertions(+), 4 deletions(-) diff --git a/autogen.sh b/autogen.sh index fd9c59a..0006de8 100755 --- a/autogen.sh +++ b/autogen.sh @@ -1,14 +1,14 @@ #! /bin/sh -srcdir=`dirname $0` +srcdir=`dirname "$0"` test -z "$srcdir" && srcdir=. ORIGDIR=`pwd` -cd $srcdir +cd "$srcdir" autoreconf -v --install || exit 1 -cd $ORIGDIR || exit $? +cd "$ORIGDIR" || exit $? if test -z "$NOCONFIGURE"; then -exec $srcdir/configure "$@" +exec "$srcdir"/configure "$@" fi -- 2.9.3 ___ xorg-devel@lists.x.org: X.Org development Archives: http://lists.x.org/archives/xorg-devel Info: https://lists.x.org/mailman/listinfo/xorg-devel
[PATCH 3/3 all repositories] autogen: Set a default subject prefix for patches
From: Adam Jackson Per discussion at XDC2015, we want this so we can easily distinguish which module a patch is for. There's no way to set this in the server-side config, so setting a default at autogen time is about the best we can do. Reviewed-by: Eric Anholt Signed-off-by: Adam Jackson Signed-off-by: Peter Hutterer --- autogen.sh | 3 +++ 1 file changed, 3 insertions(+) diff --git a/autogen.sh b/autogen.sh index 0006de8..cda30cc 100755 --- a/autogen.sh +++ b/autogen.sh @@ -9,6 +9,9 @@ cd "$srcdir" autoreconf -v --install || exit 1 cd "$ORIGDIR" || exit $? +git config --local --get format.subjectPrefix || +git config --local format.subjectPrefix "PATCH bigreqsproto" + if test -z "$NOCONFIGURE"; then exec "$srcdir"/configure "$@" fi -- 2.9.3 ___ xorg-devel@lists.x.org: X.Org development Archives: http://lists.x.org/archives/xorg-devel Info: https://lists.x.org/mailman/listinfo/xorg-devel
[PATCH 0/3 all repositories] autogen.sh: sync with server
Heads up: I'll be pushing the following three patches into all xorg repositories that need it in the next few days. Speak up now or be forever silent, etc. etc. Cheers, Peter ___ xorg-devel@lists.x.org: X.Org development Archives: http://lists.x.org/archives/xorg-devel Info: https://lists.x.org/mailman/listinfo/xorg-devel
[PATCH 1/3 all repositories] autogen.sh: use exec instead of waiting for configure to finish
Syncs the invocation of configure with the one from the server. Signed-off-by: Peter Hutterer --- autogen.sh | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/autogen.sh b/autogen.sh index fc34bd5..fd9c59a 100755 --- a/autogen.sh +++ b/autogen.sh @@ -10,5 +10,5 @@ autoreconf -v --install || exit 1 cd $ORIGDIR || exit $? if test -z "$NOCONFIGURE"; then -$srcdir/configure "$@" +exec $srcdir/configure "$@" fi -- 2.9.3 ___ xorg-devel@lists.x.org: X.Org development Archives: http://lists.x.org/archives/xorg-devel Info: https://lists.x.org/mailman/listinfo/xorg-devel
Re: [PATCH xserver 00/19] Module loader cleanup
On Mon, Jan 23, 2017 at 14:32:14 -0500, Adam Jackson wrote: > I'd like to move the module loader up to dix. In preparation for that, here's > a bunch of cleanup patches. The first three aren't mine, I just think they're > neat. > A few nits in separate mails, but for the series: Reviewed-by: Julien Cristau Cheers, Julien ___ xorg-devel@lists.x.org: X.Org development Archives: http://lists.x.org/archives/xorg-devel Info: https://lists.x.org/mailman/listinfo/xorg-devel
Re: [PATCH xserver 19/19] loader: Fix an obviously backwards comparison
On Mon, Jan 23, 2017 at 14:32:33 -0500, Adam Jackson wrote:
> Hmm, this code looks wrong to me. I know! Instead of fix it, I'll put a
> comment here that nobody will look at for decades on end.
>
> Yaaay xfree86.
>
> Signed-off-by: Adam Jackson
> ---
> hw/xfree86/loader/loadmod.c | 3 +--
> 1 file changed, 1 insertion(+), 2 deletions(-)
>
> diff --git a/hw/xfree86/loader/loadmod.c b/hw/xfree86/loader/loadmod.c
> index ca77c7a..55bec14 100644
> --- a/hw/xfree86/loader/loadmod.c
> +++ b/hw/xfree86/loader/loadmod.c
> @@ -545,8 +545,7 @@ CheckVersion(const char *module, XF86ModuleVersionInfo *
> data,
> "(%d)\n", module, maj, reqmaj);
> return FALSE;
> }
> -/* XXX Maybe this should be the other way around? */
> -if (min > reqmin) {
> +if (min < reqmin) {
> LogMessageVerb(X_WARNING, 2, "%s: module ABI minor version "
> "(%d) is newer than that available (%d)\n",
> module, min, reqmin);
I guess the message should be fixed too?
Cheers,
Julien
___
xorg-devel@lists.x.org: X.Org development
Archives: http://lists.x.org/archives/xorg-devel
Info: https://lists.x.org/mailman/listinfo/xorg-devel
Re: [PATCH xserver 06/19] loader: Remove *GetOS
On Mon, Jan 23, 2017 at 14:32:20 -0500, Adam Jackson wrote: > This API is dumb. uname(3) exists, feel free to use it, but ideally > write to the interface not to the OS. There are a couple of drivers > using this API, they could all reasonably just not. > It's uname(2) AFAICT :) Cheers, Julien ___ xorg-devel@lists.x.org: X.Org development Archives: http://lists.x.org/archives/xorg-devel Info: https://lists.x.org/mailman/listinfo/xorg-devel
Re: [PATCH xserver 02/19] xfree86: remove unused path from the LoadModule API
On 01/23/2017 01:27 PM, Eric Anholt wrote: > Adam Jackson writes: > >> From: Emil Velikov >> >> Similar to its little brothre - LoadSubModule. Currently all call sites >> provide NULL anyway ;-) > > "brother". Other than that, > > Reviewed-by: Eric Anholt Reviewed-by: Aaron Plattner We don't use LoadModule in our driver, and if we do start loading extra stuff, it'll probably be with LoadSubModule anyway. -- Aaron ___ xorg-devel@lists.x.org: X.Org development Archives: http://lists.x.org/archives/xorg-devel Info: https://lists.x.org/mailman/listinfo/xorg-devel
Re: [PATCH xserver 04/19] xfree86: flatten pathlist management in the loader
On Mon, Jan 23, 2017 at 14:32:18 -0500, Adam Jackson wrote: > From: Emil Velikov > > Now that used can set the path only via LoaderSetPath(), we can simplify > things. > used -> users, maybe? Cheers, Julien ___ xorg-devel@lists.x.org: X.Org development Archives: http://lists.x.org/archives/xorg-devel Info: https://lists.x.org/mailman/listinfo/xorg-devel
Re: [PATCH xserver 04/19] xfree86: flatten pathlist management in the loader
Adam Jackson writes: > From: Emil Velikov > > Now that used can set the path only via LoaderSetPath(), we can simplify "users"? Other than that, 4-14 are: Reviewed-by: Eric Anholt There is so much "wow, remember when that was a thing?" in this series. Time to go get some lunch before I see if I can get through the remaining few patches. signature.asc Description: PGP signature ___ xorg-devel@lists.x.org: X.Org development Archives: http://lists.x.org/archives/xorg-devel Info: https://lists.x.org/mailman/listinfo/xorg-devel
Re: [PATCH xserver 01/19] xfree86: remove references from LoadSubModule's path from the doc
On Mon, Jan 23, 2017 at 14:32:15 -0500, Adam Jackson wrote: > From: Emil Velikov > > Afaics the argument hasn't been part of the API since the documentation > has been converted to xml with commit fc6ebe1e1d3 "Convert LinuxDoc > documents to DocBook/XML" > The commit message seems confused. AFAICT the text says "LoadSubModule is like LoadModule except it doesn't take a 'path' parameter", which until patch 2 seems to be correct? Cheers, Julien > Signed-off-by: Emil Velikov > --- > hw/xfree86/doc/ddxDesign.xml | 6 +- > 1 file changed, 1 insertion(+), 5 deletions(-) > > diff --git a/hw/xfree86/doc/ddxDesign.xml b/hw/xfree86/doc/ddxDesign.xml > index f7d6628..b62272d 100644 > --- a/hw/xfree86/doc/ddxDesign.xml > +++ b/hw/xfree86/doc/ddxDesign.xml > @@ -5695,11 +5695,7 @@ the server, and may also be used from within modules. > described above, except that the module loaded is registered as a > child of the calling module. The parent parameter > is the calling module's handle. Modules loaded with this function > -are automatically unloaded when the parent module is unloaded. The > -other difference is that the path parameter may not be specified. > -The module search path used for modules loaded with this function > -is the default search path as initialised with > -LoaderSetPath(). > +are automatically unloaded when the parent module is unloaded. > > > > -- > 2.9.3 > > ___ > xorg-devel@lists.x.org: X.Org development > Archives: http://lists.x.org/archives/xorg-devel > Info: https://lists.x.org/mailman/listinfo/xorg-devel ___ xorg-devel@lists.x.org: X.Org development Archives: http://lists.x.org/archives/xorg-devel Info: https://lists.x.org/mailman/listinfo/xorg-devel
Re: [PATCH:xdm] Install xdm man pages under admin section (8), not user programs (1)
