Re: Native X (8.1/amd64) crashing with gvim-gtk3
On Mon 22 Jul 2019 at 21:01:07 +, m...@netbsd.org wrote: > Another workaround is to try UXA as an AccelMethod for the intel driver. > (I am not good at writing X.org configs.) I tried that and in my first try it seemed like it changed nothing (there didn't seem anything visibly different in the X log file), but when I retried it just now, it seemed to work (and help). So, Thanks! -Olaf. -- ___ Olaf 'Rhialto' Seibert -- "What good is a Ring of Power \X/ rhialto/at/falu.nl -- if you're unable...to Speak." - Agent Elrond signature.asc Description: PGP signature
Re: Native X (8.1/amd64) crashing with gvim-gtk3
On Tue 23 Jul 2019 at 07:48:59 -0500, John D. Baker wrote: > The X server usually crashes with "SIGSEGV" (logged in Xorg.0.log.old), > but doesn't leave a core dump. One time it got "SIGALARM" and left a > core dump. What I did to catch it, was to ssh into the machine from another one, and then run gdb -p , and (since it stops X) continue it. You obviously can't do this from an xterm on the same X server but it can probably work as well from a text console (well, if the switching to the console works, I'm not 100% sure). -Olaf. -- ___ Olaf 'Rhialto' Seibert -- "What good is a Ring of Power \X/ rhialto/at/falu.nl -- if you're unable...to Speak." - Agent Elrond signature.asc Description: PGP signature
Re: Native X (8.1/amd64) crashing with gvim-gtk3
On Tue, 23 Jul 2019, John D. Baker wrote: > Sometimes it will survive its initial encounter with 'gnumeric', but > will crash much later. As I use 'gnumeric' regularly, it's difficult > to test whether it will crash without that trigger. Another interesting feature: Once it has crashed and been re-launched by 'xdm' (or another display manager?) it will usually be trivial to make it crash right away running the trigger application. After it crashes, only a full reboot of the machine will make it less likely to crash upon launching/using the trigger application. It seems the later in the session the trigger application is launched, the more likely it will survive the initial trigger and take longer before it crashes, but it will eventually crash. -- |/"\ John D. Baker, KN5UKS NetBSD Darwin/MacOS X |\ / jdbaker[snail]consolidated[flyspeck]net OpenBSDFreeBSD | X No HTML/proprietary data in email. BSD just sits there and works! |/ \ GPGkeyID: D703 4A7E 479F 63F8 D3F4 BD99 9572 8F23 E4AD 1645
Re: Native X (8.1/amd64) crashing with gvim-gtk3
I have seen similar. The Xorg will spontaneously crash. My trigger is 'gnumeric': Launch gnumeric (might crash from just this) Move around worksheet (or it might wait until this) Modify a cell (might survive until here) Switch to another worksheet (ad nauseum) etc. Sometimes it will survive its initial encounter with 'gnumeric', but will crash much later. As I use 'gnumeric' regularly, it's difficult to test whether it will crash without that trigger. The X server usually crashes with "SIGSEGV" (logged in Xorg.0.log.old), but doesn't leave a core dump. One time it got "SIGALARM" and left a core dump. The affected machine was likewise equipped with Intel integrated graphics (G41) with the xf86-video-intel driver and SNA acceleration. Up until that point it had been rock