March 2023 - Wine-bugs - list.winehq.org

[Bug 43224] New: Freelist scan can result in O(n) time when allocating
by wine-bugs＠winehq.org 03 Mar '23

03 Mar '23

https://bugs.winehq.org/show_bug.cgi?id=43224 Bug ID: 43224 Summary: Freelist scan can result in O(n) time when allocating Product: Wine Version: unspecified Hardware: x86 OS: Linux Status: UNCONFIRMED Severity: normal Priority: P2 Component: ntdll Assignee: wine-bugs(a)winehq.org Reporter: ranma42(a)gmail.com Distribution: --- Created attachment 58517 --> https://bugs.winehq.org/attachment.cgi?id=58517 Micro-benchmark that shows the degenerate behaviour The heap implementation in ntdll can be very inefficient for some allocation patterns. Specifically, it is possible that a freelist contains thousands of entries of a given size and it is scanned while looking for a bigger entry, that will not fit in any of those. This is demonstrated by the attached program. It accepts a size as an argument and will allocate 1000000 elements of that size, deallocate half of them (the "even" ones), then allocate 1000 elements that are just one bit bigger. For most sizes the program terminates in milliseconds, but in my environment it takes several seconds for sizes 16, 32, 48, 64, 80, 88, 96, 112, 120, 128. A possible workaround (that I had implemented in the past, for short-lived applications) is to allocate from a freelist whose entries are all at least as big as the block that is being allocated, but that can cause inefficient memory usage. I noticed that wine includes an implementation of a red-black tree. Would it make sense to organise the free entries as lists in such a tree? That should ensure O(log N) in the worst case. If performance is a concern, it might be possible to use "direct" freelists for small sizes (ensuring that all of the alignment-compatible sizes are covered) and use a tree for bigger sizes. I am willing to work on this, but I need some directions on what is the best path forward (namely, if we should prefer the memory-inefficient but simpler approach of just allocating from "bigger" freelists or if the tree approach is the desired one). The bug https://bugs.winehq.org/show_bug.cgi?id=37773 might be related. -- 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.

2 9

[Bug 54534] New: dbghelp:dbghelp - The test_loaded_modules() enumeration fails on Windows 10 1607
by WineHQ Bugzilla 03 Mar '23

03 Mar '23

https://bugs.winehq.org/show_bug.cgi?id=54534 Bug ID: 54534 Summary: dbghelp:dbghelp - The test_loaded_modules() enumeration fails on Windows 10 1607 Product: Wine Version: unspecified Hardware: x86-64 OS: Windows Status: NEW Severity: normal Priority: P2 Component: dbghelp Assignee: wine-bugs(a)winehq.org Reporter: fgouget(a)codeweavers.com dbghelp:dbghelp - The test_loaded_modules() enumeration fails on Windows 10 1607: dbghelp.c:517: Test failed: SymLoadModuleExW failed on C:\WINDOWS\System32\bcryptPrimitives.dll: 0 See https://test.winehq.org/data/patterns.html#dbghelp:dbghelp This is the only Windows 10 version which fails. A bisect confirms that this failure was introduced by the commit below: 8be5c10ff893a22330ad94ae18dfb5fa497b09fd Author: Eric Pouech <eric.pouech(a)gmail.com> AuthorDate: Thu Feb 16 10:48:53 2023 +0100 dbghelp/tests: Add tests for 'module' name in EnumLoadedModules() callback. Signed-off-by: Eric Pouech <eric.pouech(a)gmail.com> -- 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.

1 4

[Bug 54584] New: kernel32:locale - The NtGetNlsSectionPtr() test fails on Windows 11
by WineHQ Bugzilla 03 Mar '23

03 Mar '23

https://bugs.winehq.org/show_bug.cgi?id=54584 Bug ID: 54584 Summary: kernel32:locale - The NtGetNlsSectionPtr() test fails on Windows 11 Product: Wine Version: unspecified Hardware: x86-64 OS: Windows Status: NEW Severity: normal Priority: P2 Component: kernel32 Assignee: wine-bugs(a)winehq.org Reporter: fgouget(a)codeweavers.com kernel32:locale - The NtGetNlsSectionPtr() test fails on Windows 11: locale.c:4978: Test failed: 14: failed 0 See https://test.winehq.org/data/patterns.html#kernel32:locale This means Windows 11 supports a new type of NLS section (14). Also NtGetNlsSectionPtr() returns success so presumably that section type does not take an id parameter. -- 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.

1 4

[Bug 54583] New: kernel32:locale - The non-breaking space GetNumberFormatEx() test fails on Windows 11
by WineHQ Bugzilla 03 Mar '23

03 Mar '23

https://bugs.winehq.org/show_bug.cgi?id=54583 Bug ID: 54583 Summary: kernel32:locale - The non-breaking space GetNumberFormatEx() test fails on Windows 11 Product: Wine Version: unspecified Hardware: x86-64 OS: Windows Status: NEW Severity: normal Priority: P2 Component: kernel32 Assignee: wine-bugs(a)winehq.org Reporter: fgouget(a)codeweavers.com kernel32:locale - The non-breaking space GetNumberFormatEx() test fails on Windows 11: locale.c:1820: Test failed: Expected L"-12\00a0345,00", got L"-12\202f345,00" See https://test.winehq.org/data/patterns.html#kernel32:locale So Windows 11 returns a narrow non-breaking space instead of a 'plain' non-breaking space separator when using the French format. The narrow non-breaking space is probably more correct. -- 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.