On Mon, Jan 16, 2017 at 16:26:08 -0800, Alan Coopersmith wrote: > As best I can tell, it was historically under section 1 mainly because > the old X Consortium Imake configs only supporting installing program > man pages there, and didn't have an option for using other sections. > > Signed-off-by: Alan Coopersmith > --- > > I'm sending this out not just for review that I made the right changes, > but to see if there's general agreement this change is right. It was > pointed out by one of our users who thought it odd that gdm's man page > is in section 8, but xdm's is in section 1. > That seems right to me. Cheers, Julien ___ xorg-devel@lists.x.org: X.Org development Archives: http://lists.x.org/archives/xorg-devel Info: https://lists.x.org/mailman/listinfo/xorg-devel
Re: [PATCH xserver 02/19] xfree86: remove unused path from the LoadModule API
Adam Jackson writes: > From: Emil Velikov > > Similar to its little brothre - LoadSubModule. Currently all call sites > provide NULL anyway ;-) "brother". Other than that, Reviewed-by: Eric Anholt signature.asc Description: PGP signature ___ xorg-devel@lists.x.org: X.Org development Archives: http://lists.x.org/archives/xorg-devel Info: https://lists.x.org/mailman/listinfo/xorg-devel
Re: [PATCH xserver 01/19] xfree86: remove references from LoadSubModule's path from the doc
Adam Jackson writes: > From: Emil Velikov > > Afaics the argument hasn't been part of the API since the documentation > has been converted to xml with commit fc6ebe1e1d3 "Convert LinuxDoc > documents to DocBook/XML" In the subject, "remove references *to* LoadSubModule's path"? Potential followup: ddxDesign.xml's LoaderSetPath() docs refer to these functions taking a path, when they don't any more. Reviewed-by: Eric Anholt signature.asc Description: PGP signature ___ xorg-devel@lists.x.org: X.Org development Archives: http://lists.x.org/archives/xorg-devel Info: https://lists.x.org/mailman/listinfo/xorg-devel
[PATCH xserver 00/19] Module loader cleanup
I'd like to move the module loader up to dix. In preparation for that, here's a bunch of cleanup patches. The first three aren't mine, I just think they're neat. b/hw/xfree86/common/xf86.h |2 b/hw/xfree86/common/xf86Config.c| 78 - b/hw/xfree86/common/xf86Config.h|1 b/hw/xfree86/common/xf86Configure.c | 103 +-- b/hw/xfree86/common/xf86Helper.c| 19 - b/hw/xfree86/common/xf86Init.c |6 b/hw/xfree86/common/xf86Module.h| 31 -- b/hw/xfree86/common/xf86str.h | 10 b/hw/xfree86/doc/ddxDesign.xml | 81 - b/hw/xfree86/loader/Makefile.am |3 b/hw/xfree86/loader/loader.c| 47 +-- b/hw/xfree86/loader/loader.h|2 b/hw/xfree86/loader/loaderProcs.h | 10 b/hw/xfree86/loader/loadmod.c | 495 +--- hw/xfree86/loader/os.c | 72 - 15 files changed, 262 insertions(+), 698 deletions(-) - ajax ___ xorg-devel@lists.x.org: X.Org development Archives: http://lists.x.org/archives/xorg-devel Info: https://lists.x.org/mailman/listinfo/xorg-devel
[PATCH xserver 02/19] xfree86: remove unused path from the LoadModule API
From: Emil Velikov
Similar to its little brothre - LoadSubModule. Currently all call sites
provide NULL anyway ;-)
Cc: Aaron Plattner
Signed-off-by: Emil Velikov
---
hw/xfree86/common/xf86Helper.c | 2 +-
hw/xfree86/common/xf86Init.c| 2 +-
hw/xfree86/doc/ddxDesign.xml| 8 +---
hw/xfree86/loader/loaderProcs.h | 2 +-
hw/xfree86/loader/loadmod.c | 7 +++
5 files changed, 7 insertions(+), 14 deletions(-)
diff --git a/hw/xfree86/common/xf86Helper.c b/hw/xfree86/common/xf86Helper.c
index 6f3a608..9388436 100644
--- a/hw/xfree86/common/xf86Helper.c
+++ b/hw/xfree86/common/xf86Helper.c
@@ -1615,7 +1615,7 @@ xf86LoadOneModule(const char *name, void *opt)
return NULL;
}
-mod = LoadModule(Name, NULL, NULL, NULL, opt, NULL, &errmaj, &errmin);
+mod = LoadModule(Name, NULL, NULL, opt, NULL, &errmaj, &errmin);
if (!mod)
LoaderErrorMsg(NULL, Name, errmaj, errmin);
free(Name);
diff --git a/hw/xfree86/common/xf86Init.c b/hw/xfree86/common/xf86Init.c
index a544b65..f6f77c0 100644
--- a/hw/xfree86/common/xf86Init.c
+++ b/hw/xfree86/common/xf86Init.c
@@ -1503,7 +1503,7 @@ xf86LoadModules(const char **list, void **optlist)
else
opt = NULL;
-if (!LoadModule(name, NULL, NULL, NULL, opt, NULL, &errmaj, &errmin)) {
+if (!LoadModule(name, NULL, NULL, opt, NULL, &errmaj, &errmin)) {
LoaderErrorMsg(NULL, name, errmaj, errmin);
failed = TRUE;
}
diff --git a/hw/xfree86/doc/ddxDesign.xml b/hw/xfree86/doc/ddxDesign.xml
index b62272d..2ff2894 100644
--- a/hw/xfree86/doc/ddxDesign.xml
+++ b/hw/xfree86/doc/ddxDesign.xml
@@ -5210,7 +5210,7 @@ XFree86 common layer.
-pointer LoadModule(const char *module, const char *path,
+pointer LoadModule(const char *module,
const char **subdirlist, const char **patternlist,
pointer options, const XF86ModReqInfo * modreq,
int *errmaj, int *errmin);
@@ -5226,12 +5226,6 @@ XFree86 common layer.
This might change. The other parameters are:
-
- path
-
- An optional comma-separated list of module search paths.
- When NULL, the default search path is
used.
-
diff --git a/hw/xfree86/loader/loaderProcs.h b/hw/xfree86/loader/loaderProcs.h
index cfc4d80..8d7872f 100644
--- a/hw/xfree86/loader/loaderProcs.h
+++ b/hw/xfree86/loader/loaderProcs.h
@@ -74,7 +74,7 @@ void LoaderInit(void);
ModuleDescPtr LoadDriver(const char *, const char *, int, void *, int *,
int *);
-ModuleDescPtr LoadModule(const char *, const char *, const char **,
+ModuleDescPtr LoadModule(const char *, const char **,
const char **, void *, const XF86ModReqInfo *,
int *, int *);
ModuleDescPtr DuplicateModule(ModuleDescPtr mod, ModuleDescPtr parent);
diff --git a/hw/xfree86/loader/loadmod.c b/hw/xfree86/loader/loadmod.c
index 940f5fc..5e4d7da 100644
--- a/hw/xfree86/loader/loadmod.c
+++ b/hw/xfree86/loader/loadmod.c
@@ -751,7 +751,7 @@ LoadSubModule(void *_parent, const char *module,
return NULL;
}
-submod = LoadModule(module, NULL, subdirlist, patternlist, options,
+submod = LoadModule(module, subdirlist, patternlist, options,
modreq, errmaj, errmin);
if (submod && submod != (ModuleDescPtr) 1) {
parent->child = AddSibling(parent->child, submod);
@@ -821,7 +821,6 @@ static const char *compiled_in_modules[] = {
* module The module name. Normally this is not a filename but the
* module's "canonical name. A full pathname is, however,
* also accepted.
- * path A comma separated list of module directories.
* subdirlist A NULL terminated list of subdirectories to search. When
* NULL, the default "stdSubdirs" list is used. The default
* list is also substituted for entries with value DEFAULT_LIST.
@@ -853,7 +852,7 @@ static const char *compiled_in_modules[] = {
*
*/
ModuleDescPtr
-LoadModule(const char *module, const char *path, const char **subdirlist,
+LoadModule(const char *module, const char **subdirlist,
const char **patternlist, void *options,
const XF86ModReqInfo * modreq, int *errmaj, int *errmin)
{
@@ -909,7 +908,7 @@ LoadModule(const char *module, const char *path, const char
**subdirlist,
goto LoadModule_fail;
}
-pathlist = InitPathList(path);
+pathlist = InitPathList(NULL);
if (!pathlist) {
/* This could be a malloc failure too */
if (errmaj)
--
2.9.3
___
xorg-devel@lists.x.org: X.Org development
Archives: http://lists.x.org/archives/xorg-devel
Info: https://lists.x.org/mailman/listinfo/xorg-devel
[PATCH xserver 18/19] loader: Learn about the joy of snprintf
Signed-off-by: Adam Jackson
---
hw/xfree86/loader/loadmod.c | 11 ++-
1 file changed, 2 insertions(+), 9 deletions(-)
diff --git a/hw/xfree86/loader/loadmod.c b/hw/xfree86/loader/loadmod.c
index cb86925..ca77c7a 100644
--- a/hw/xfree86/loader/loadmod.c
+++ b/hw/xfree86/loader/loadmod.c
@@ -335,10 +335,7 @@ FindModule(const char *module, const char *dirname,
PatternPtr patterns)
return NULL;
for (s = stdSubdirs; *s; s++) {
-if ((strlen(dirname) + strlen(*s)) > PATH_MAX)
-continue;
-strcpy(buf, dirname);
-strcat(buf, *s);
+snprintf(buf, PATH_MAX, "%s%s", dirname, *s);
if ((name = FindModuleInSubdir(buf, module)))
break;
}
@@ -371,11 +368,7 @@ LoaderListDir(const char *subdir, const char **patternlist)
goto bail;
for (elem = pathlist; *elem; elem++) {
-if ((dirlen = strlen(*elem) + strlen(subdir) + 1) > PATH_MAX)
-continue;
-strcpy(buf, *elem);
-strcat(buf, "/");
-strcat(buf, subdir);
+dirlen = snprintf(buf, PATH_MAX, "%s/%s", *elem, subdir);
fp = buf + dirlen;
if (stat(buf, &stat_buf) == 0 && S_ISDIR(stat_buf.st_mode) &&
(d = opendir(buf))) {
--
2.9.3
___
xorg-devel@lists.x.org: X.Org development
Archives: http://lists.x.org/archives/xorg-devel
Info: https://lists.x.org/mailman/listinfo/xorg-devel
[PATCH xserver 17/19] loader: Remove unused path and name from ModuleDescPtr
Just a waste of memory. Path was never referenced at all, and name was
only used when unloading the module; we can just as well get the
module's internal idea of its name from VersionInfo.
