[Bug 56530] New: Final Fantasy XI Online: Memory leak when Wine is built with CFLAGS="-g -mno-avx".
https://bugs.winehq.org/show_bug.cgi?id=56530 Bug ID: 56530 Summary: Final Fantasy XI Online: Memory leak when Wine is built with CFLAGS="-g -mno-avx". Product: Wine Version: 8.13 Hardware: x86-64 URL: https://web.archive.org/web/20210810150839/http://www. playonline.com/ff11eu/download/media/install_win.html OS: Linux Status: UNCONFIRMED Severity: normal Priority: P2 Component: d3d Assignee: wine-bugs(a)winehq.org Reporter: chiitoo(a)gentoo.org Regression SHA1: c065b4fe0b3622e0c737a4c7c1c514273ad2d8a7 Distribution: --- After c065b4fe0b3 [1], a memory leak can be seen in Final Fantasy XI Online, but only when building with CFLAGS="-g" instead of CFLAGS="-march=znver4 -O2 -pipe -mindirect-branch=thunk -mno-avx" ---- err:d3d:wined3d_debug_callback 01DBBCF8: "GL_OUT_OF_MEMORY in glGetTexImage". ---- Nothing too specific needs to be done, aside from simply logging in to the game, and standing still, which will make the address space fill up in a few minutes (some in-game area requirements may be present, or at least some make it happen faster than others). Thank you! 1. https://source.winehq.org/git/wine.git/commitdiff/c065b4fe0b3622e0c737a4c7c1... -- 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.
https://bugs.winehq.org/show_bug.cgi?id=56530 Chiitoo <chiitoo(a)gentoo.org> changed: What |Removed |Added ---------------------------------------------------------------------------- Distribution|--- |Gentoo Keywords| |download, regression -- 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.
https://bugs.winehq.org/show_bug.cgi?id=56530 Zeb Figura <z.figura12(a)gmail.com> changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |z.figura12(a)gmail.com --- Comment #1 from Zeb Figura <z.figura12(a)gmail.com> --- That's extremely weird. Does it not leak memory at all without those flags? Or is it just starting from a lower baseline presumably due to smaller executable sizes (which would make a lot more sense)? -- 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.
https://bugs.winehq.org/show_bug.cgi?id=56530 --- Comment #2 from Chiitoo <chiitoo(a)gentoo.org> --- So far I have still not seen it happen without the '-g' flag even after hours/days of the game running (and need to use '-mno-avx' due to the GCC bug [1] but no idea if it has anything to do with it). Memory seems to get released properly without '-g' (just eye-balling it though; could probably test that better), but enabling it leads to out-of-memory in just a few minutes depending on the in-game area. 1. https://bugs.winehq.org/show_bug.cgi?id=55007 -- 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.
https://bugs.winehq.org/show_bug.cgi?id=56530 --- Comment #3 from Zeb Figura <z.figura12(a)gmail.com> --- (In reply to Chiitoo from comment #2)
So far I have still not seen it happen without the '-g' flag even after hours/days of the game running (and need to use '-mno-avx' due to the GCC bug [1] but no idea if it has anything to do with it).
Memory seems to get released properly without '-g' (just eye-balling it though; could probably test that better), but enabling it leads to out-of-memory in just a few minutes depending on the in-game area.
Ah interesting, -g makes the difference. I wonder if there's a leak in dbghelp then. -- 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.
https://bugs.winehq.org/show_bug.cgi?id=56530 --- Comment #4 from Zeb Figura <z.figura12(a)gmail.com> --- (In reply to Zeb Figura from comment #3)
Ah interesting, -g makes the difference. I wonder if there's a leak in dbghelp then.
That wouldn't explain why c065b4fe0b is to blame, though. Are you sure that's not just bug 55654? -- 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.
https://bugs.winehq.org/show_bug.cgi?id=56530 --- Comment #5 from Chiitoo <chiitoo(a)gentoo.org> --- (In reply to Zeb Figura from comment #4)
(In reply to Zeb Figura from comment #3)
Ah interesting, -g makes the difference. I wonder if there's a leak in dbghelp then.
That wouldn't explain why c065b4fe0b is to blame, though. Are you sure that's not just bug 55654?
I'll try to re-test and make sure during this week. Thanks! -- 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.
https://bugs.winehq.org/show_bug.cgi?id=56530 Chiitoo <chiitoo(a)gentoo.org> changed: What |Removed |Added ---------------------------------------------------------------------------- URL|https://web.archive.org/web |https://web.archive.org/web |/20210810150839/http://www. |/20230801110859/http://dl.s |playonline.com/ff11eu/downl |quare-enix.co.jp/taru/FFXiB |oad/media/install_win.html |3Setup.exe --- Comment #6 from Chiitoo <chiitoo(a)gentoo.org> --- Changing the download URL to the benchmark application which seems to be producing the same results as the actual game. -- 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.
https://bugs.winehq.org/show_bug.cgi?id=56530 --- Comment #7 from Chiitoo <chiitoo(a)gentoo.org> --- Seems like the benchmark applications indeed does trigger the same issue, hopefully, and it's a bunch easier to get into testing without already playing the game (fun account management related things etc.). ^^; -- 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.
https://bugs.winehq.org/show_bug.cgi?id=56530 Chiitoo <chiitoo(a)gentoo.org> changed: What |Removed |Added ---------------------------------------------------------------------------- Resolution|--- |FIXED Status|UNCONFIRMED |RESOLVED Fixed by SHA1| |870d9d3e8a7a1bdb2820d0dc922 | |eed0a668d8a96 --- Comment #8 from Chiitoo <chiitoo(a)gentoo.org> --- While testing for a different issue, I noticed that I could run Final Fantasy XI Online for hours without issues, using a debug Wine build. So... to the bisect-mobile! What might have fixed it!? It seems that the improvements (or at least, slowing the leak(s) down) started at https://gitlab.winehq.org/wine/wine/-/commit/3c1aa7f74adbaf430074003da2a149a... where the benchmark application started taking around 3 minutes and 11 seconds to run out of memory. Right before that though, at https://gitlab.winehq.org/wine/wine/-/commit/5bf3698f8dac4dd4936a3199d73d88e... there was a bit of a dip where it would run out of memory already at around 50 seconds, down from the original 1 minute and 11 seconds or so. These commits would be in the 9.14 release, but it took a bit more until things really settled. At commit https://gitlab.winehq.org/wine/wine/-/commit/11390796260f8fb54b19024f1093848... the run-time is extended to around 4 minutes and 11, seconds, and finally, after https://gitlab.winehq.org/wine/wine/-/commit/870d9d3e8a7a1bdb2820d0dc922eed0... which is included in release 9.15, the virtual memory size seems to have become pretty much completely stable again from what I've seen so far, so I guess we can say that this issue was fixed by 870d9d3e8a7. Thank you! -- 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.
https://bugs.winehq.org/show_bug.cgi?id=56530 --- Comment #9 from Zeb Figura <z.figura12(a)gmail.com> --- That is... even weirder, but if it works, it works... -- 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.
https://bugs.winehq.org/show_bug.cgi?id=56530 Alexandre Julliard <julliard(a)winehq.org> changed: What |Removed |Added ---------------------------------------------------------------------------- Status|RESOLVED |CLOSED --- Comment #10 from Alexandre Julliard <julliard(a)winehq.org> --- Closing bugs fixed in 10.1. -- 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)
-
WineHQ Bugzilla