1 4

[Bug 24256] New: 3D Sexvilla 2: extremely long loading times
by wine-bugs＠winehq.org 03 Mar '23

03 Mar '23

http://bugs.winehq.org/show_bug.cgi?id=24256 Summary: 3D Sexvilla 2: extremely long loading times Product: Wine Version: 1.2 Platform: x86 URL: http://www.3d-sexgames.com/ OS/Version: Linux Status: UNCONFIRMED Severity: normal Priority: P2 Component: ntdll AssignedTo: wine-bugs(a)winehq.org ReportedBy: mailme667(a)yahoo.co.uk Loading time makes the game unusable (version 2.093). I just watched the loading screen for half an hour. But I know it works, because I tried starting the game, turning the monitor off, having dinner and see the game running when I came back a long time later. Clean .wine tested. Other users in appdb have reported the same problem, it's not just me. Bug is tested to be present on wine 1.2 and 1.3.1. sysprof makes it clear ntdll is consuming nearly all the time. Using wine 1.2 the following is seen repeating over and over with winedebug during loading: 002e:Call KERNEL32.GetTickCount() ret=7da7655d 002e:Ret KERNEL32.GetTickCount() retval=00032601 ret=7da7655d 002e:Call KERNEL32.GetTickCount() ret=7e7997bb 002e:Ret KERNEL32.GetTickCount() retval=00032601 ret=7e7997bb 002e:Call KERNEL32.GetTickCount() ret=7e7997bb 002e:Ret KERNEL32.GetTickCount() retval=00032607 ret=7e7997bb 002e:Call KERNEL32.GetTickCount() ret=7da76415 002e:Ret KERNEL32.GetTickCount() retval=00032607 ret=7da76415 002e:Call ntdll.RtlAcquireResourceShared(00187fa4,00000001) ret=7da76541 002e:Ret ntdll.RtlAcquireResourceShared() retval=00000001 ret=7da76541 002e:Call ntdll.RtlReleaseResource(00187fa4) ret=7da76555 002e:Ret ntdll.RtlReleaseResource() retval=00000000 ret=7da76555 I got the following from wine-dbg also using 1.2: Wine-dbg>bt 0x0000002e Backtrace: =>0 0xb7705430 (0x04e0e9d8) 1 0x7e78ea4a TIME_MMSysTimeThread+0x329(arg=0x4e0ea34) [/build/buildd/wine1.2-1.2/dlls/winmm/time.c:218] in winmm (0x04e0ea68) 2 0x7e78ea4a TIME_MMSysTimeThread+0x329(arg=0x7e770000) [/build/buildd/wine1.2-1.2/dlls/winmm/time.c:218] in winmm (0x04e0ea78) 3 0x7bc6f8f0 call_thread_func+0xb() in ntdll (0x04e0eb48) 4 0x7bc6fac0 call_thread_entry_point+0x6f(entry=0x7e78e720, arg=0x7e770000) [/build/buildd/wine1.2-1.2/dlls/ntdll/signal_i386.c:2473] in ntdll (0x04e0f398) 5 0x7bc780b5 start_thread+0xf4(info=0x7ff9cfb8) [/build/buildd/wine1.2-1.2/dlls/ntdll/thread.c:399] in ntdll (0x04e0f498) 6 0xb757780e start_thread+0xbd() in libpthread.so.0 (0x00000000) I hope I've provided enough information. -- 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.

2 39

[Bug 54338] New: Swift crashes due unimplemented QueryUnbiasedInterruptTimePrecise function
by WineHQ Bugzilla 03 Mar '23

03 Mar '23

https://bugs.winehq.org/show_bug.cgi?id=54338 Bug ID: 54338 Summary: Swift crashes due unimplemented QueryUnbiasedInterruptTimePrecise function Product: Wine Version: 8.0-rc4 Hardware: x86-64 OS: Linux Status: UNCONFIRMED Severity: normal Priority: P2 Component: -unknown Assignee: wine-bugs(a)winehq.org Reporter: mikrutrafal(a)protonmail.com Distribution: --- Swift language after installing, crashes with this message ``` wine: Call from 000000017002FA08 to unimplemented function api-ms-win-core-realtime-l1-1-1.dll.QueryUnbiasedInterruptTimePrecise, aborting wine: Unimplemented function api-ms-win-core-realtime-l1-1-1.dll.QueryUnbiasedInterruptTimePrecise called at address 000000017002FA08 (thread 0110), starting debugger... ``` Steps to reproduce - install Swift 5.7.2 - https://download.swift.org/swift-5.7.2-release/windows10/swift-5.7.2-RELEAS… - execute `wine swift` -- 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.