Signed-off-by: Adam Jackson
---
hw/xfree86/loader/loaderProcs.h | 2 --
hw/xfree86/loader/loadmod.c | 31 +--
2 files changed, 9 insertions(+), 24 deletions(-)
diff --git a/hw/xfree86/loader/loaderProcs.h b/hw/xfree86/loader/loaderProcs.h
index 84ed103..8b16cb7 100644
--- a/hw/xfree86/loader/loaderProcs.h
+++ b/hw/xfree86/loader/loaderProcs.h
@@ -59,8 +59,6 @@ typedef struct module_desc {
struct module_desc *child;
struct module_desc *sib;
struct module_desc *parent;
-char *name;
-char *path;
void *handle;
ModuleSetupProc SetupProc;
ModuleTearDownProc TearDownProc;
diff --git a/hw/xfree86/loader/loadmod.c b/hw/xfree86/loader/loadmod.c
index fc982c7..cb86925 100644
--- a/hw/xfree86/loader/loadmod.c
+++ b/hw/xfree86/loader/loadmod.c
@@ -600,17 +600,6 @@ LoadSubModule(void *_parent, const char *module,
return submod;
}
-static ModuleDescPtr
-NewModuleDesc(const char *name)
-{
-ModuleDescPtr mdp = calloc(1, sizeof(ModuleDesc));
-
-if (mdp)
-mdp->name = xstrdup(name);
-
-return mdp;
-}
-
ModuleDescPtr
DuplicateModule(ModuleDescPtr mod, ModuleDescPtr parent)
{
@@ -619,7 +608,7 @@ DuplicateModule(ModuleDescPtr mod, ModuleDescPtr parent)
if (!mod)
return NULL;
-ret = NewModuleDesc(mod->name);
+ret = calloc(1, sizeof(ModuleDesc));
if (ret == NULL)
return NULL;
@@ -632,7 +621,6 @@ DuplicateModule(ModuleDescPtr mod, ModuleDescPtr parent)
ret->sib = DuplicateModule(mod->sib, parent);
ret->parent = parent;
ret->VersionInfo = mod->VersionInfo;
-ret->path = strdup(mod->path);
return ret;
}
@@ -726,7 +714,7 @@ LoadModule(const char *module, void *options, const
XF86ModReqInfo *modreq,
*errmaj = LDR_BADUSAGE;
goto LoadModule_fail;
}
-ret = NewModuleDesc(name);
+ret = calloc(1, sizeof(ModuleDesc));
if (!ret) {
if (errmaj)
*errmaj = LDR_NOMEM;
@@ -773,7 +761,6 @@ LoadModule(const char *module, void *options, const
XF86ModReqInfo *modreq,
ret->handle = LoaderOpen(found, errmaj);
if (ret->handle == NULL)
goto LoadModule_fail;
-ret->path = strdup(found);
/* drop any explicit suffix from the module name */
p = strchr(name, '.');
@@ -857,31 +844,31 @@ void
UnloadModule(void *_mod)
{
ModuleDescPtr mod = _mod;
+const char *name;
if (mod == (ModuleDescPtr) 1)
return;
-if (mod == NULL || mod->name == NULL)
+if (mod == NULL)
return;
+name = mod->VersionInfo->modname;
+
if (mod->parent)
-LogMessageVerbSigSafe(X_INFO, 3, "UnloadSubModule: \"%s\"\n",
- mod->name);
+LogMessageVerbSigSafe(X_INFO, 3, "UnloadSubModule: \"%s\"\n", name);
else
-LogMessageVerbSigSafe(X_INFO, 3, "UnloadModule: \"%s\"\n", mod->name);
+LogMessageVerbSigSafe(X_INFO, 3, "UnloadModule: \"%s\"\n", name);
if (mod->TearDownData != ModuleDuplicated) {
if ((mod->TearDownProc) && (mod->TearDownData))
mod->TearDownProc(mod->TearDownData);
-LoaderUnload(mod->name, mod->handle);
+LoaderUnload(name, mod->handle);
}
if (mod->child)
UnloadModule(mod->child);
if (mod->sib)
UnloadModule(mod->sib);
-free(mod->path);
-free(mod->name);
free(mod);
}
--
2.9.3
___
xorg-devel@lists.x.org: X.Org development
Archives: http://lists.x.org/archives/xorg-devel
Info: https://lists.x.org/mailman/listinfo/xorg-devel
[PATCH xserver 14/19] loader: Move loader list details to internal header
There's no reason a driver should ever care about this.
Signed-off-by: Adam Jackson
---
hw/xfree86/common/xf86Config.c| 2 +-
hw/xfree86/common/xf86Configure.c | 1 +
hw/xfree86/common/xf86Module.h| 2 --
hw/xfree86/loader/loaderProcs.h | 3 +++
4 files changed, 5 insertions(+), 3 deletions(-)
diff --git a/hw/xfree86/common/xf86Config.c b/hw/xfree86/common/xf86Config.c
index 21daf1a..3a8f0e1 100644
--- a/hw/xfree86/common/xf86Config.c
+++ b/hw/xfree86/common/xf86Config.c
@@ -60,8 +60,8 @@
#include "globals.h"
#include "extension.h"
#include "xf86pciBus.h"
-
#include "xf86Xinput.h"
+#include "loaderProcs.h"
#include "xkbsrv.h"
diff --git a/hw/xfree86/common/xf86Configure.c
b/hw/xfree86/common/xf86Configure.c
index 0d7a127..e5889b4 100644
--- a/hw/xfree86/common/xf86Configure.c
+++ b/hw/xfree86/common/xf86Configure.c
@@ -40,6 +40,7 @@
#include "xf86Sbus.h"
#endif
#include "misc.h"
+#include "loaderProcs.h"
typedef struct _DevToConfig {
GDevRec GDev;
diff --git a/hw/xfree86/common/xf86Module.h b/hw/xfree86/common/xf86Module.h
index faea07f..a3b2c93 100644
--- a/hw/xfree86/common/xf86Module.h
+++ b/hw/xfree86/common/xf86Module.h
@@ -156,8 +156,6 @@ extern _X_EXPORT void *LoadSubModule(void *, const char *,
const char **,
extern _X_EXPORT void UnloadSubModule(void *);
extern _X_EXPORT void UnloadModule(void *);
extern _X_EXPORT void *LoaderSymbol(const char *);
-extern _X_EXPORT const char **LoaderListDirs(const char **, const char **);
-extern _X_EXPORT void LoaderFreeDirList(char **);
extern _X_EXPORT void LoaderErrorMsg(const char *, const char *, int, int);
extern _X_EXPORT Bool LoaderShouldIgnoreABI(void);
extern _X_EXPORT int LoaderGetABIVersion(const char *abiclass);
diff --git a/hw/xfree86/loader/loaderProcs.h b/hw/xfree86/loader/loaderProcs.h
index ee8a557..eee9a27 100644
--- a/hw/xfree86/loader/loaderProcs.h
+++ b/hw/xfree86/loader/loaderProcs.h
@@ -85,6 +85,9 @@ unsigned long LoaderGetModuleVersion(ModuleDescPtr mod);
void LoaderResetOptions(void);
void LoaderSetOptions(unsigned long);
+const char **LoaderListDirs(const char **, const char **);
+void LoaderFreeDirList(char **);
+
/* Options for LoaderSetOptions */
#define LDR_OPT_ABI_MISMATCH_NONFATAL 0x0001
--
2.9.3
___
xorg-devel@lists.x.org: X.Org development
Archives: http://lists.x.org/archives/xorg-devel
Info: https://lists.x.org/mailman/listinfo/xorg-devel
[PATCH xserver 13/19] xfree86: Remove a stray reference to font modules
Signed-off-by: Adam Jackson
---
hw/xfree86/common/xf86Configure.c | 3 +--
1 file changed, 1 insertion(+), 2 deletions(-)
diff --git a/hw/xfree86/common/xf86Configure.c
b/hw/xfree86/common/xf86Configure.c
index 8c8e49e..0d7a127 100644
--- a/hw/xfree86/common/xf86Configure.c
+++ b/hw/xfree86/common/xf86Configure.c
@@ -399,10 +399,9 @@ configureModuleSection(void)
{
const char **elist, **el;
-/* Find the list of extension & font modules. */
+/* Find the list of extension modules. */
const char *esubdirs[] = {
"extensions",
-"fonts",
NULL
};
parsePrologue(XF86ConfModulePtr, XF86ConfModuleRec);
--
2.9.3
___
xorg-devel@lists.x.org: X.Org development
Archives: http://lists.x.org/archives/xorg-devel
Info: https://lists.x.org/mailman/listinfo/xorg-devel
[PATCH xserver 19/19] loader: Fix an obviously backwards comparison
Hmm, this code looks wrong to me. I know! Instead of fix it, I'll put a
comment here that nobody will look at for decades on end.
Yaaay xfree86.
Signed-off-by: Adam Jackson
---
hw/xfree86/loader/loadmod.c | 3 +--
1 file changed, 1 insertion(+), 2 deletions(-)
diff --git a/hw/xfree86/loader/loadmod.c b/hw/xfree86/loader/loadmod.c
index ca77c7a..55bec14 100644
--- a/hw/xfree86/loader/loadmod.c
+++ b/hw/xfree86/loader/loadmod.c
@@ -545,8 +545,7 @@ CheckVersion(const char *module, XF86ModuleVersionInfo *
data,
"(%d)\n", module, maj, reqmaj);
return FALSE;
}
-/* XXX Maybe this should be the other way around? */
-if (min > reqmin) {
+if (min < reqmin) {
LogMessageVerb(X_WARNING, 2, "%s: module ABI minor version "
"(%d) is newer than that available (%d)\n",
module, min, reqmin);
--
2.9.3
___
xorg-devel@lists.x.org: X.Org development
Archives: http://lists.x.org/archives/xorg-devel
Info: https://lists.x.org/mailman/listinfo/xorg-devel
[PATCH xserver 15/19] loader: Turn LoaderListDirs into LoaderListDir
Callers only ever use this for a single directory anyway.
While we're at it, also move xf86DriverListFromCompile near its only
user in the X -configure code (and inline it out of existence), and
remove LoaderFreeDirList as it's unused (since X -configure is just
going to exit anyway, none of that code cares about cleanup).