solid. I'd been suspecting hardware beginning to fail, but your post would suggest otherwise. It's been a couple of months since this started and I've switched to an older, smaller, slower machine w/radeon video in the meantime. I couldn't afford having SSH sessions broken due to X crashing. I think I've seen similar using -current with the xf86-video-intel-2014 driver (needed to get working SNA acceleration on my HW). I don't recall any changes to 8.1 that should have affected X-on-intel like this. -- |/"\ John D. Baker, KN5UKS NetBSD Darwin/MacOS X |\ / jdbaker[snail]consolidated[flyspeck]net OpenBSDFreeBSD | X No HTML/proprietary data in email. BSD just sits there and works! |/ \ GPGkeyID: D703 4A7E 479F 63F8 D3F4 BD99 9572 8F23 E4AD 1645
Re: Native X (8.1/amd64) crashing with gvim-gtk3
On Mon 22 Jul 2019 at 21:11:23 +0200, Tobias Nygren wrote: > On Mon, 22 Jul 2019 21:02:21 +0200 > Rhialto wrote: > > > On Mon 22 Jul 2019 at 20:49:12 +0200, Tobias Nygren wrote: > > > On Mon, 22 Jul 2019 20:38:14 +0200 > > > Rhialto wrote: > > > > > > > at > > > > /usr/xsrc/external/mit/xf86-video-intel/dist/src/sna/gen4_vertex.c:854 > > > > > > Does 8.1 ship with modesetting support for intel? I had a similar crash > > > some time back and the solution was to switch Xorg from > > > xf86-video-intel to the modesetting driver. > > > > If so, how do I switch? I'd like to try that. > > mkdir -p /etc/X11/xorg.conf.d > cat << EOF > /etc/X11/xorg.conf.d/10-modesetting.conf > Section "device" > # Option "NoAccel" "true" > Identifier "Card0" > Driver "modesetting" > EndSection > EOF Thanks. I tried it, and X tried it, but it didn't seem to work. [ 6320.041] (II) LoadModule: "modesetting" [ 6320.042] (II) Loading /usr/X11R7/lib/modules/drivers/modesetting_drv.so [ 6320.043] (II) Module modesetting: vendor="X.Org Foundation" [ 6320.043]compiled for 1.18.4, module version = 1.18.4 [ 6320.043]Module class: X.Org Video Driver [ 6320.043]ABI class: X.Org Video Driver, version 20.0 [ 6320.043] (II) LoadModule: "mouse" [ 6320.044] (II) Loading /usr/X11R7/lib/modules/drivers/mouse_drv.so [ 6320.044] (II) Module mouse: vendor="X.Org Foundation" [ 6320.044]compiled for 1.18.4, module version = 1.9.2 [ 6320.044]Module class: X.Org XInput Driver [ 6320.044]ABI class: X.Org XInput driver, version 22.1 [ 6320.044] (II) LoadModule: "kbd" [ 6320.045] (II) Loading /usr/X11R7/lib/modules/drivers/kbd_drv.so [ 6320.045] (II) Module kbd: vendor="X.Org Foundation" [ 6320.045]compiled for 1.18.4, module version = 1.9.0 [ 6320.046]Module class: X.Org XInput Driver [ 6320.046]ABI class: X.Org XInput driver, version 22.1 [ 6320.046] (II) modesetting: Driver for Modesetting Kernel Drivers: kms [ 6320.046] (--) Using wscons driver on /dev/ttyE4 in pcvt compatibility mode (version 3.32) [ 6320.046] (--) using VT number 5 [ 6320.062] (**) modeset(0): claimed PCI slot 0@0:2:0 [ 6320.062] (II) modeset(0): using default device [ 6320.062] (WW) VGA arbiter: cannot open kernel arbiter, no multi-card support [ 6320.062] (II) UnloadModule: "modesetting" [ 6320.062] (EE) Screen(s) found, but none have a usable configuration. [ 6320.062] (EE) Fatal server error: [ 6320.063] (EE) no screens found(EE) I also reduced the rest of my xorg.conf file to nothing, but that made no differnce. -Olaf. -- ___ Olaf 'Rhialto' Seibert -- "What good is a Ring of Power \X/ rhialto/at/falu.nl -- if you're unable...to Speak." - Agent Elrond signature.asc Description: PGP signature
