[Bug 9878] New: Team Fortress 2 crashes when joining team
http://bugs.winehq.org/show_bug.cgi?id=9878 Summary: Team Fortress 2 crashes when joining team Product: Wine Version: CVS/GIT Platform: Other URL: http://steampowered.com/ OS/Version: other Status: NEW Keywords: regression Severity: normal Priority: P2 Component: wine-kernel AssignedTo: wine-bugs(a)winehq.org ReportedBy: vitaliy(a)kievinfo.com CC: julliard(a)winehq.org Start the TF2, create server, select CP_GRANARY map, after server starts, try joining any team - crash. This is a regression. git bisect pointed to this patch: commit 1d063ae18d990343fc077dcbf650add797924018 Author: Alexandre Julliard <julliard(a)winehq.org> Date: Mon Oct 1 15:28:50 2007 +0200 ntdll: Remove assumptions that the subheap is at the beginning of the memory block. -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=9878 Vitaliy Margolen <vitaliy(a)kievinfo.com> changed: What |Removed |Added ---------------------------------------------------------------------------- OS/Version|other |Linux Platform|Other |PC -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=9878 --- Comment #1 from Vitaliy Margolen <vitaliy(a)kievinfo.com> 2007-10-03 08:41:10 --- Minidump of not much use: vitaliy(a)dragon:Steam $wine winedbg "SteamApps/dumb2/team fortress 2/Steam__3250__2007_10_3T13_31_18C91711.mdmp" WineDbg starting on minidump on pid 0044 hl2.exe was running on #1 Intel Pentium Pro/II or AMD Athlon (0.2560) CPU on Windows 2000 (2195) Register dump: CS:0073 SS:007b DS:007b ES:007b FS:0033 GS:003b EIP:0d3fed72 ESP:0034d5b0 EBP:0034d690 EFLAGS:00010212( - 00 - RIA1) EAX:3f2948bb EBX:0034d908 ECX:3f414a1b EDX:35d629c8 ESI:0034d600 EDI:35d629c8 Stack dump: 0x0034d5b0: 0d456680 00000015 ff000000 00000000 0x0034d5c0: 00000000 00000000 06375488 00000000 0x0034d5d0: 35d629c8 2e48d850 3f800000 2e4c62d0 0x0034d5e0: 0ce5b310 00000018 7c98c54e 0dd9e060 0x0034d5f0: 00000028 0dd9e060 0dd9e060 0034d600 0x0034d600: 43a66231 4032b0ce 41f31ab1 3f800000 Backtrace: =>1 0x0d3fed72 in studiorender (+0xed72) (0x0034d690) WineDbg starting on pid 0044 -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=9878 --- Comment #2 from Simon <sileon(a)gmail.com> 2007-10-04 11:02:55 --- Created an attachment (id=8385) --> (http://bugs.winehq.org/attachment.cgi?id=8385) Output of the ET:QW crash Enemy Territory: Quake Wars is affected by this bug too and it crashes during startup. I did a regression test myself and got the same result. Attached is the output of a dying ET:QW. -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=9878 Simon <sileon(a)gmail.com> changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |sileon(a)gmail.com -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=9878 Vitaliy Margolen <vitaliy(a)kievinfo.com> changed: What |Removed |Added ---------------------------------------------------------------------------- Attachment #8385|application/octet-stream |text/plain mime type| | -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=9878 Vitaliy Margolen <vitaliy(a)kievinfo.com> changed: What |Removed |Added ---------------------------------------------------------------------------- Severity|normal |major Summary|Team Fortress 2 crashes when|Most Steam games crash when |joining team |starting/loading a game --- Comment #3 from Vitaliy Margolen <vitaliy(a)kievinfo.com> 2007-10-11 23:34:22 --- Now it affects: Team Fortress 2 Half Life 2 Half Life 2 ep2 Portal Lost Coast -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=9878 --- Comment #4 from Alexandre Julliard <julliard(a)winehq.org> 2007-10-12 02:25:47 --- You should run with warn+heap and try to figure out where the heap gets corrupted. -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=9878 Stefan Dösinger <stefandoesinger(a)gmx.at> changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |stefandoesinger(a)gmx.at --- Comment #5 from Stefan Dösinger <stefandoesinger(a)gmx.at> 2007-10-12 13:58:04 --- a warn+heap trace doesn't change anything for me. The crash occurs at the same place ingame, nothing written on the warn channel, backtrace is the same. -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=9878 Flexo <spokehedz(a)gmail.com> changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |spokehedz(a)gmail.com --- Comment #6 from Flexo <spokehedz(a)gmail.com> 2007-10-13 11:54:02 --- I have found that by changing the resolution down that this error goes away entirely. I used to play at 1280x1024 but now since changing to 1024x768 the issue has stopped entirely. I am currently playing Portals, and I have not tested TF2. -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=9878 --- Comment #7 from Stefan Dösinger <stefandoesinger(a)gmx.at> 2007-10-13 13:03:57 --- (In reply to comment #6)