Signed-off-by: Adam Jackson
---
hw/xfree86/common/xf86Config.c| 76 --
hw/xfree86/common/xf86Config.h| 1 -
hw/xfree86/common/xf86Configure.c | 75 +
hw/xfree86/loader/loaderProcs.h | 3 +-
hw/xfree86/loader/loadmod.c | 99 +--
5 files changed, 110 insertions(+), 144 deletions(-)
diff --git a/hw/xfree86/common/xf86Config.c b/hw/xfree86/common/xf86Config.c
index 3a8f0e1..f03acf3 100644
--- a/hw/xfree86/common/xf86Config.c
+++ b/hw/xfree86/common/xf86Config.c
@@ -518,82 +518,6 @@ xf86InputDriverlistFromConfig(void)
return modulearray;
}
-static int
-is_fallback(const char *s)
-{
-/* later entries are less preferred */
-const char *fallback[5] = { "modesetting", "fbdev", "vesa", "wsfb", NULL
};
-int i;
-
-for (i = 0; fallback[i]; i++)
- if (strstr(s, fallback[i]))
- return i;
-
-return -1;
-}
-
-static int
-driver_sort(const void *_l, const void *_r)
-{
-const char *l = *(const char **)_l;
-const char *r = *(const char **)_r;
-int left = is_fallback(l);
-int right = is_fallback(r);
-
-/* neither is a fallback, asciibetize */
-if (left == -1 && right == -1)
- return strcmp(l, r);
-
-/* left is a fallback */
-if (left >= 0)
- return 1;
-
-/* right is a fallback */
-if (right >= 0)
- return -1;
-
-/* both are fallbacks, which is worse */
-return left - right;
-}
-
-static void
-fixup_video_driver_list(const char **drivers)
-{
-const char **end;
-
-/* walk to the end of the list */
-for (end = drivers; *end && **end; end++);
-end--;
-
-qsort(drivers, end - drivers, sizeof(const char *), driver_sort);
-}
-
-static const char **
-GenerateDriverlist(const char *dirname)
-{
-const char **ret;
-const char *subdirs[] = { dirname, NULL };
-static const char *patlist[] = { "(.*)_drv\\.so", NULL };
-ret = LoaderListDirs(subdirs, patlist);
-
-/* fix up the probe order for video drivers */
-if (strstr(dirname, "drivers") && ret != NULL)
-fixup_video_driver_list(ret);
-
-return ret;
-}
-
-const char **
-xf86DriverlistFromCompile(void)
-{
-static const char **driverlist = NULL;
-
-if (!driverlist)
-driverlist = GenerateDriverlist("drivers");
-
-return driverlist;
-}
-
static void
configFiles(XF86ConfFilesPtr fileconf)
{
diff --git a/hw/xfree86/common/xf86Config.h b/hw/xfree86/common/xf86Config.h
index 23fb383..bbcb252 100644
--- a/hw/xfree86/common/xf86Config.h
+++ b/hw/xfree86/common/xf86Config.h
@@ -61,7 +61,6 @@ typedef struct _ModuleDefault {
*/
const char **xf86ModulelistFromConfig(void ***);
const char **xf86DriverlistFromConfig(void);
-const char **xf86DriverlistFromCompile(void);
const char **xf86InputDriverlistFromConfig(void);
Bool xf86BuiltinInputDriver(const char *);
ConfigStatus xf86HandleConfigFile(Bool);
diff --git a/hw/xfree86/common/xf86Configure.c
b/hw/xfree86/common/xf86Configure.c
index e5889b4..f975b98 100644
--- a/hw/xfree86/common/xf86Configure.c
+++ b/hw/xfree86/common/xf86Configure.c
@@ -400,14 +400,9 @@ configureModuleSection(void)
{
const char **elist, **el;
-/* Find the list of extension modules. */
-const char *esubdirs[] = {
-"extensions",
-NULL
-};
parsePrologue(XF86ConfModulePtr, XF86ConfModuleRec);
-elist = LoaderListDirs(esubdirs, NULL);
+elist = LoaderListDir("extensions", NULL);
if (elist) {
for (el = elist; *el; el++) {
XF86LoadPtr module;
@@ -534,6 +529,70 @@ configureDDCMonitorSection(int screennum)
return ptr;
}
+static int
+is_fallback(const char *s)
+{
+/* later entries are less preferred */
+const char *fallback[5] = { "modesetting", "fbdev", "vesa", "wsfb", NULL
};
+int i;
+
+for (i = 0; fallback[i]; i++)
+ if (strstr(s, fallback[i]))
+ return i;
+
+return -1;
+}
+
+static int
+driver_sort(const void *_l, const void *_r)
+{
+const char *l = *(const char **)_l;
+const char *r = *(const char **)_r;
+int left = is_fallback(l);
+int right = is_fallback(r);
+
+/* neither is a fallback, asciibetize */
+if (left == -1 && right == -1)
+ return strcmp(l, r);
+
+/* left is a fallback */
+if (left >= 0)
+ return 1;
+
+/* right is a fallback */
+if (right >= 0)
+ return -1;
+
+/* both are fallbacks, which is worse */
+return left - right;
+}
+
+static void
+fixup_video_driver_list(const char **drivers)
+{
+const char **end;
+
+/* walk to the end of the list */
+for (end
[PATCH xserver 16/19] loader: Remove unused arguments from LoadModule
Nobody was ever calling this with a non-null argument for subdir list or
pattern list. Having done this, InitSubdirs is only ever called with a
NULL argument, so it's really just a complicated way of duplicating the
default list; we can remove that and just walk the list directly.
The minor error code was only ever used to distinguish among two cases
of LDR_BADUSAGE. Whatever.
Signed-off-by: Adam Jackson
---
hw/xfree86/common/xf86Helper.c | 6 +-
hw/xfree86/common/xf86Init.c| 6 +-
hw/xfree86/doc/ddxDesign.xml| 49 +
hw/xfree86/loader/loader.c | 4 +-
hw/xfree86/loader/loader.h | 2 +-
hw/xfree86/loader/loaderProcs.h | 4 +-
hw/xfree86/loader/loadmod.c | 149 +++-
7 files changed, 21 insertions(+), 199 deletions(-)
diff --git a/hw/xfree86/common/xf86Helper.c b/hw/xfree86/common/xf86Helper.c
index b464864..7d6a374 100644
--- a/hw/xfree86/common/xf86Helper.c
+++ b/hw/xfree86/common/xf86Helper.c
@@ -1590,7 +1590,7 @@ xf86LoadSubModule(ScrnInfoPtr pScrn, const char *name)
void *
xf86LoadOneModule(const char *name, void *opt)
{
-int errmaj, errmin;
+int errmaj;
char *Name;
void *mod;
@@ -1608,9 +1608,9 @@ xf86LoadOneModule(const char *name, void *opt)
return NULL;
}
-mod = LoadModule(Name, NULL, NULL, opt, NULL, &errmaj, &errmin);
+mod = LoadModule(Name, opt, NULL, &errmaj);
if (!mod)
-LoaderErrorMsg(NULL, Name, errmaj, errmin);
+LoaderErrorMsg(NULL, Name, errmaj, 0);
free(Name);
return mod;
}
diff --git a/hw/xfree86/common/xf86Init.c b/hw/xfree86/common/xf86Init.c
index f6f77c0..e61fe66 100644
--- a/hw/xfree86/common/xf86Init.c
+++ b/hw/xfree86/common/xf86Init.c
@@ -1473,7 +1473,7 @@ ddxUseMsg(void)
Bool
xf86LoadModules(const char **list, void **optlist)
{
-int errmaj, errmin;
+int errmaj;
void *opt;
int i;
char *name;
@@ -1503,8 +1503,8 @@ xf86LoadModules(const char **list, void **optlist)
else
opt = NULL;
-if (!LoadModule(name, NULL, NULL, opt, NULL, &errmaj, &errmin)) {
-LoaderErrorMsg(NULL, name, errmaj, errmin);
+if (!LoadModule(name, opt, NULL, &errmaj)) {
+LoaderErrorMsg(NULL, name, errmaj, 0);
failed = TRUE;
}
free(name);
diff --git a/hw/xfree86/doc/ddxDesign.xml b/hw/xfree86/doc/ddxDesign.xml
index 05ee042..89a85a4 100644
--- a/hw/xfree86/doc/ddxDesign.xml
+++ b/hw/xfree86/doc/ddxDesign.xml
@@ -5210,10 +5210,8 @@ XFree86 common layer.
-pointer LoadModule(const char *module,
- const char **subdirlist, const char **patternlist,
- pointer options, const XF86ModReqInfo * modreq,
- int *errmaj, int *errmin);
+pointer LoadModule(const char *module, pointer options,
+ const XF86ModReqInfo * modreq, int *errmaj);
The LoadModule() function loads the module called
@@ -5227,42 +5225,6 @@ XFree86 common layer.
-
-
- subdirlist
-
- An optional NULL terminated list of
- subdirectories to search. When NULL,
- the default built-in list is used (refer to
- stdSubdirs in
loadmod.c).
- The default list is also substituted for entries in
- subdirlist with the value
- DEFAULT_LIST. This makes is possible
- to augment the default list instead of replacing it.
- Subdir elements must be relative, and must not contain
- "..". If any violate this
requirement,
- the load fails.
-
-
-
-
- patternlist
-
- An optional NULL terminated list of
- POSIX regular expressions used to connect module
- filenames with canonical module names. Each regex
- should contain exactly one subexpression that corresponds
- to the canonical module name. When NULL,
- the default built-in list is used (refer to
- stdPatterns in
- loadmod.c). The default list is also
- substituted for entries in patternlist
- with the value DEFAULT_LIST. This
- makes it possible to augment the default list instead
- of replacing it.
-
-
-
options
@@ -5371,13 +5333,6 @@ typedef struct {
LoadModule() fails.
-
- errmin
-
- Like errmaj, but for the minor part
- of the error code.
-
-
diff --git a/hw/xfree86/load
[PATCH xserver 11/19] xfree86: Remove DriverRec1 compat struct
The idea here is that the driver might have once been old enough to not
have the driverFunc slot in DriverRec, with the module ABI not having
changed when it was added. That was ages ago, and drivers always declare
themselves with DriverRec not DriverRec1, so uninitialized slots will
simply be zero.
Signed-off-by: Adam Jackson
---
hw/xfree86/common/xf86Helper.c | 9 +
hw/xfree86/common/xf86str.h| 10 --
2 files changed, 1 insertion(+), 18 deletions(-)
diff --git a/hw/xfree86/common/xf86Helper.c b/hw/xfree86/common/xf86Helper.c
index f48af75..b464864 100644
--- a/hw/xfree86/common/xf86Helper.c
+++ b/hw/xfree86/common/xf86Helper.c
@@ -79,14 +79,7 @@ xf86AddDriver(DriverPtr driver, void *module, int flags)
xf86DriverList = xnfreallocarray(xf86DriverList,
xf86NumDrivers, sizeof(DriverPtr));
xf86DriverList[xf86NumDrivers - 1] = xnfalloc(sizeof(DriverRec));
-if (flags & HaveDriverFuncs)
-*xf86DriverList[xf86NumDrivers - 1] = *driver;
-else {
-(void) memset(xf86DriverList[xf86NumDrivers - 1], 0,
sizeof(DriverRec));
-(void) memcpy(xf86DriverList[xf86NumDrivers - 1], driver,
- sizeof(DriverRec1));
-
-}
+*xf86DriverList[xf86NumDrivers - 1] = *driver;
xf86DriverList[xf86NumDrivers - 1]->module = module;
xf86DriverList[xf86NumDrivers - 1]->refCount = 0;
}
diff --git a/hw/xfree86/common/xf86str.h b/hw/xfree86/common/xf86str.h
index bfcb75e..74c65ba 100644
--- a/hw/xfree86/common/xf86str.h
+++ b/hw/xfree86/common/xf86str.h
@@ -209,16 +209,6 @@ typedef CARD32 xorgHWFlags;
*/
struct _DriverRec;
-typedef struct {
-int driverVersion;
-const char *driverName;
-void (*Identify) (int flags);
-Bool (*Probe) (struct _DriverRec * drv, int flags);
-const OptionInfoRec *(*AvailableOptions) (int chipid, int bustype);
-void *module;
-int refCount;
-} DriverRec1;
-
struct _SymTabRec;
struct _PciChipsets;
--
2.9.3
___
xorg-devel@lists.x.org: X.Org development
Archives: http://lists.x.org/archives/xorg-devel
Info: https://lists.x.org/mailman/listinfo/xorg-devel
[PATCH xserver 12/19] xfree86: Fix up some bad indentation
indent(1) gets confused by function-like macros with no trailing
semicolon, which is fair enough really.