Re: Native X (8.1/amd64) crashing with gvim-gtk3
On Mon, 22 Jul 2019 21:02:21 +0200 Rhialto wrote: > On Mon 22 Jul 2019 at 20:49:12 +0200, Tobias Nygren wrote: > > On Mon, 22 Jul 2019 20:38:14 +0200 > > Rhialto wrote: > > > > > at /usr/xsrc/external/mit/xf86-video-intel/dist/src/sna/gen4_vertex.c:854 > > > > Does 8.1 ship with modesetting support for intel? I had a similar crash > > some time back and the solution was to switch Xorg from > > xf86-video-intel to the modesetting driver. > > If so, how do I switch? I'd like to try that. mkdir -p /etc/X11/xorg.conf.d cat << EOF > /etc/X11/xorg.conf.d/10-modesetting.conf Section "device" # Option "NoAccel" "true" Identifier "Card0" Driver "modesetting" EndSection EOF
Re: Native X (8.1/amd64) crashing with gvim-gtk3
On Mon 22 Jul 2019 at 20:49:12 +0200, Tobias Nygren wrote: > On Mon, 22 Jul 2019 20:38:14 +0200 > Rhialto wrote: > > > at /usr/xsrc/external/mit/xf86-video-intel/dist/src/sna/gen4_vertex.c:854 > > Does 8.1 ship with modesetting support for intel? I had a similar crash > some time back and the solution was to switch Xorg from > xf86-video-intel to the modesetting driver. If so, how do I switch? I'd like to try that. > -Tobias -Olaf. -- ___ Olaf 'Rhialto' Seibert -- "What good is a Ring of Power \X/ rhialto/at/falu.nl -- if you're unable...to Speak." - Agent Elrond signature.asc Description: PGP signature
Re: Native X (8.1/amd64) crashing with gvim-gtk3
On Mon, 22 Jul 2019 20:38:14 +0200 Rhialto wrote: > at /usr/xsrc/external/mit/xf86-video-intel/dist/src/sna/gen4_vertex.c:854 Does 8.1 ship with modesetting support for intel? I had a similar crash some time back and the solution was to switch Xorg from xf86-video-intel to the modesetting driver. -Tobias
Native X (8.1/amd64) crashing with gvim-gtk3
(or maybe just any gnome3 program) It is quite repeatable for me: start gvim with a file, move cursor down so that it scrolls. Sometimes it crashes even earlier. Stack backtrace (great to have full debug files available!) Program received signal SIGSEGV, Segmentation fault. 0x76ca23b0ee90 in emit_primitive_identity_mask (sna=, op=0x7f7fff788ce0, r=0x7f7fff788cd0) at /usr/xsrc/external/mit/xf86-video-intel/dist/src/sna/gen4_vertex.c:854 854 /usr/xsrc/external/mit/xf86-video-intel/dist/src/sna/gen4_vertex.c: No such file or directory. (gdb) bt #0 0x76ca23b0ee90 in emit_primitive_identity_mask (sna=, op=0x7f7fff788ce0, r=0x7f7fff788cd0) at /usr/xsrc/external/mit/xf86-video-intel/dist/src/sna/gen4_vertex.c:854 #1 0x76ca23a6feaf in glyphs0_to_dst (sna=sna@entry=0x76ca2d301000, op=op@entry=3 '\003', src=src@entry=0x76ca27908b00, dst=dst@entry=0x76ca27908980, src_x=0, src_x@entry=2, src_y=0, src_y@entry=69, nlist=, nlist@entry=1, list=list@entry=0x7f7fff7890d0, glyphs=0x7f7fff7895e0, glyphs@entry=0x7f7fff7894d0) at /usr/xsrc/external/mit/xf86-video-intel/dist/src/sna/sna_glyphs.c:906 #2 0x76ca23a74862 in sna_glyphs (op=3 '\003', src=0x76ca27908b00, dst=0x76ca27908980, mask=0x0, src_x=, src_y=, nlist=1, list=0x7f7fff7890d0, glyphs=0x7f7fff7894d0) at /usr/xsrc/external/mit/xf86-video-intel/dist/src/sna/sna_glyphs.c:1998 #3 0x00015b56ebbf in damageGlyphs (op=, pSrc=0x76ca27908b00, pDst=0x76ca27908980, maskFormat=0x0, xSrc=, ySrc=, nlist=1, list=0x7f7fff7890d0, glyphs=0x7f7fff7894d0) at /usr/xsrc/external/mit/xorg-server/dist/miext/damage/damage.c:569 #4 0x00015b554b71 in ?? () #5 0x00015b486b84 in Dispatch () #6 0x00015b45a505 in dix_main () #7 0x00015b45a05b in ___start () #8 0x76ca2de31000 in ?? () #9 0x0005 in ?? () #10 0x7f7fff78a3c8 in ?? () #11 0x7f7fff78a3d9 in ?? () #12 0x7f7fff78a3dc in ?? () #13 0x7f7fff78a3e1 in ?? () #14 0x7f7fff78a3e7 in ?? () #15 0x in ?? () (gdb) info locals dst = {p = {x = 274, y = 69}, f = 6.33703319e-39} msk_x = 584 msk_y = 32 w = h = v = 0x76ca2d6dcff0 (gdb) print *op $1 = {blt = 0x76ca23aeb52c , box = 0x76ca23aeb895 , boxes = 0x76ca23aea420 , thread_boxes = 0x0, done = 0x76ca23ae69ae , damage = 0x0, op = 8, dst = {pixmap = 0x76ca2818df00, format = 537036936, bo = 0x76ca2a347800, x = 0, y = 0, width = 644, height = 704}, src = {bo = 0x76ca2cf8ea60, transform = 0x0, width = 1, height = 1, pict_format = 537036936, card_format = 192, filter = 0, repeat = 1, is_affine = 1, is_solid = 1, is_linear = 0, is_opaque = 1, alpha_fixup = 0, rb_reversed = 0, offset = { 0, 0}, scale = {1, 1}, embedded_transform = {matrix = {{0, 0, 0}, {0, 0, 0}, {0, 0, 0}}}, u = {linear = {dx = 0, dy = 0, offset = 0}, gen2 = { pixel = 0}, gen3 = {type = 0, mode = 0, constants = 0}}}, mask = { bo = 0x76ca2d7d4840, transform = 0x0, width = 1024, height = 1024, pict_format = 537036936, card_format = 192, filter = 0, repeat = 0, is_affine = 1, is_solid = 0, is_linear = 0, is_opaque = 0, alpha_fixup = 0, rb_reversed = 0, offset = {0, 0}, scale = {0.0009765625, 0.0009765625}, embedded_transform = {matrix = {{0, 0, 0}, {0, 0, 0}, {0, 0, 0}}}, u = {linear = {dx = 0, dy = 0, offset = 0}, gen2 = { pixel = 0}, gen3 = {type = 0, mode = 0, constants = 0}}}, is_affine = 1, has_component_alpha = 1, need_magic_ca_pass = 1, rb_reversed = 0, floats_per_vertex = 4, floats_per_rect = 12, prim_emit = 0x76ca23b0ede9 , emit_boxes = 0x76ca23b0eeeb , redirect = { real_bo = 0x0, real_damage = 0x0, damage = 0x0, box = {x1 = 0, y1 = 0, x2 = 0, y2 = 0}}, u = {blt = {src_pixmap = 0x90006, sx = 0, sy = 0, inplace = 0, overwrites = 0, bpp = 0, alu = 0, cmd = 0, br13 = 0, pitch = {0, 0}, pixel = 0, bo = {0x0, 0x0}}, gen3 = {constants = { 8.40779079e-45, 1.26116862e-44, 0, 0, 0, 0, 0, 0}, num_constants = 0}, gen4 = {wm_kernel = 6, ve_id = 9}, gen5 = {wm_kernel = 6, ve_id = 0}, gen6 = {flags = 6}, gen7 = {flags = 6}, gen8 = {flags = 6}}, priv = 0x0} (gdb) print v[7] $2 = 0 (gdb) print v[3] $3 = 0.0390625 (gdb) print op->mask.scale[1] $4 = 0.0009765625 (gdb) In version 1.1.1.3 there is this, and all subexpressions from line 854 seem to be ok, at least gdb can print values. 850 dst.p.x = r->dst.x + r->width; 851 dst.p.y = r->dst.y + r->height; 852 v[0] = dst.f; 853 v[2] = (msk_x + w) * op->mask.scale[0]; 854 v[7] = v[3] = (msk_y + h) * op->mask.scale[1]; 855 856 dst.p.x = r->dst.x; -Olaf. -- ___ Olaf 'Rhialto' Seibert -- "What good is a Ring of Power \X/ rhialto/at/falu.nl -- if you're unable...to Speak." - Agent Elrond signature.asc Description: PGP signature