I have found that by changing the resolution down that this error goes away entirely. I used to play at 1280x1024 but now since changing to 1024x768 the issue has stopped entirely. I am currently playing Portals, and I have not tested TF2. I think you see a different problem here: A crash due to font loading. This bug is resolution independent for me(plain HL2 and HL2EP2 crash at 800x600).
-- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=9878 --- Comment #8 from Flexo <spokehedz(a)gmail.com> 2007-10-13 16:49:16 --- (In reply to comment #7)
(In reply to comment #6)
I have found that by changing the resolution down that this error goes away entirely. I used to play at 1280x1024 but now since changing to 1024x768 the issue has stopped entirely. I am currently playing Portals, and I have not tested TF2. I think you see a different problem here: A crash due to font loading. This bug is resolution independent for me(plain HL2 and HL2EP2 crash at 800x600).
Then why have I been playing all day, up until just 5 min ago when I changed the resolution and it crashed? -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=9878 --- Comment #9 from Vitaliy Margolen <vitaliy(a)kievinfo.com> 2007-10-14 11:11:17 --- Not resolution dependent. Crashes in 640x480, 800x600, 1024x768. This seems to be a memory corruption, at it can manifest itself differently on every setup. And Flexo, please if you don't have this bug - good, you can continue on playing and ignore this bug. -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=9878 Vitaliy Margolen <vitaliy(a)kievinfo.com> changed: What |Removed |Added ---------------------------------------------------------------------------- CC|spokehedz(a)gmail.com | -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=9878 --- Comment #10 from Flexo <spokehedz(a)gmail.com> 2007-10-14 11:54:36 --- Created an attachment (id=8564) --> (http://bugs.winehq.org/attachment.cgi?id=8564) Picture of the Error when changing the resolution to 1280x1024 Well, if this isn't the bug, what do I have to do in order to help you guys determine what bug I have? -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=9878 --- Comment #11 from Vitaliy Margolen <vitaliy(a)kievinfo.com> 2007-10-14 12:11:43 --- (In reply to comment #10) That usually means "out of memory" - reboot your PC. Video driver leaked enough video ram. Or you setting too high of the settings for the amount of VRAM you have. Or you have not set VideoMemorySize registry settings to the size of the VRAM your card has. In either case it is not related to this bug. There are lots of bugs that affect Half-Life 2 and most of them lead to a crash. -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=9878 Vitaliy Margolen <vitaliy(a)kievinfo.com> changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |waninkoko(a)gmail.com --- Comment #12 from Vitaliy Margolen <vitaliy(a)kievinfo.com> 2007-10-15 09:23:49 --- *** Bug 10044 has been marked as a duplicate of this bug. *** -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=9878 tsubasa <sylvain.pasquet(a)gmail.com> changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |sylvain.pasquet(a)gmail.com --- Comment #13 from tsubasa <sylvain.pasquet(a)gmail.com> 2007-10-15 15:23:21 --- For recovering source game that crashing on startup we need to following this : You need to disable steam community things. I have a french steam client but by the way for me its : 1-Go in file menu and select the parameters. 2- Go in "In the game" tab 3- Uncheck the Enable community steam. Restart steam and all will work like before. -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=9878 Viktor Varland <viktor(a)varland.se> changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |viktor(a)varland.se --- Comment #14 from Viktor Varland <viktor(a)varland.se> 2007-10-16 06:25:27 --- (In reply to comment #13)
For recovering source game that crashing on startup we need to following this :
You need to disable steam community things. I have a french steam client but by the way for me its : 1-Go in file menu and select the parameters. 2- Go in "In the game" tab 3- Uncheck the Enable community steam.