Signed-off-by: Adam Jackson
---
hw/xfree86/common/xf86Configure.c | 28 ++--
1 file changed, 14 insertions(+), 14 deletions(-)
diff --git a/hw/xfree86/common/xf86Configure.c
b/hw/xfree86/common/xf86Configure.c
index 59d275e..8c8e49e 100644
--- a/hw/xfree86/common/xf86Configure.c
+++ b/hw/xfree86/common/xf86Configure.c
@@ -157,7 +157,7 @@ configureInputSection(void)
{
XF86ConfInputPtr mouse = NULL;
-parsePrologue(XF86ConfInputPtr, XF86ConfInputRec)
+parsePrologue(XF86ConfInputPtr, XF86ConfInputRec);
ptr->inp_identifier = xnfstrdup("Keyboard0");
ptr->inp_driver = xnfstrdup("kbd");
@@ -196,7 +196,7 @@ configureScreenSection(int screennum)
int i;
int depths[] = { 1, 4, 8, 15, 16, 24 /*, 32 */ };
char *tmp;
-parsePrologue(XF86ConfScreenPtr, XF86ConfScreenRec)
+parsePrologue(XF86ConfScreenPtr, XF86ConfScreenRec);
XNFasprintf(&tmp, "Screen%d", screennum);
ptr->scrn_identifier = tmp;
@@ -254,9 +254,9 @@ configureDeviceSection(int screennum)
int i = 0;
char *identifier;
-parsePrologue(XF86ConfDevicePtr, XF86ConfDeviceRec)
+parsePrologue(XF86ConfDevicePtr, XF86ConfDeviceRec);
-/* Move device info to parser structure */
+/* Move device info to parser structure */
if (asprintf(&identifier, "Card%d", screennum) == -1)
identifier = NULL;
ptr->dev_identifier = identifier;
@@ -326,9 +326,9 @@ configureLayoutSection(void)
{
int scrnum = 0;
-parsePrologue(XF86ConfLayoutPtr, XF86ConfLayoutRec)
+parsePrologue(XF86ConfLayoutPtr, XF86ConfLayoutRec);
-ptr->lay_identifier = "X.org Configured";
+ptr->lay_identifier = "X.org Configured";
{
XF86ConfInputrefPtr iptr;
@@ -389,9 +389,9 @@ configureLayoutSection(void)
static XF86ConfFlagsPtr
configureFlagsSection(void)
{
-parsePrologue(XF86ConfFlagsPtr, XF86ConfFlagsRec)
+parsePrologue(XF86ConfFlagsPtr, XF86ConfFlagsRec);
-return ptr;
+return ptr;
}
static XF86ConfModulePtr
@@ -405,9 +405,9 @@ configureModuleSection(void)
"fonts",
NULL
};
-parsePrologue(XF86ConfModulePtr, XF86ConfModuleRec)
+parsePrologue(XF86ConfModulePtr, XF86ConfModuleRec);
-elist = LoaderListDirs(esubdirs, NULL);
+elist = LoaderListDirs(esubdirs, NULL);
if (elist) {
for (el = elist; *el; el++) {
XF86LoadPtr module;
@@ -427,9 +427,9 @@ configureModuleSection(void)
static XF86ConfFilesPtr
configureFilesSection(void)
{
-parsePrologue(XF86ConfFilesPtr, XF86ConfFilesRec)
+parsePrologue(XF86ConfFilesPtr, XF86ConfFilesRec);
-if (xf86ModulePath)
+if (xf86ModulePath)
ptr->file_modulepath = xnfstrdup(xf86ModulePath);
if (defaultFontPath)
ptr->file_fontpath = xnfstrdup(defaultFontPath);
@@ -441,7 +441,7 @@ static XF86ConfMonitorPtr
configureMonitorSection(int screennum)
{
char *tmp;
-parsePrologue(XF86ConfMonitorPtr, XF86ConfMonitorRec)
+parsePrologue(XF86ConfMonitorPtr, XF86ConfMonitorRec);
XNFasprintf(&tmp, "Monitor%d", screennum);
ptr->mon_identifier = tmp;
@@ -486,7 +486,7 @@ configureDDCMonitorSection(int screennum)
int displaySizeLen;
char *tmp;
-parsePrologue(XF86ConfMonitorPtr, XF86ConfMonitorRec)
+parsePrologue(XF86ConfMonitorPtr, XF86ConfMonitorRec);
XNFasprintf(&tmp, "Monitor%d", screennum);
ptr->mon_identifier = tmp;
--
2.9.3
___
xorg-devel@lists.x.org: X.Org development
Archives: http://lists.x.org/archives/xorg-devel
Info: https://lists.x.org/mailman/listinfo/xorg-devel
[PATCH xserver 04/19] xfree86: flatten pathlist management in the loader
From: Emil Velikov
Now that used can set the path only via LoaderSetPath(), we can simplify
things.
Signed-off-by: Emil Velikov
---
hw/xfree86/loader/loadmod.c | 16 ++--
1 file changed, 2 insertions(+), 14 deletions(-)
diff --git a/hw/xfree86/loader/loadmod.c b/hw/xfree86/loader/loadmod.c
index 5e4d7da..fd61a82 100644
--- a/hw/xfree86/loader/loadmod.c
+++ b/hw/xfree86/loader/loadmod.c
@@ -127,9 +127,6 @@ InitPathList(const char *path)
int addslash;
int n = 0;
-if (!path)
-return defaultPathList;
-
fullpath = strdup(path);
if (!fullpath)
return NULL;
@@ -171,13 +168,6 @@ InitPathList(const char *path)
return list;
}
-static void
-FreePathList(char **pathlist)
-{
-if (pathlist && pathlist != defaultPathList)
-FreeStringList(pathlist);
-}
-
void
LoaderSetPath(const char *path)
{
@@ -498,7 +488,7 @@ LoaderListDirs(const char **subdirlist, const char
**patternlist)
char **ret = NULL;
int n = 0;
-if (!(pathlist = InitPathList(NULL)))
+if (!(pathlist = defaultPathList))
return NULL;
if (!(subdirs = InitSubdirs(subdirlist)))
goto bail;
@@ -565,7 +555,6 @@ LoaderListDirs(const char **subdirlist, const char
**patternlist)
bail:
FreePatterns(patterns);
FreeSubdirs(subdirs);
-FreePathList(pathlist);
return (const char **) ret;
}
@@ -908,7 +897,7 @@ LoadModule(const char *module, const char **subdirlist,
goto LoadModule_fail;
}
-pathlist = InitPathList(NULL);
+pathlist = defaultPathList;
if (!pathlist) {
/* This could be a malloc failure too */
if (errmaj)
@@ -1033,7 +1022,6 @@ LoadModule(const char *module, const char **subdirlist,
ret = NULL;
LoadModule_exit:
-FreePathList(pathlist);
FreePatterns(patterns);
free(found);
free(name);
--
2.9.3
___
xorg-devel@lists.x.org: X.Org development
Archives: http://lists.x.org/archives/xorg-devel
Info: https://lists.x.org/mailman/listinfo/xorg-devel
[PATCH xserver 09/19] loader: Remove unused loader error codes and dead enum
The enum has been unused since at least the removal of elfloader.
Signed-off-by: Adam Jackson
---
hw/xfree86/common/xf86Module.h | 15 +--
hw/xfree86/loader/loadmod.c| 18 --
2 files changed, 1 insertion(+), 32 deletions(-)
diff --git a/hw/xfree86/common/xf86Module.h b/hw/xfree86/common/xf86Module.h
index 7d09a1b..ff0e23e 100644
--- a/hw/xfree86/common/xf86Module.h
+++ b/hw/xfree86/common/xf86Module.h
@@ -49,13 +49,6 @@
#define NULL ((void *)0)
#endif
-typedef enum {
-LD_RESOLV_IFDONE = 0, /* only check if no more
- delays pending */
-LD_RESOLV_NOW = 1, /* finish one delay step */
-LD_RESOLV_FORCE = 2 /* force checking... */
-} LoaderResolveOptions;
-
#define DEFAULT_LIST ((char *)-1)
/* Built-in ABI classes. These definitions must not be changed. */
@@ -92,19 +85,13 @@ typedef enum {
#define MODULEVENDORSTRING "X.Org Foundation"
#endif
-/* Error return codes for errmaj. New codes must only be added at the end. */
+/* Error return codes for errmaj */
typedef enum {
LDR_NOERROR = 0,
LDR_NOMEM, /* memory allocation failed */
LDR_NOENT, /* Module file does not exist */
-LDR_NOSUBENT, /* pre-requsite file to be sub-loaded does not
exist */
-LDR_NOSPACE,/* internal module array full */
-LDR_NOMODOPEN, /* module file could not be opened (check
errmin) */
-LDR_UNKTYPE,/* file is not a recognized module type */
LDR_NOLOAD, /* type specific loader failed */
LDR_ONCEONLY, /* Module should only be loaded once (not an
error) */
-LDR_NOPORTOPEN, /* could not open port (check errmin) */
-LDR_NOHARDWARE, /* could not query/initialize the hardware
device */
LDR_MISMATCH, /* the module didn't match the spec'd
requirments */
LDR_BADUSAGE, /* LoadModule is called with bad arguments */
LDR_INVALID,/* The module doesn't have a valid ModuleData
object */
diff --git a/hw/xfree86/loader/loadmod.c b/hw/xfree86/loader/loadmod.c
index d326d9d..85689be 100644
--- a/hw/xfree86/loader/loadmod.c
+++ b/hw/xfree86/loader/loadmod.c
@@ -1094,18 +1094,6 @@ LoaderErrorMsg(const char *name, const char *modname,
int errmaj, int errmin)
case LDR_NOENT:
msg = "module does not exist";
break;
-case LDR_NOSUBENT:
-msg = "a required submodule could not be loaded";
-break;
-case LDR_NOSPACE:
-msg = "too many modules";
-break;
-case LDR_NOMODOPEN:
-msg = "open failed";
-break;
-case LDR_UNKTYPE:
-msg = "unknown module type";
-break;
case LDR_NOLOAD:
msg = "loader failed";
break;
@@ -1113,12 +1101,6 @@ LoaderErrorMsg(const char *name, const char *modname,
int errmaj, int errmin)
msg = "already loaded";
type = X_INFO;
break;
-case LDR_NOPORTOPEN:
-msg = "port open failed";
-break;
-case LDR_NOHARDWARE:
-msg = "no hardware found";
-break;
case LDR_MISMATCH:
msg = "module requirement mismatch";
break;
--
2.9.3
___
xorg-devel@lists.x.org: X.Org development
Archives: http://lists.x.org/archives/xorg-devel
Info: https://lists.x.org/mailman/listinfo/xorg-devel
[PATCH xserver 06/19] loader: Remove *GetOS
This API is dumb. uname(3) exists, feel free to use it, but ideally
write to the interface not to the OS. There are a couple of drivers
using this API, they could all reasonably just not.
This also removes the OS name from the loader subdirectory path search.
Having /usr/lib/xorg shared across OSes is a non-goal here.
