forwarded 622653 https://bugs.freedesktop.org/show_bug.cgi?id=36319 tags 622653 upstream fixed-upstream patch thanks Hello, On trečiadienis 13 Balandis 2011 19:58:06 Julian Andres Klode wrote: > Package: xserver-xorg-video-intel > Version: 2:2.14.902-1 > Severity: grave > > X crashed about 4 times today. It did not crash directly after the > upgrade or yesterday, as far as I can remember, but now it seems > to crash. I am running GNOME 3 with gnome-shell, and thus use > compositing. > > > I attached upgrade logs of the past days, as well as > a X log showing the bug. The relevant part is: > > Backtrace: > [ 5306.687] 0: /usr/bin/Xorg (xorg_backtrace+0x28) [0x4aacc8] > [ 5306.687] 1: /usr/bin/Xorg (0x400000+0x61e59) [0x461e59] > [ 5306.687] 2: /lib/libpthread.so.0 (0x7f52090f9000+0xef60) > [0x7f5209107f60] [ 5306.687] 3: /usr/lib/libdrm_intel.so.1 > (drm_intel_bo_emit_reloc+0x0) [0x7f5205c2adb0] [ 5306.687] 4: > /usr/lib/xorg/modules/drivers/intel_drv.so (0x7f5205e34000+0x227a0) > [0x7f5205e567a0] [ 5306.687] 5: > /usr/lib/xorg/modules/drivers/intel_drv.so (0x7f5205e34000+0x2797f) > [0x7f5205e5b97f] [ 5306.687] 6: > /usr/lib/xorg/modules/drivers/intel_drv.so (0x7f5205e34000+0x29d5f) > [0x7f5205e5dd5f] [ 5306.687] 7: > /usr/lib/xorg/modules/drivers/intel_drv.so (0x7f5205e34000+0x27473) > [0x7f5205e5b473] [ 5306.687] 8: /usr/bin/Xorg (miPolyRectangle+0xd8) > [0x550228] > [ 5306.687] 9: /usr/bin/Xorg (0x400000+0xdf142) [0x4df142] > [ 5306.687] 10: /usr/bin/Xorg (0x400000+0x46082) [0x446082] > [ 5306.687] 11: /usr/bin/Xorg (0x400000+0x48909) [0x448909] > [ 5306.687] 12: /usr/bin/Xorg (0x400000+0x257ab) [0x4257ab] > [ 5306.687] 13: /lib/libc.so.6 (__libc_start_main+0xfd) [0x7f5207e55c4d] > [ 5306.687] 14: /usr/bin/Xorg (0x400000+0x25339) [0x425339] > [ 5306.687] Segmentation fault at address (nil) > [ 5306.687] > Fatal server error: > [ 5306.687] Caught signal 11 (Segmentation fault). Server aborting > [ 5306.687] > [ 5306.687] The bug is i965 specific. Upstream has just fixed it hence the fix is NOT included in the latest 2.15.0 release. It would great if you could add the attached patch (which is a backport from upstream) to the 2.15.0 package. -- Modestas Vainius <modestas@vainius.eu>
From: Chris Wilson <chris@chris-wilson.co.uk>
Subject: i965/video: We need 150 dwords of space for video state emission
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=36319
Bug-Debian: http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=622653
Origin: backport, commit:c9fb69cb2502917dfb2828c90802de7766072899,
commit:a51cd83d25f2f9f2107219d5671194f931601244
Last-Update: 2011-04-17
(Actually around 131, with additional 10% just for safety.)
intel: Beware the unsigned promotion when checking for batch overflows
diff --git a/src/i965_video.c b/src/i965_video.c
index c757681..53a9394 100644
--- a/src/i965_video.c
+++ b/src/i965_video.c
@@ -1210,7 +1210,7 @@ I965DisplayVideoTextured(ScrnInfoPtr scrn,
intel_batch_submit(scrn);
}
- intel_batch_start_atomic(scrn, 100);
+ intel_batch_start_atomic(scrn, 150);
i965_emit_video_setup(scrn, surface_state_binding_table_bo, n_src_surf, pixmap);
diff --git a/src/intel_batchbuffer.h b/src/intel_batchbuffer.h
index 605932a..2403a38 100644
--- a/src/intel_batchbuffer.h
+++ b/src/intel_batchbuffer.h
@@ -50,14 +50,14 @@ static inline int intel_vertex_space(intel_screen_private *intel)
}
static inline void
-intel_batch_require_space(ScrnInfoPtr scrn, intel_screen_private *intel, unsigned int sz)
+intel_batch_require_space(ScrnInfoPtr scrn, intel_screen_private *intel, int sz)
{
assert(sz < intel->batch_bo->size - 8);
if (intel_batch_space(intel) < sz)
intel_batch_submit(scrn);
}
-static inline void intel_batch_start_atomic(ScrnInfoPtr scrn, unsigned int sz)
+static inline void intel_batch_start_atomic(ScrnInfoPtr scrn, int sz)
{
intel_screen_private *intel = intel_get_screen_private(scrn);
Attachment:
signature.asc
Description: This is a digitally signed message part.