Restart steam and all will work like before.
This does not fix the bug for me. I've had the Steam Community deactivated since wine 0.9.44 with Source-based games working. When upgrading to 0.9.47 the games (primarily TF2) started crashing when I click "Join team", up until then everything is fine. -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=9878 --- Comment #15 from tsubasa <sylvain.pasquet(a)gmail.com> 2007-10-16 06:59:44 --- (In reply to comment #14)
(In reply to comment #13)
For recovering source game that crashing on startup we need to following this :
You need to disable steam community things. I have a french steam client but by the way for me its : 1-Go in file menu and select the parameters. 2- Go in "In the game" tab 3- Uncheck the Enable community steam.
Restart steam and all will work like before.
This does not fix the bug for me. I've had the Steam Community deactivated since wine 0.9.44 with Source-based games working. When upgrading to 0.9.47 the games (primarily TF2) started crashing when I click "Join team", up until then everything is fine.
I have this bug to but you need to set the detail for the model to low, if not game will crash when joining a team. -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=9878 --- Comment #16 from Viktor Varland <viktor(a)varland.se> 2007-10-16 08:35:08 --- (In reply to comment #15)
(In reply to comment #14)
(In reply to comment #13)
For recovering source game that crashing on startup we need to following this :
You need to disable steam community things. I have a french steam client but by the way for me its : 1-Go in file menu and select the parameters. 2- Go in "In the game" tab 3- Uncheck the Enable community steam.
Restart steam and all will work like before.
This does not fix the bug for me. I've had the Steam Community deactivated since wine 0.9.44 with Source-based games working. When upgrading to 0.9.47 the games (primarily TF2) started crashing when I click "Join team", up until then everything is fine.
I have this bug to but you need to set the detail for the model to low, if not game will crash when joining a team.
Doing this results in that the game starts, I can join servers and teams (hitting enter), but the entire screen is light blue from joining the game to quit. I can escape into the main menu, no problem, but changing settings again results in a lock. -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=9878 --- Comment #17 from tsubasa <sylvain.pasquet(a)gmail.com> 2007-10-16 11:19:18 --- (In reply to comment #16)
(In reply to comment #15)
(In reply to comment #14)
(In reply to comment #13)
For recovering source game that crashing on startup we need to following this :
You need to disable steam community things. I have a french steam client but by the way for me its : 1-Go in file menu and select the parameters. 2- Go in "In the game" tab 3- Uncheck the Enable community steam.
Restart steam and all will work like before.
This does not fix the bug for me. I've had the Steam Community deactivated since wine 0.9.44 with Source-based games working. When upgrading to 0.9.47 the games (primarily TF2) started crashing when I click "Join team", up until then everything is fine.
I have this bug to but you need to set the detail for the model to low, if not game will crash when joining a team.
Doing this results in that the game starts, I can join servers and teams (hitting enter), but the entire screen is light blue from joining the game to quit. I can escape into the main menu, no problem, but changing settings again results in a lock.
If you have set in regedit the OffscreenRenderingMode key, try to delete it, the only custom reg key i have is the GLSL -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=9878 --- Comment #18 from Stefan Dösinger <stefandoesinger(a)gmx.at> 2007-10-16 12:02:57 --- I have the feeling that a number of different bugs are mixed here. Source based game crash != Source based game crash. The community feature problem is definitly not related to this patch, I have had a number of customer reports that this breaks in Crossover too, which is based on .34. -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=9878 --- Comment #19 from Stefan Dösinger <stefandoesinger(a)gmx.at> 2007-10-16 12:05:02 --- Also, please note that bug 7698 is a completely different problem, but can lead to very simmilar symptoms. -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=9878 --- Comment #20 from Viktor Varland <viktor(a)varland.se> 2007-10-16 12:58:55 --- (In reply to comment #17)
(In reply to comment #16)
(In reply to comment #15)
(In reply to comment #14)
(In reply to comment #13)
For recovering source game that crashing on startup we need to following this :
You need to disable steam community things. I have a french steam client but by the way for me its : 1-Go in file menu and select the parameters. 2- Go in "In the game" tab 3- Uncheck the Enable community steam.