Signed-off-by: Adam Jackson
---
hw/xfree86/common/xf86.h | 2 --
hw/xfree86/common/xf86Helper.c | 4 ---
hw/xfree86/common/xf86Module.h | 2 --
hw/xfree86/loader/Makefile.am | 3 +-
hw/xfree86/loader/loadmod.c| 14 +---
hw/xfree86/loader/os.c | 72 --
6 files changed, 2 insertions(+), 95 deletions(-)
delete mode 100644 hw/xfree86/loader/os.c
diff --git a/hw/xfree86/common/xf86.h b/hw/xfree86/common/xf86.h
index f724688..828bff1 100644
--- a/hw/xfree86/common/xf86.h
+++ b/hw/xfree86/common/xf86.h
@@ -343,8 +343,6 @@ xf86SetSilkenMouse(ScreenPtr pScreen);
extern _X_EXPORT void *
xf86FindXvOptions(ScrnInfoPtr pScrn, int adapt_index, const char *port_name,
const char **adaptor_name, void **adaptor_options);
-extern _X_EXPORT void
-xf86GetOS(const char **name, int *major, int *minor, int *teeny);
extern _X_EXPORT ScrnInfoPtr
xf86ConfigFbEntity(ScrnInfoPtr pScrn, int scrnFlag,
int entityIndex, EntityProc init,
diff --git a/hw/xfree86/common/xf86Helper.c b/hw/xfree86/common/xf86Helper.c
index 9388436..f48af75 100644
--- a/hw/xfree86/common/xf86Helper.c
+++ b/hw/xfree86/common/xf86Helper.c
@@ -1752,10 +1752,6 @@ xf86FindXvOptions(ScrnInfoPtr pScrn, int adaptor_index,
const char *port_name,
return NULL;
}
-/* Rather than duplicate loader's get OS function, just include it directly */
-#define LoaderGetOS xf86GetOS
-#include "loader/os.c"
-
static void
xf86ConfigFbEntityInactive(EntityInfoPtr pEnt, EntityProc init,
EntityProc enter, EntityProc leave, void *private)
diff --git a/hw/xfree86/common/xf86Module.h b/hw/xfree86/common/xf86Module.h
index e0212cf..cd4227a 100644
--- a/hw/xfree86/common/xf86Module.h
+++ b/hw/xfree86/common/xf86Module.h
@@ -176,8 +176,6 @@ extern _X_EXPORT void *LoaderSymbol(const char *);
extern _X_EXPORT const char **LoaderListDirs(const char **, const char **);
extern _X_EXPORT void LoaderFreeDirList(char **);
extern _X_EXPORT void LoaderErrorMsg(const char *, const char *, int, int);
-extern _X_EXPORT void LoaderGetOS(const char **name, int *major, int *minor,
- int *teeny);
extern _X_EXPORT Bool LoaderShouldIgnoreABI(void);
extern _X_EXPORT int LoaderGetABIVersion(const char *abiclass);
diff --git a/hw/xfree86/loader/Makefile.am b/hw/xfree86/loader/Makefile.am
index 9218cab..3529a7a 100644
--- a/hw/xfree86/loader/Makefile.am
+++ b/hw/xfree86/loader/Makefile.am
@@ -14,7 +14,6 @@ EXTRA_DIST = \
libloader_la_SOURCES = \
loader.c \
loaderProcs.h \
-loadmod.c \
- os.c
+loadmod.c
libloader_la_LIBADD = $(DLOPEN_LIBS)
diff --git a/hw/xfree86/loader/loadmod.c b/hw/xfree86/loader/loadmod.c
index 03c9966..95a37fc 100644
--- a/hw/xfree86/loader/loadmod.c
+++ b/hw/xfree86/loader/loadmod.c
@@ -269,9 +269,7 @@ InitSubdirs(const char **subdirlist)
const char **tmp_subdirlist = NULL;
char **subdirs = NULL;
const char **s, **stmp = NULL;
-const char *osname;
-const char *slash;
-int oslen = 0, len;
+int len;
Bool indefault;
if (subdirlist == NULL) {
@@ -282,9 +280,6 @@ InitSubdirs(const char **subdirlist)
subdirlist[1] = NULL;
}
-LoaderGetOS(&osname, NULL, NULL, NULL);
-oslen = strlen(osname);
-
{
/* Count number of entries and check for invalid paths */
for (i = 0, s = subdirlist; *s; i++, s++) {
@@ -323,12 +318,8 @@ InitSubdirs(const char **subdirlist)
}
len = strlen(*s);
if (**s && (*s)[len - 1] != '/') {
-slash = "/";
len++;
}
-else
-slash = "";
-len += oslen + 2;
if (!(subdirs[i] = malloc(len))) {
while (--i >= 0)
free(subdirs[i]);
@@ -336,9 +327,6 @@ InitSubdirs(const char **subdirlist)
free(tmp_subdirlist);
return NULL;
}
-/* tack on the OS name */
-sprintf(subdirs[i], "%s%s%s/", *s, slash, osname);
-i++;
/* path as given */
subdirs[i] = strdup(*s);
i++;
diff --git a/hw/xfree86/loader/os.c b/hw/xfree86/loader/os.c
deleted file mode 100644
index 8d03721..000
--- a/hw/xfree86/loader/os.c
+++ /dev/null
@@ -1,72 +0,0 @@
-/*
- * Copyright (c) 1999-2002 by The XFree86 Project, Inc.
- *
- * Permission is hereby granted, free of charge, to any person obtaining a
- * copy of this software and associated documentation files (the "Software"),
- * to deal in the Software without
[PATCH xserver 05/19] loader: Port from xfree86 to dix API
Signed-off-by: Adam Jackson
---
hw/xfree86/loader/loader.c | 43 +-
hw/xfree86/loader/loadmod.c | 141
2 files changed, 81 insertions(+), 103 deletions(-)
diff --git a/hw/xfree86/loader/loader.c b/hw/xfree86/loader/loader.c
index c8d7540..fded87b 100644
--- a/hw/xfree86/loader/loader.c
+++ b/hw/xfree86/loader/loader.c
@@ -50,21 +50,10 @@
#include
#endif
-#include
-#include
-#include
-#include
-#include
-#include
-#include
#include
-#include
-
#include "os.h"
#include "loader.h"
#include "loaderProcs.h"
-#include "xf86.h"
-#include "xf86Priv.h"
#ifdef HAVE_DLFCN_H
@@ -80,20 +69,20 @@ extern void *xorg_symbols[];
void
LoaderInit(void)
{
-xf86MsgVerb(X_INFO, 2, "Loader magic: %p\n", (void *) xorg_symbols);
-xf86MsgVerb(X_INFO, 2, "Module ABI versions:\n");
-xf86ErrorFVerb(2, "\t%s: %d.%d\n", ABI_CLASS_ANSIC,
- GET_ABI_MAJOR(LoaderVersionInfo.ansicVersion),
- GET_ABI_MINOR(LoaderVersionInfo.ansicVersion));
-xf86ErrorFVerb(2, "\t%s: %d.%d\n", ABI_CLASS_VIDEODRV,
- GET_ABI_MAJOR(LoaderVersionInfo.videodrvVersion),
- GET_ABI_MINOR(LoaderVersionInfo.videodrvVersion));
-xf86ErrorFVerb(2, "\t%s : %d.%d\n", ABI_CLASS_XINPUT,
- GET_ABI_MAJOR(LoaderVersionInfo.xinputVersion),
- GET_ABI_MINOR(LoaderVersionInfo.xinputVersion));
-xf86ErrorFVerb(2, "\t%s : %d.%d\n", ABI_CLASS_EXTENSION,
- GET_ABI_MAJOR(LoaderVersionInfo.extensionVersion),
- GET_ABI_MINOR(LoaderVersionInfo.extensionVersion));
+LogMessageVerb(X_INFO, 2, "Loader magic: %p\n", (void *) xorg_symbols);
+LogMessageVerb(X_INFO, 2, "Module ABI versions:\n");
+LogWrite(2, "\t%s: %d.%d\n", ABI_CLASS_ANSIC,
+ GET_ABI_MAJOR(LoaderVersionInfo.ansicVersion),
+ GET_ABI_MINOR(LoaderVersionInfo.ansicVersion));
+LogWrite(2, "\t%s: %d.%d\n", ABI_CLASS_VIDEODRV,
+ GET_ABI_MAJOR(LoaderVersionInfo.videodrvVersion),
+ GET_ABI_MINOR(LoaderVersionInfo.videodrvVersion));
+LogWrite(2, "\t%s : %d.%d\n", ABI_CLASS_XINPUT,
+ GET_ABI_MAJOR(LoaderVersionInfo.xinputVersion),
+ GET_ABI_MINOR(LoaderVersionInfo.xinputVersion));
+LogWrite(2, "\t%s : %d.%d\n", ABI_CLASS_EXTENSION,
+ GET_ABI_MAJOR(LoaderVersionInfo.extensionVersion),
+ GET_ABI_MINOR(LoaderVersionInfo.extensionVersion));
}
@@ -108,10 +97,10 @@ LoaderOpen(const char *module, int *errmaj, int *errmin)
ErrorF("LoaderOpen(%s)\n", module);
#endif
-xf86Msg(X_INFO, "Loading %s\n", module);
+LogMessage(X_INFO, "Loading %s\n", module);
if (!(ret = dlopen(module, RTLD_LAZY | RTLD_GLOBAL))) {
-xf86Msg(X_ERROR, "Failed to load %s: %s\n", module, dlerror());
+LogMessage(X_ERROR, "Failed to load %s: %s\n", module, dlerror());
if (errmaj)
*errmaj = LDR_NOLOAD;
if (errmin)
diff --git a/hw/xfree86/loader/loadmod.c b/hw/xfree86/loader/loadmod.c
index fd61a82..03c9966 100644
--- a/hw/xfree86/loader/loadmod.c
+++ b/hw/xfree86/loader/loadmod.c
@@ -51,18 +51,11 @@
#endif
#include "os.h"
-/* For stat() and related stuff */
-#define NO_OSLIB_PROTOTYPES
-#include "xf86_OSlib.h"
-#define LOADERDECLARATIONS
#include "loaderProcs.h"
-#include "misc.h"
-#include "xf86.h"
-#include "xf86Priv.h"
-#include "xf86Xinput.h"
+#include "xf86Module.h"
#include "loader.h"
-#include "xf86Optrec.h"
+#include
#include
#include
#include
@@ -307,7 +300,7 @@ InitSubdirs(const char **subdirlist)
*/
if (**s == '/' || **s == '\\' || strchr(*s, ':') ||
strstr(*s, "..")) {
-xf86Msg(X_ERROR, "InitSubdirs: Bad subdir: \"%s\"\n", *s);
+LogMessage(X_ERROR, "InitSubdirs: Bad subdir \"%s\"\n",
*s);
free(tmp_subdirlist);
return NULL;
}
@@ -572,22 +565,22 @@ CheckVersion(const char *module, XF86ModuleVersionInfo *
data,
long ver = data->xf86version;
MessageType errtype;
-xf86Msg(X_INFO, "Module %s: vendor=\"%s\"\n",
-data->modname ? data->modname : "UNKNOWN!",
-data->vendor ? data->vendor : "UNKNOWN!");
+LogMessage(X_INFO, "Module %s: vendor=\"%s\"\n",
+ data->modname ? data->modname : "UNKNOWN!",
+ data->vendor ? data->vendor : "UNKNOWN!");
vercode[0] = ver / 1000;
vercode[1] = (ver / 10) % 100;
vercode[2] = (ver / 1000) % 100;
vercode[3] = ver % 1000;
-xf86ErrorF("\tcompiled for %d.%d.%d", vercode[0], vercode[1], vercode[2]);
+LogWrite(1, "\tcompiled for %d.%d.%d", vercode[0], vercode[1], vercode[2]);
if (vercode[3] != 0)
-xf86ErrorF(".%d", vercode[3]);
-xf86ErrorF(", module version = %d.%d.%d\n", data->majorversion,
- data->minorversi
[PATCH xserver 10/19] loader: Remove silly "unspecified" version handling
Nobody who is using this functionality is ever not specifying a major
version, which makes sense. If you don't care about a minor version,
that's equivalent to saying you require minor >= 0, so just say so;
likewise patch level.
