similar to: [ANNOUNCE] xf86-video-nouveau 1.0.16

This causes nouveau_drv.so to not be loadable due to a missing wfbPictureInit. I'm not 100% sure what causes it -- it happens at runtime in the middle of probe, and I don't have wfb enabled. Either way, I'm going to revert this until we get a better idea of what's going on. On Sun, Jan 20, 2019 at 7:20 PM Rhys Kidd <rhyskidd at gmail.com> wrote: > > Fixes warning with

Fixes warning with xserver 1.20 and gcc 8.2: nv_driver.c:1443:9: warning: implicit declaration of function ‘wfbScreenInit’; did you mean ‘fbScreenInit’? [-Wimplicit-function-declaration] ret = wfbScreenInit(pScreen, FBStart, pScrn->virtualX, ^~~~~~~~~~~~~ fbScreenInit (See xserver 706e6d9cd074da606016ed4ecff51e9c2a822087) Signed-off-by: Rhys Kidd <rhyskidd at

https://bugs.freedesktop.org/show_bug.cgi?id=63263 Priority: medium Bug ID: 63263 Assignee: nouveau at lists.freedesktop.org Summary: X server crash in nouveau_xv.c:NVPutImage (NVCopyNV12ColorPlanes) QA Contact: xorg-team at lists.x.org Severity: normal Classification: Unclassified OS: Linux (All)

int i is accessed outside immediate scope so leave declaration at the highest common scope level: 1073: int ret, i; ... // Highest common scope ... 1193: if (newTTSize <= destination_buffer->size) { ... // Used in this scope ... 1248: } else { ... // Used in this scope ... 1316: }

- added lots of comments, discussed comments with ahuillet and refined comments - replaced two constants with MACROs - no changes to actual code diff --git a/src/nv_video.c b/src/nv_video.c index cf4f88d..d9ee700 100644 --- a/src/nv_video.c +++ b/src/nv_video.c @@ -25,6 +25,15 @@ #include "nv_include.h" #include "nv_dma.h" +/* +2046 is the maximum image size in one

Not much has changed, except that i've test run these patches for a few weeks now. I'm actually surprised i've avoided Out-Of-Memory issues. Patch 2 remains a temporary hack, awaiting a structural fix by darktama. The whole thing works fine, but XSHM is an issue (http://stillunknown.livejournal.com/928.html). With it disabled most apps are fine, although a few issues remain. As

Warning reported by gcc 8.2: nouveau_xv.c: In function ‘NVPutImage’: nouveau_xv.c:1369:7: warning: declaration of ‘ret’ shadows a previous local [-Wshadow] int ret = BadImplementation; ^~~ nouveau_xv.c:1073:6: note: shadowed declaration is here int ret, i; ^~~ Fixes: a50bba6 ("xv: misc cleanups") Cc: Ben Skeggs <bskeggs at redhat.com> Signed-off-by: Rhys Kidd

http://bugs.freedesktop.org/show_bug.cgi?id=22253 Summary: KMS Nouveau - wfb crash Product: xorg Version: unspecified Platform: x86-64 (AMD64) OS/Version: Linux (All) Status: NEW Severity: major Priority: medium Component: Driver/nouveau AssignedTo: nouveau at lists.freedesktop.org

- Only for sufficiently new xserver's and exa_driver_pixmaps. --- src/nouveau_exa.c | 217 +++++++++++++++++++++++++++++++++++++++++++++++++++-- src/nv_driver.c | 51 +++++++++++-- src/nv_proto.h | 4 + src/nv_type.h | 12 +++- 4 files changed, 267 insertions(+), 17 deletions(-) diff --git a/src/nouveau_exa.c b/src/nouveau_exa.c index 93fc3c5..074a226 100644 ---

These two patches make nv50 usable with driver allocated pixmaps. I am aware that some cards (NV9X?) have a slightly different texture layout somehow, i would encourage people to use the test program i sent a few days ago to try and figure it out, or at least sent in the output (see the last email for details). Sorry for all the bitmasking, shifting and such in the first patch but is absolutely

The 'w' argument to NVCopyNV12ColorPlanes is used to index into the source image. line_len is rounded up to 8 on NV50+, so if the source image (+ left offset) is not rounded to 8, NVCopyNV12ColorPlanes could read past the end of the array and crash X. This change can cause the last few horizontal pixels of dst to not be initialized, but they should be truncated by the renderer anyways.

memcpy() goes from taking 45% to 66% of total function time, which translates to a 30% decrease in NVPutImage runtime. Signed-off-by: Ilia Mirkin <imirkin at alum.mit.edu> --- src/nouveau_xv.c | 33 ++++++++++++++++++++++++++------- 1 file changed, 26 insertions(+), 7 deletions(-) diff --git a/src/nouveau_xv.c b/src/nouveau_xv.c index 567e30c..5569b7c 100644 --- a/src/nouveau_xv.c +++

This is the first release candidate for xserver 1.20. Notable changes since 1.19 include: - RANDR 1.6, which enables leasing RANDR resources to a client for its exclusive use (e.g. head mounted displays) - Depth 30 support in glamor and the modesetting driver - A meson-based build system, parallel to autotools - Pageflipping support for PRIME output sinks - OutputClass device matching for

http://bugs.freedesktop.org/show_bug.cgi?id=19860 Summary: xv texture adaptor kills X Product: xorg Version: git Platform: x86-64 (AMD64) OS/Version: Linux (All) Status: NEW Severity: normal Priority: medium Component: Driver/nouveau AssignedTo: nouveau at lists.freedesktop.org ReportedBy:

The first patch (for drm) fixes a very annoying bug for nv50 gart. The 2nd is just a few cleanups in libnouveau_drm. I also pushed a few changes to the ng ddx, and I'm at the moment thinking on how to deal with the issue of nv50. We are no longer garuanteed that pixmaps are really migrated upon mapping, so that is the first thing to do. Probably a good idea to begin with making the bo copy

I've got a preliminary wfb implementation, but it still has issues. In the meantime i've added what i know to the wiki. Maarten.

A short series of compiler visibility warning fixes that I prepared whilst trialing improvements to xf86-video-nouveau's use of the core xorg-server utility macros. Rhys Kidd (4): wfb: Remove declaration for undefined function nouveau_wfb_init() dri2: Mark local create/destroy buffer and copy region functions as static xv: Mark local NVSetupTexturedVideo function as static

With pleasure we announce the release of version 1.0.16. Here is a summary of the changes: * Fixed a performance issue with TCP communication under Windows. * Fixed code that, during network outages, would cause tinc to exit when it thought two nodes with identical Names were on the VPN. This version of tinc is compatible with 1.0pre8, 1.0 and later, but not with earlier version of tinc.

With pleasure we announce the release of version 1.0.16. Here is a summary of the changes: * Fixed a performance issue with TCP communication under Windows. * Fixed code that, during network outages, would cause tinc to exit when it thought two nodes with identical Names were on the VPN. This version of tinc is compatible with 1.0pre8, 1.0 and later, but not with earlier version of tinc.

The control flow was: if (!y) { ppix = ... } if (y) { ... } else if (x) { use ppix for something } else { use ppix for something } Merge the if(!y) block with the two else branches. This avoids a false-positive in the clang static analyzer, it can't know that !y and x are mutually exclusive. The result looks something like this: if (y) { ... } else { ppix = ... if (x) {