Restart steam and all will work like before.
This does not fix the bug for me. I've had the Steam Community deactivated since wine 0.9.44 with Source-based games working. When upgrading to 0.9.47 the games (primarily TF2) started crashing when I click "Join team", up until then everything is fine.
I have this bug to but you need to set the detail for the model to low, if not game will crash when joining a team.
Doing this results in that the game starts, I can join servers and teams (hitting enter), but the entire screen is light blue from joining the game to quit. I can escape into the main menu, no problem, but changing settings again results in a lock.
If you have set in regedit the OffscreenRenderingMode key, try to delete it, the only custom reg key i have is the GLSL
I have no custom keys, and it still crashes when I join a team on a server. -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=9878 Vitaliy Margolen <vitaliy(a)kievinfo.com> changed: What |Removed |Added ---------------------------------------------------------------------------- CC|viktor(a)varland.se, | |sylvain.pasquet(a)gmail.com | --- Comment #21 from Vitaliy Margolen <vitaliy(a)kievinfo.com> 2007-10-16 20:32:21 --- DO NOT TURN bugzilla into forum. This bug is a memory corruption. It can manifest itself in lots of different ways. You "fix" doesn't fix anything. -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=9878 --- Comment #22 from Allan Tong <actong88(a)gmail.com> 2007-10-17 00:41:30 --- I find that if I pad the subheap header by 8 bytes, I no longer get the crash. In fact, making sure that the first free block in the subheap is 16 byte aligned seems to fix the issue for me. E.g. padding the subheap header by 8, 24, or 40 bytes fixes the problem, but if I pad by 16 or 32 bytes I still get the crash. FWIW the first free block used to be 16 byte aligned pre-patch. In my case, there seems to be one heap allocation that's causing the problem. HL2 calls CreateVertexBuffer to create a 2MB buffer in the default memory pool (D3DPOOL_DEFAULT), which as far as I understand means that it should be created in video memory. The current wine implementation ends up creating a 2MB buffer in the process heap. I've found that if I just make sure this one allocation is 16 byte aligned, then I don't get the crash. Is it possible that the game is assuming a 16-byte aligned buffer? Is that normal for a buffer allocated in video memory? -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=9878 --- Comment #23 from Stefan Dösinger <stefandoesinger(a)gmx.at> 2007-10-17 01:08:43 --- Could Steam of VAC mess with the heap data structures? Steam is some sort of copy protection system, and VAC is a cheat protection tool, so both are candidates for messing with things they shouldn't touch. -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=9878 Vitaliy Margolen <vitaliy(a)kievinfo.com> changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |abc.mikey(a)googlemail.com --- Comment #24 from Vitaliy Margolen <vitaliy(a)kievinfo.com> 2007-10-17 08:04:57 --- *** Bug 10070 has been marked as a duplicate of this bug. *** -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=9878 Harry Bullen <hbullen(a)fit.edu> changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |hbullen(a)fit.edu --- Comment #25 from Harry Bullen <hbullen(a)fit.edu> 2007-10-18 17:21:21 --- I have noticed that Bug 9983 appears to be related to this one, but I'm not really sure. -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=9878 --- Comment #26 from Stefan Dösinger <stefandoesinger(a)gmx.at> 2007-10-19 15:19:05 --- (In reply to comment #22) Allan, I didn't read your comment properly before, and I missed the bits about the vertex buffer. WineD3D creates a shadow copy of the vertex buffer on the process Heap, but also allocates an opengl vbo to store the data in video memory. What is the exact size of the problematic buffer requested by hl2? How