Likewise nobody using this functionality is ever not specifying an ABI
class.
Signed-off-by: Adam Jackson
---
hw/xfree86/common/xf86Module.h | 6 --
hw/xfree86/doc/ddxDesign.xml | 20 +-
hw/xfree86/loader/loadmod.c| 47 +-
3 files changed, 29 insertions(+), 44 deletions(-)
diff --git a/hw/xfree86/common/xf86Module.h b/hw/xfree86/common/xf86Module.h
index ff0e23e..faea07f 100644
--- a/hw/xfree86/common/xf86Module.h
+++ b/hw/xfree86/common/xf86Module.h
@@ -141,12 +141,6 @@ typedef struct {
const char *moduleclass;/* module class */
} XF86ModReqInfo;
-/* values to indicate unspecified fields in XF86ModReqInfo. */
-#define MAJOR_UNSPEC 0xFF
-#define MINOR_UNSPEC 0xFF
-#define PATCH_UNSPEC 0x
-#define ABI_VERS_UNSPEC0x
-
#define MODULE_VERSION_NUMERIC(maj, min, patch) \
maj) & 0xFF) << 24) | (((min) & 0xFF) << 16) | (patch & 0x))
#define GET_MODULE_MAJOR_VERSION(vers) (((vers) >> 24) & 0xFF)
diff --git a/hw/xfree86/doc/ddxDesign.xml b/hw/xfree86/doc/ddxDesign.xml
index 2ff2894..05ee042 100644
--- a/hw/xfree86/doc/ddxDesign.xml
+++ b/hw/xfree86/doc/ddxDesign.xml
@@ -5293,12 +5293,12 @@ XFree86 common layer.
as follows:
typedef struct {
- CARD8majorversion; /* MAJOR_UNSPEC */
- CARD8minorversion; /* MINOR_UNSPEC */
- CARD16 patchlevel;/* PATCH_UNSPEC */
- const char * abiclass; /* ABI_CLASS_NONE */
- CARD32 abiversion;/* ABI_VERS_UNSPEC */
- const char * moduleclass; /* MOD_CLASS_NONE */
+ CARD8majorversion;
+ CARD8minorversion;
+ CARD16 patchlevel;
+ const char * abiclass;
+ CARD32 abiversion;
+ const char * moduleclass;
} XF86ModReqInfo;
@@ -5323,8 +5323,8 @@ typedef struct {
The module's minor version must be
no less than this value. This
comparison is only made if
- majorversion is
- specified and matches.
+ majorversion
+ matches.
@@ -5333,8 +5333,8 @@ typedef struct {
The module's patchlevel must be no
less than this value. This comparison
is only made if
- minorversion is
- specified and matches.
+ minorversion
+ matches.
diff --git a/hw/xfree86/loader/loadmod.c b/hw/xfree86/loader/loadmod.c
index 85689be..528cc88 100644
--- a/hw/xfree86/loader/loadmod.c
+++ b/hw/xfree86/loader/loadmod.c
@@ -617,32 +617,24 @@ CheckVersion(const char *module, XF86ModuleVersionInfo *
data,
/* Check against requirements that the caller has specified */
if (req) {
-if (req->majorversion != MAJOR_UNSPEC) {
-if (data->majorversion != req->majorversion) {
-LogMessageVerb(X_WARNING, 2, "%s: module major version (%d) "
- "doesn't match required major version (%d)\n",
- module, data->majorversion, req->majorversion);
-return FALSE;
-}
-else if (req->minorversion != MINOR_UNSPEC) {
-if (data->minorversion < req->minorversion) {
-LogMessageVerb(X_WARNING, 2, "%s: module minor version "
- "(%d) is less than the required minor "
- "version (%d)\n", module,
- data->minorversion, req->minorversion);
-return FALSE;
-}
-else if (data->minorversion == req->minorversion &&
- req->patchlevel != PATCH_UNSPEC) {
-if (data->patchlevel < req->patchlevel) {
-LogMessageVerb(X_WARNING, 2, "%sL module patch level "
- "(%d) is less than the required patch "
- "level (%d)\n", module,
data->patchlevel,
- req->patchlevel);
-return FALSE;
-}
-}
-}
+if (data->majorversion != req->majorversion) {
+
[PATCH xserver 07/19] loader: Don't add internal/ to the search path
Signed-off-by: Adam Jackson
---
hw/xfree86/loader/loadmod.c | 1 -
1 file changed, 1 deletion(-)
diff --git a/hw/xfree86/loader/loadmod.c b/hw/xfree86/loader/loadmod.c
index 95a37fc..d326d9d 100644
--- a/hw/xfree86/loader/loadmod.c
+++ b/hw/xfree86/loader/loadmod.c
@@ -177,7 +177,6 @@ static const char *stdSubdirs[] = {
"input/",
"drivers/",
"extensions/",
-"internal/",
NULL
};
--
2.9.3
___
xorg-devel@lists.x.org: X.Org development
Archives: http://lists.x.org/archives/xorg-devel
Info: https://lists.x.org/mailman/listinfo/xorg-devel
[PATCH xserver 08/19] loader: Include fewer headers from xf86Module.h
This looks like more, but only if you don't compare it to the number pulled in by misc.h. Signed-off-by: Adam Jackson --- hw/xfree86/common/xf86Module.h | 6 -- 1 file changed, 4 insertions(+), 2 deletions(-) diff --git a/hw/xfree86/common/xf86Module.h b/hw/xfree86/common/xf86Module.h index cd4227a..7d09a1b 100644 --- a/hw/xfree86/common/xf86Module.h +++ b/hw/xfree86/common/xf86Module.h @@ -41,8 +41,10 @@ #ifndef _XF86MODULE_H #define _XF86MODULE_H -#include "misc.h" -#include "extension.h" +#include +#include +#include + #ifndef NULL #define NULL ((void *)0) #endif -- 2.9.3 ___ xorg-devel@lists.x.org: X.Org development Archives: http://lists.x.org/archives/xorg-devel Info: https://lists.x.org/mailman/listinfo/xorg-devel
[PATCH xserver 01/19] xfree86: remove references from LoadSubModule's path from the doc
From: Emil Velikov Afaics the argument hasn't been part of the API since the documentation has been converted to xml with commit fc6ebe1e1d3 "Convert LinuxDoc documents to DocBook/XML" Signed-off-by: Emil Velikov --- hw/xfree86/doc/ddxDesign.xml | 6 +- 1 file changed, 1 insertion(+), 5 deletions(-) diff --git a/hw/xfree86/doc/ddxDesign.xml b/hw/xfree86/doc/ddxDesign.xml index f7d6628..b62272d 100644 --- a/hw/xfree86/doc/ddxDesign.xml +++ b/hw/xfree86/doc/ddxDesign.xml @@ -5695,11 +5695,7 @@ the server, and may also be used from within modules. described above, except that the module loaded is registered as a child of the calling module. The parent parameter is the calling module's handle. Modules loaded with this function -are automatically unloaded when the parent module is unloaded. The -other difference is that the path parameter may not be specified. -The module search path used for modules loaded with this function -is the default search path as initialised with -LoaderSetPath(). +are automatically unloaded when the parent module is unloaded. -- 2.9.3 ___ xorg-devel@lists.x.org: X.Org development Archives: http://lists.x.org/archives/xorg-devel Info: https://lists.x.org/mailman/listinfo/xorg-devel
[PATCH xserver 03/19] xfree86: remove dummy/dead function prototype for LoadDriver
From: Emil Velikov
Signed-off-by: Emil Velikov
---
hw/xfree86/loader/loaderProcs.h | 2 --
1 file changed, 2 deletions(-)
diff --git a/hw/xfree86/loader/loaderProcs.h b/hw/xfree86/loader/loaderProcs.h
index 8d7872f..ee8a557 100644
--- a/hw/xfree86/loader/loaderProcs.h
+++ b/hw/xfree86/loader/loaderProcs.h
@@ -72,8 +72,6 @@ typedef struct module_desc {
void LoaderInit(void);
-ModuleDescPtr LoadDriver(const char *, const char *, int, void *, int *,
- int *);
ModuleDescPtr LoadModule(const char *, const char **,
const char **, void *, const XF86ModReqInfo *,
int *, int *);
--
2.9.3
___
xorg-devel@lists.x.org: X.Org development
Archives: http://lists.x.org/archives/xorg-devel
Info: https://lists.x.org/mailman/listinfo/xorg-devel
[PATCH xserver] tests: shuffle around the linking order to please GNU ld
Classic GNU ld resolves symbol dependencies only forward, while GOLD seems to work regardless of the specified library order. I assume that the original author of the changes (just like me), has been using GNU ld with GOLD enabled (default on ArchLinux), and did not check on older distros like Ubuntu 16.04 which shows the issue. Suggested-by: Eero Tamminen Signed-off-by: Martin Peres --- test/Makefile.am | 6 +- 1 file changed, 5 insertions(+), 1 deletion(-) diff --git a/test/Makefile.am b/test/Makefile.am index 111b60788b..e7fe587bb8 100644 --- a/test/Makefile.am +++ b/test/Makefile.am @@ -64,7 +64,7 @@ tests_LDFLAGS = \ -Wl,-wrap,GrabButton \ $() -tests_LDADD = $(XORG_SYS_LIBS) $(XSERVER_SYS_LIBS) $(GLX_SYS_LIBS) +tests_LDADD = tests_SOURCES = \ tests-common.c \ @@ -225,6 +225,10 @@ endif endif !XORG +# GNU LD scans only in one direction, add the following dependencies at the end +# so as they get picked up by the previously-linked libraries +tests_LDADD += $(XORG_SYS_LIBS) $(XSERVER_SYS_LIBS) $(GLX_SYS_LIBS) + endif ENABLE_UNIT_TESTS EXTRA_DIST = \ -- 2.11.0 ___ xorg-devel@lists.x.org: X.Org development Archives: http://lists.x.org/archives/xorg-devel Info: https://lists.x.org/mailman/listinfo/xorg-devel
Re: [PATCH] libXinerama: Set number to 0 on error.