1 5

[Bug 53504] New: Sacred:unhandled exception in Wine 7.14
by WineHQ Bugzilla 03 Mar '23

03 Mar '23

https://bugs.winehq.org/show_bug.cgi?id=53504 Bug ID: 53504 Summary: Sacred:unhandled exception in Wine 7.14 Product: Wine Version: 7.14 Hardware: x86-64 OS: Linux Status: UNCONFIRMED Severity: normal Priority: P2 Component: -unknown Assignee: wine-bugs(a)winehq.org Reporter: chebanenkoigor93(a)gmail.com Distribution: --- Created attachment 72868 --> https://bugs.winehq.org/attachment.cgi?id=72868 screenshot Wine 7.14 gives me exception after launching Sacred. -- 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.

1 4

[Bug 54582] New: kernel32:locale - test_NLSVersion() fails on Windows 10 22H2
by WineHQ Bugzilla 03 Mar '23

03 Mar '23

https://bugs.winehq.org/show_bug.cgi?id=54582 Bug ID: 54582 Summary: kernel32:locale - test_NLSVersion() fails on Windows 10 22H2 Product: Wine Version: unspecified Hardware: x86-64 OS: Windows Status: NEW Severity: normal Priority: P2 Component: kernel32 Assignee: wine-bugs(a)winehq.org Reporter: fgouget(a)codeweavers.com kernel32:locale - test_NLSVersion() fails on Windows 10 22H2: locale.c:7871: Test failed: IsValidNLSVersion succeeded See https://test.winehq.org/data/patterns.html#kernel32:locale Apparently the point of this set of test is to verify that IsValidNLSVersion() rejects any version different from that returned by GetNLSVersion(). But sometimes that's not the case. Indeed a similar failure happened on Windows 10 2004 back in the day, see bug 51159. In that case IsValidNLSVersion() accepted versions 0x100 lower than the current one, presumably because that version of Windows was keeping compatibility with older NLS versions. That one was fixed by making bigger changes to the version. In this case Windows 10 22H2 accepts NLS versions that are 0x100 greater than the current one. So presumably we can just use a larger increment to get the expected failure. -- 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.

1 5

[Bug 51259] New: 6.0.1 Introduces error causing Wavelab to close when loading presets
by WineHQ Bugzilla 03 Mar '23

03 Mar '23

https://bugs.winehq.org/show_bug.cgi?id=51259 Bug ID: 51259 Summary: 6.0.1 Introduces error causing Wavelab to close when loading presets Product: Wine Version: 6.0.1 Hardware: x86-64 OS: Linux Status: UNCONFIRMED Severity: normal Priority: P2 Component: -unknown Assignee: wine-bugs(a)winehq.org Reporter: shagooserver(a)gmail.com Distribution: --- Update 6.0.1 (7/6/21) introduces an error whereby Wavelab 6 closes/crashes when a preset is loaded (into the master section). The terminal outputs: X Error of failed request: BadValue (integer parameter out of range for operation) Major opcode of failed request: 151 (GLX) Minor opcode of failed request: 3 (X_GLXCreateContext) Value in failed request: 0x0 Serial number of failed request: 6199 Current serial number in output stream: 6200 -- 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.

1 3

[Bug 53094] New: ntdll:rtlstr test crashes on win32 arch with hi-IN locale
by WineHQ Bugzilla 03 Mar '23

03 Mar '23

https://bugs.winehq.org/show_bug.cgi?id=53094 Bug ID: 53094 Summary: ntdll:rtlstr test crashes on win32 arch with hi-IN locale Product: Wine Version: unspecified Hardware: x86-64 OS: Linux Status: NEW Severity: normal Priority: P2 Component: ntdll Assignee: wine-bugs(a)winehq.org Reporter: rbernon(a)codeweavers.com Regression SHA1: d8c973ad95ba5e8a9a51df0dd9be587950179ec3 Distribution: --- As seen on the testbot nightly runs there for instance: https://test.winehq.org/data/c1e793f1119de0c0ef7d4bd6d9fefbafdb5dbbe5/linux… Locally reproducible with: LC_ALL=hi-IN.UTF-8 LC_LANG=hi-IN.UTF-8 WINEARCH=win32 WINEPREFIX=$PWD/pfx wine32/wine wine32/dlls/ntdll/tests/ntdll_test.exe rtlstr Bisected to: commit d8c973ad95ba5e8a9a51df0dd9be587950179ec3 Author: Alexandre Julliard <julliard(a)winehq.org> Date: Mon May 30 22:38:28 2022 +0200 kernelbase: Reimplement CompareStringEx using the sortkey generation code. Signed-off-by: Alexandre Julliard <julliard(a)winehq.org> dlls/kernel32/tests/locale.c | 117 +++++++++----------- dlls/kernelbase/locale.c | 247 +++++++++++++++++-------------------------- 2 files changed, 147 insertions(+), 217 deletions(-) -- 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.

1 2