many bytes does your 16 byte alignment change add? Very old DirectDraw apps(Diablo 1) were found to write 4 extra bytes at the end of surfaces. Maybe we should keep some space free for vertex buffers too. I will do some tests with HL2. If HL2 writes beyond the vertex buffer end, I should be able to see that. -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=9878 --- Comment #27 from Allan Tong <actong88(a)gmail.com> 2007-10-19 22:06:15 --- Created an attachment (id=8670) --> (http://bugs.winehq.org/attachment.cgi?id=8670) alignment patch HL2 is requesting exactly 2MiB = 2097152 bytes. I tried allocating a bigger buffer but that didn't seem to do anything. The attached patch is a hack that hopefully should illustrate what I'm seeing with respect to the alignment. It creates a new heap flag HEAP_TESTALIGN that when passed to HeapAlloc, will cause it to return a buffer aligned to a 16-byte boundary (actually it only does that if it has to create a new subheap, but that should be good enough for this problem). The alignment is done by adjusting the size of the subheap and the subheap header. Actually the simplest patch would be to just add 8 bytes to the SUBHEAP structure in ntdll/heap.c, but the attached patch isolates the problem to the heap allocation in CreateVertexBuffer. In either case, I no longer get the crash in HL2. You can adjust the padding by changing the value of ALIGN_PADDING in ntdll/heap.c. I've found that values of 8, 24, and 40 work, while values of 0, 16, 32, and 48 will cause the crash. The first set all cause the first block in the subheap to be aligned on a 16-byte boundary. -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=9878 --- Comment #28 from Stefan Dösinger <stefandoesinger(a)gmx.at> 2007-10-20 05:58:39 --- Hmm, let me make sure I understand this correctly: When the vertex buffer is HeapAlloced, the memory is directly after the subheap structure, and doesn't start at a multiple of 16 bytes: (1) [ SH ][ BUFFER ][SOMETHINGELSE] | This causes a crash, propably because HL2 overwrites SH or SOMETHINGELSE. Now with the padding, you get the start address to a multiple of 16 bytes: (2) [ SH ][PAD][ BUFFER ][SOMETHINGELSE] | And this works. Now if you pad a bit more, it crashes again (3) [ SH ][ PAD ][ BUFFER ][SOMETHINGELSE] | So since (2) works, but (3) fails, this can't be because HL2 overwrites the subheap header. ie it is really the alignment of the start address, not the distance from some other structure. Some SSE instructions work on 16 byte blocks as far as I know, but do they require 16 byte alignment? But HL2 itself could assume alignment in some internal calculations. -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=9878 --- Comment #29 from Allan Tong <actong88(a)gmail.com> 2007-10-20 09:32:05 --- (In reply to comment #28) Yes, that's what I'm seeing. I went as far as padding with 48 bytes, and it just alternated crashing and not crashing. Regarding SSE, an example is MOVAPD, an aligned version of MOVUPD that requires an address aligned on a 16-byte boundary when accessing memory. -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=9878 --- Comment #30 from Vitaliy Margolen <vitaliy(a)kievinfo.com> 2007-10-20 11:49:36 --- Indeed always making the subheap 8 bytes bigger fixes the problem here. -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=9878 Vitaliy Margolen <vitaliy(a)kievinfo.com> changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |asksteved.com(a)gmail.com --- Comment #31 from Vitaliy Margolen <vitaliy(a)kievinfo.com> 2007-10-20 12:52:02 --- *** Bug 10110 has been marked as a duplicate of this bug. *** -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=9878 --- Comment #32 from Harry Bullen <hbullen(a)fit.edu> 2007-10-22 20:53:18 --- The posted patch fixes the crash for me, on my Linux 2.6.22-gentoo-r8 #1 PREEMPT Mon Oct 15 16:42:49 EST 2007 i686 Intel(R) Pentium(R) 4 CPU 2.53GHz GenuineIntel