On Sun, 2017-01-22 at 15:18 +0100, Tobias Stoeckmann wrote:
> @@ -286,6 +286,7 @@ XineramaQueryScreens(
> if (!_XReply (dpy, (xReply *) &rep, 0, xFalse)) {
> UnlockDisplay (dpy);
> SyncHandle ();
> + *number = 0;
Not that any caller has likely made this mistake, but you want an if
(number) before this, otherwise you turn a protocol error into a
segfault.
- ajax
___
xorg-devel@lists.x.org: X.Org development
Archives: http://lists.x.org/archives/xorg-devel
Info: https://lists.x.org/mailman/listinfo/xorg-devel
[PATCH xserver] parser: Fix crash when xf86nameCompare(s1 = x, s2 = NULL)
Signed-off-by: Adam Jackson
---
hw/xfree86/parser/scan.c | 2 ++
1 file changed, 2 insertions(+)
diff --git a/hw/xfree86/parser/scan.c b/hw/xfree86/parser/scan.c
index 81a454b..3356224 100644
--- a/hw/xfree86/parser/scan.c
+++ b/hw/xfree86/parser/scan.c
@@ -1046,6 +1046,8 @@ xf86nameCompare(const char *s1, const char *s2)
return 0;
else
return 1;
+} else if (!s2 || *s2 == 0) {
+return -1;
}
while (*s1 == '_' || *s1 == ' ' || *s1 == '\t')
--
2.9.3
___
xorg-devel@lists.x.org: X.Org development
Archives: http://lists.x.org/archives/xorg-devel
Info: https://lists.x.org/mailman/listinfo/xorg-devel
[PATCH xserver] glamor: keep gl_fbo and fbo consistent
If the pixmap type is neither GLAMOR_TEXTURE_ONLY nor GLAMOR_TEXTURE_DRM
we might have the fbo field set but the gl_fbo still set to the default
GLAMOR_FBO_UNATTACHED, which later may fail an assert in
glamor_upload_picture_to_texture().
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=99346
Signed-off-by: Olivier Fourdan
---
glamor/glamor_render.c | 8
glamor/glamor_utils.h | 4
2 files changed, 12 insertions(+)
diff --git a/glamor/glamor_render.c b/glamor/glamor_render.c
index e04dd21..a9ab971 100644
--- a/glamor/glamor_render.c
+++ b/glamor/glamor_render.c
@@ -916,6 +916,10 @@ glamor_composite_choose_shader(CARD8 op,
}
if (source_pixmap_priv->gl_fbo == GLAMOR_FBO_UNATTACHED) {
#ifdef GLAMOR_PIXMAP_DYNAMIC_UPLOAD
+if (!GLAMOR_PIXMAP_PRIV_HAS_TEXTURE(source_pixmap_priv)) {
+glamor_fallback("no texture in source\n");
+goto fail;
+}
source_needs_upload = TRUE;
#else
glamor_fallback("no texture in source\n");
@@ -932,6 +936,10 @@ glamor_composite_choose_shader(CARD8 op,
}
if (mask_pixmap_priv->gl_fbo == GLAMOR_FBO_UNATTACHED) {
#ifdef GLAMOR_PIXMAP_DYNAMIC_UPLOAD
+if (!GLAMOR_PIXMAP_PRIV_HAS_TEXTURE(mask_pixmap_priv)) {
+glamor_fallback("no texture in mask\n");
+goto fail;
+}
mask_needs_upload = TRUE;
#else
glamor_fallback("no texture in mask\n");
diff --git a/glamor/glamor_utils.h b/glamor/glamor_utils.h
index 6b88527..636c095 100644
--- a/glamor/glamor_utils.h
+++ b/glamor/glamor_utils.h
@@ -585,6 +585,10 @@
#define GLAMOR_PIXMAP_PRIV_HAS_FBO(pixmap_priv)(pixmap_priv->gl_fbo ==
GLAMOR_FBO_NORMAL)
+#define GLAMOR_PIXMAP_PRIV_HAS_TEXTURE(pixmap_priv) ( \
+ pixmap_priv->type == GLAMOR_TEXTURE_DRM \
+ || pixmap_priv->type == GLAMOR_TEXTURE_ONLY)
+
/**
* Borrow from uxa.
*/
--
2.9.3
___
xorg-devel@lists.x.org: X.Org development
Archives: http://lists.x.org/archives/xorg-devel
Info: https://lists.x.org/mailman/listinfo/xorg-devel
[PATCH util-modular] WIP: HACK: invoke configure from a subdir
---
Educated guess since I cannot reproduce it. If you're seing any of the
install/changelog issues mentioned in my email, just drop the relevant
lines from Makefile.am
release.sh | 4 +++-
1 file changed, 3 insertions(+), 1 deletion(-)
diff --git a/release.sh b/release.sh
index e0209bc..74b9bae 100755
--- a/release.sh
+++ b/release.sh
@@ -391,7 +391,9 @@ process_module() {
return 1
fi
-./configure >/dev/null
+mkdir -p foo
+cd foo
+../configure >/dev/null
if [ $? -ne 0 ]; then
echo "Error: failed to configure module."
cd $top_src
--
2.11.0
___
xorg-devel@lists.x.org: X.Org development
Archives: http://lists.x.org/archives/xorg-devel
Info: https://lists.x.org/mailman/listinfo/xorg-devel
Re: [PATCH util-modular 4/4] release.sh: implement around git worktree
On 23 January 2017 at 04:03, Peter Hutterer wrote:
> On Fri, Jan 20, 2017 at 02:19:06PM +, Emil Velikov wrote:
>> On 20 January 2017 at 02:49, Peter Hutterer wrote:
>> > On Thu, Jan 19, 2017 at 07:30:10PM +, Emil Velikov wrote:
>> >> From: Emil Velikov
>> >>
>> >> Months ago, before my work in here there the script had a number of
>> >> assumptions:
>> >> - autoreconf or alike is ran by the user - slight nuisance esp. when
>> >> releasing multiple packages.
>> >> - the generated files were compatible with the ones on the server -
>> >> rarely and issue, but still
>> >> - config.status means 'all the autotools bits were generated correctly'
>> >> - failed to consider if user has multiple config.status - using
>> >> ./build-64bit/ and ./build-32bit/ anyone ?
>> >> - ...
>> >>
>> >> With 663364cda5e316a0509ff5869293e3a815b9945f we mitigated a lot of that
>> >> but did not consider
>> >> - forcing people to `git clean' is annoying and if not careful one can
>> >> purge local files that they want to keep
>> >> - the files generated by autotools were left behind
>> >>
>> >> In order to mitigate those use git worktree. This is more efficient
>> >> [both bandwidth and storage wise] than pulling/generating git tarballs,
>> >> git clone and friends.
>> >>
>> >> Cc: Peter Hutterer
>> >> Signed-off-by: Emil Velikov
>> >> ---
>> >> There's a git of noise produced by git worktree and autoreconf. I'm fine
>> >> with 2>/dev/null but I'd leave that to others to decide
>> >> ---
>> >> release.sh | 34 +++---
>> >> 1 file changed, 19 insertions(+), 15 deletions(-)
>> >>
>> >> diff --git a/release.sh b/release.sh
>> >> index 8ed179d..0b3fcbf 100755
>> >> --- a/release.sh
>> >> +++ b/release.sh
>> >> @@ -357,25 +357,24 @@ process_module() {
>> >> return 1
>> >> fi
>> >>
>> >> -if [ -e configure ]; then
>> >> -echo "Error: the git repository contains configure"
>> >> -echo "Did you forget to run git clean -fXd (and git clean -fxd)
>> >> ?"
>> >> -return 1
>> >> -fi
>> >> -
>> >> -# Create tmpdir for the build
>> >> +# Create tmpdir for the release
>> >> build_dir=`mktemp -d -p . build.XX`
>> >> if [ $? -ne 0 ]; then
>> >> -echo "Error: could not create a temporary directory for the
>> >> build."
>> >> +echo "Error: could not create a temporary directory for the
>> >> release"
>> >> echo "Do you have coreutils' mktemp ?"
>> >> return 1
>> >> fi
>> >>
>> >> -echo "Info: generating configure."
>> >> -autoreconf --force --install >/dev/null
>> >> +# Worktree removal is intentionally left to the user, due to:
>> >> +# - currently we cannot select only one worktree to prune
>> >> +# - requires to removal of $build_dir which might contradict with
>> >> the
>> >> +# user decision to keep some artefacts like tarballs or other
>> >> +echo "Info: creating new git worktree."
>> >> +git worktree add $build_dir
>> >> if [ $? -ne 0 ]; then
>> >> -echo "Error: failed to generate configure."
>> >> -return 1
>> >> + echo "Error: failed to create a git worktree."
>> >> + cd $top_src
>> >> + return 1
>> >
>> > there's an indentation error,
>> Funky indentation was there (it was a code movement) - we still have a
>> handful of tabs that are meant to be 8 spaces.
>>
>> > but otherwise the series looks good and is
>> > Reviewed-by: Peter Hutterer
>> >
>> Note: after another thought I've went with v2 for 3/4 (having a
>> version check and building from there).
>>
>> > I guess I'll test this on the next release ;)
>> >
>> Thank you very much. Please let me know if you spot anything odd.
>
> Ok, tested it on the libXi release and something is off, but haven't found
> out what it is yet. You can reproduce it by running a --dry-run on the libXi
> repo, for some reason make distcleancheck fails with
>
> ERROR: files left in build directory after distclean:
> ./config.sub
> ./ltmain.sh
> ./config.guess
> ./install-sh
> ./missing
> ./depcomp
> ./compile
> make[1]: *** [distcleancheck] Error 1
>
> since these are standard files, it's not a libXi issue, must be something to
> do with the worktrees or how we invoke configure. Reverting this patch
> makes it work again.
>
That's strange I haven't seen issues like the ones you mentioned.
Fetching a clean copy of libXi and doing "../path/to/release.sh .
--dry-run" brings up other issues :-]
Which version of autotools are you using ? See libtool --help
Using Arch here, packages patched as mentioned:
version:libtool (GNU libtool) 2.4.6
automake: automake (GNU automake) 1.15
autoconf: autoconf (GNU Autoconf) 2.69
https://git.archlinux.org/svntogit/packages.git/tree/trunk/automake-1.15-perl-regex.patch?h=packages/automake
- fix perl regular expression
https://git.archlinux.org/svntogit/packages.git/tree/trunk/autoconf-2.69-perl-5.22-autoscan.patch?h=p