GNU/Linux. -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=9878 Vitaliy Margolen <vitaliy(a)kievinfo.com> changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |j.lathian(a)gmail.com --- Comment #33 from Vitaliy Margolen <vitaliy(a)kievinfo.com> 2007-10-22 21:23:50 --- *** Bug 10143 has been marked as a duplicate of this bug. *** -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=9878 --- Comment #34 from Stefan Dösinger <stefandoesinger(a)gmx.at> 2007-10-23 04:52:46 --- I have written a test, and Windows uses 32 byte alignment. I've implemented a fix for that which allocates 32 extra bytes and skips the first bytes to get to the alignment boundary. This fixes the bug, and I will send it in the next days. It depends on my surface memory management cleanups, so a few other patches are needed first. -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=9878 Douglas E. Warner <silfreed-winehq(a)silfreed.net> changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |silfreed-winehq(a)silfreed.net -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=9878 Rodrigo Saboya <saboya(a)gmail.com> changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |saboya(a)gmail.com --- Comment #35 from Rodrigo Saboya <saboya(a)gmail.com> 2007-10-25 10:09:28 --- If it helps: With 0.9.46, Portal and TF2 works fine. With 0.9.47 and CVS (tested today @ 3 am) Portal freezes when I open portals and TF2 freezes when I join a team or I get to a certain part of the map in the developer commented maps. -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=9878 Renato Caldas <seventhguardian(a)gmail.com> changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |seventhguardian(a)gmail.com --- Comment #36 from Renato Caldas <seventhguardian(a)gmail.com> 2007-10-26 20:15:21 --- I'm not sure if this helps, but I managed to reproduce my crash "when I want to". I'm running wine cvs (in practice, 0.9.47) on x86_64, nvidia Go 7300, kernel 2.6.23. In dx8, as soon as I try to look towards Alyx the game crashes. I never get to see her in dx8. If I save and restart the game with dx9 I can proceed in the game, and Alyx always has the correct colors (no color bug on her). Then I save the game, looking the other way and restart the it in dx8, it works! Until the time where I look towards her again, then it crashes a few degrees before I would actually see her. I don't have this problem with other npc's. I'm not sure if this has to do with her proximity, some gfx techniques used on her but not on other npc's etc. It may even be some problem not related to this bug.. If this may interest you I can provide more info. If not, I apologize.. -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=9878 --- Comment #37 from Renato Caldas <seventhguardian(a)gmail.com> 2007-10-26 20:25:19 --- (In reply to comment #36)
I'm not sure if this helps, but I managed to reproduce my crash "when I want to". I'm running wine cvs (in practice, 0.9.47) on x86_64, nvidia Go 7300, kernel 2.6.23.
I'm sorry for being too hasty.. The patch actually solves it on the cvs version. -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=9878 Vitaliy Margolen <vitaliy(a)kievinfo.com> changed: What |Removed |Added ---------------------------------------------------------------------------- Component|wine-kernel |wine-directx-d3d --- Comment #38 from Vitaliy Margolen <vitaliy(a)kievinfo.com> 2007-10-26 20:26:28 --- As per Stefan this is a problem with d3d. -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=9878 Michael <baby.lueshi(a)gmail.com> changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |baby.lueshi(a)gmail.com --- Comment #39 from Michael <baby.lueshi(a)gmail.com> 2007-10-28 10:23:02 --- (In reply to comment #35)
If it helps:
With 0.9.46, Portal and TF2 works fine.
With 0.9.47 and CVS (tested today @ 3 am) Portal freezes when I open portals and TF2 freezes when I join a team or I get to a certain part of the map in the developer commented maps.
I've found that setting the Windows version for hl2.exe to Win98 fixes at least the crash opening portals. -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=9878 --- Comment #40 from Douglas E. Warner <silfreed-winehq(a)silfreed.net> 2007-10-30 12:23:04 --- The patch in attachment #8670 allows Half-life 2 to run for me. -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=9878 Vitaliy Margolen <vitaliy(a)kievinfo.com> changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |FIXED --- Comment #41 from Vitaliy Margolen <vitaliy(a)kievinfo.com> 2007-11-02 21:26:00 --- Tested Team Fortress 2 and Episode2. Bot appears to work again. Thanks Stefan for the patches. Fixed in GIT, should be available in 0.9.49 -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=9878 Toby Brommerich <Snub_Fighter(a)snubfighter.com> changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |Snub_Fighter(a)snubfighter.com --- Comment #42 from Toby Brommerich <Snub_Fighter(a)snubfighter.com> 2007-11-06 09:21:28 --- (In reply to comment #34)
I have written a test, and Windows uses 32 byte alignment. I've implemented a fix for that which allocates 32 extra bytes and skips the first bytes to get to the alignment boundary. This fixes the bug, and I will send it in the next days. It depends on my surface memory management cleanups, so a few other patches are needed first.
So I'm having this issue should I just grab the source through git or is there a way to manually apply the patch to 0.9.48.0? -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=9878 --- Comment #43 from Stefan Dösinger <stefandoesinger(a)gmx.at> 2007-11-06 12:29:04 --- You should either grab the git source or wait until friday for .49. The fix depends on a number of other patches, so its not easy to apply it by hand. -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=9878 Toby Brommerich <Snub_Fighter(a)snubfighter.com> changed: What |Removed |Added ---------------------------------------------------------------------------- CC|baby.lueshi(a)gmail.com, | |sileon(a)gmail.com, | |saboya(a)gmail.com, | |abc.mikey(a)googlemail.com, | |stefandoesinger(a)gmx.at, | |hbullen(a)fit.edu, | |asksteved.com(a)gmail.com, | |waninkoko(a)gmail.com, | |Snub_Fighter(a)snubfighter.com| |, silfreed- | |winehq(a)silfreed.net, | |julliard(a)winehq.org | CC|seventhguardian(a)gmail.com, | |j.lathian(a)gmail.com | --- Comment #44 from Toby Brommerich <Snub_Fighter(a)snubfighter.com> 2007-11-06 13:27:24 --- (In reply to comment #43)
You should either grab the git source or wait until friday for .49. The fix depends on a number of other patches, so its not easy to apply it by hand.
Alright thanx for the update, guess I'll make due and wait till friday. -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=9878 Vitaliy Margolen <vitaliy(a)kievinfo.com> changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |baby.lueshi(a)gmail.com, | |saboya(a)gmail.com, | |sileon(a)gmail.com, | |abc.mikey(a)googlemail.com, | |stefandoesinger(a)gmx.at, | |asksteved.com(a)gmail.com, | |hbullen(a)fit.edu, | |waninkoko(a)gmail.com, | |silfreed- | |winehq(a)silfreed.net, | |Snub_Fighter(a)snubfighter.com | |, julliard(a)winehq.org CC| |seventhguardian(a)gmail.com, | |j.lathian(a)gmail.com --- Comment #45 from Vitaliy Margolen <vitaliy(a)kievinfo.com> 2007-11-06 13:42:45 --- Toby Brommerich, why did you removed all the people from the CC? Did anyone asked _YOU_ to do that? -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=9878 --- Comment #46 from Toby Brommerich <Snub_Fighter(a)snubfighter.com> 2007-11-06 15:04:41 --- (In reply to comment #45)
Toby Brommerich, why did you removed all the people from the CC? Did anyone asked _YOU_ to do that?
oh sorry i thought that ment sending out emails for my reply.... :( -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=9878 Dan Kegel <dank(a)kegel.com> changed: What |Removed |Added ---------------------------------------------------------------------------- Status|RESOLVED |CLOSED --- Comment #47 from Dan Kegel <dank(a)kegel.com> 2008-01-28 05:40:37 --- Closing all RESOLVED FIXED bugs older than four weeks. -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email Do not reply to this email, post in Bugzilla using the above URL to reply. ------- You are receiving this mail because: ------- You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=9878 Austin English <austinenglish(a)gmail.com> changed: What |Removed |Added ---------------------------------------------------------------------------- Version|CVS/GIT |unspecified -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email Do not reply to this email, post in Bugzilla using the above URL to reply. ------- You are receiving this mail because: ------- You are watching all bug changes.
participants (1)
-
wine-bugs＠winehq.org