February 2021 - Wine-Bugs - list.winehq.org

[Bug 36012] New: OllyDbg 2.x debugger crashes with Wine builtin dbghelp
by wine-bugs＠winehq.org Feb. 11, 2021

Feb. 11, 2021

https://bugs.winehq.org/show_bug.cgi?id=36012 Bug ID: 36012 Summary: OllyDbg 2.x debugger crashes with Wine builtin dbghelp Product: Wine Version: 1.7.16 Hardware: x86 OS: Linux Status: NEW Severity: normal Priority: P2 Component: dbghelp Assignee: wine-bugs(a)winehq.org Reporter: focht(a)gmx.net Hello folks, noticed this a while ago ... reporting now :) Ollydbg v2.x crashes for me with Wine builtin dbghelp. Interestingly the old Ollydbg v1.x doesn't crash with builtin. So I ended up with: Ollydbg 1.x debugging Ollydbg 2.x debugging Wine builtin 'notepad' :) Relevant part of tracelog (with some helper traces added by me): --- snip --- $ WINEDEBUG=+tid,+seh,+dbghelp,+dbghelp_symt wine ./ollydbg.exe >>log2.txt 2>&1 ... 0023:trace:dbghelp_symt:symt_new_function Adding global function L"kernel32<elf>":WCEL_InsertChar @7b83839a-7b8383c8 0023:trace:dbghelp_symt:symt_add_func_local Adding local symbol (L"kernel32<elf>":WCEL_InsertChar): ctx 0x20acd44 0023:trace:dbghelp_symt:symt_add_func_local Adding local symbol (L"kernel32<elf>":WCEL_InsertChar): c 0x20aaba0 0023:trace:dbghelp_symt:symt_add_func_local Adding local symbol (L"kernel32<elf>":WCEL_InsertChar): buffer 0x20ad7f0 ... 0023:trace:dbghelp_symt:symt_add_func_line (0x20ae608)WCEL_InsertChar:0 /home/focht/projects/wine/wine.repo/src/dlls/kernel32/editline.c:328 0023:trace:dbghelp_symt:symt_add_func_line (0x20ae608)WCEL_InsertChar:d /home/focht/projects/wine/wine.repo/src/dlls/kernel32/editline.c:331 0023:trace:dbghelp_symt:symt_add_func_line (0x20ae608)WCEL_InsertChar:15 /home/focht/projects/wine/wine.repo/src/dlls/kernel32/editline.c:332 0023:trace:dbghelp_symt:symt_add_func_line (0x20ae608)WCEL_InsertChar:1b /home/focht/projects/wine/wine.repo/src/dlls/kernel32/editline.c:333 0023:trace:dbghelp_symt:symt_add_func_line (0x20ae608)WCEL_InsertChar:2d /home/focht/projects/wine/wine.repo/src/dlls/kernel32/editline.c:334 ... 0023:trace:dbghelp_symt:symt_new_public Adding public symbol L"kernel32<elf>":WCEL_InsertChar @7b83839a ... 0023:trace:dbghelp:SymMatchStringW L"WCEL_InsertChar" L"*" N 0023:trace:dbghelp_symt:symt_fill_sym_info 0x20ae608 => WCEL_InsertChar 47 7b83839a 0023:trace:dbghelp:send_symbol se->cb=0x7d7f4fd1 se->sym_info=0x335fd8, se->sym_info->Size=47 se->user=0x33683c 0023:trace:dbghelp:sym_enumW sew->cb=0x4c90bc sew->sym_info=0x336848 size=47 sew->ctx=0x337098 0023:trace:dbghelp:SymSetContext Setting same frame {rtn=0 frm=0 stk=0} 0023:trace:dbghelp:symt_enum_locals_helper pair=0x3312fc match=0x7d800bd2 se=0x331398 func=0x20ae608 v=0x20ae63c vlen=4 0023:trace:dbghelp:symt_enum_locals_helper i=0 lsym=0x20ae66c, lsym->tag=0x7 0023:trace:dbghelp:SymMatchStringW L"ctx" L"*" Y 0023:trace:dbghelp_symt:symt_fill_sym_info 0x20ae66c => ctx 4 4 0023:trace:dbghelp:send_symbol se->cb=0x7d7f4fd1 se->sym_info=0x3313b8, se->sym_info->Size=4 se->user=0x331c1c 0023:trace:dbghelp:sym_enumW sew->cb=0x4c8f54 sew->sym_info=0x331c28 size=4 sew->ctx=0x33250c 0023:trace:dbghelp:symt_enum_locals_helper i=1 lsym=0x20ae6e4, lsym->tag=0x7 0023:trace:dbghelp:SymMatchStringW L"c" L"*" Y 0023:trace:dbghelp_symt:symt_fill_sym_info 0x20ae6e4 => c 2 ffffffe0 0023:trace:dbghelp:send_symbol se->cb=0x7d7f4fd1 se->sym_info=0x3313b8, se->sym_info->Size=2 se->user=0x331c1c 0023:trace:dbghelp:sym_enumW sew->cb=0x4c8f54 sew->sym_info=0x331c28 size=2 sew->ctx=0x33250c 0023:trace:seh:raise_exception code=c0000005 flags=0 addr=0x7d7f1208 ip=7d7f1208 tid=0023 0023:trace:seh:raise_exception info[0]=00000000 0023:trace:seh:raise_exception info[1]=00000064 0023:trace:seh:raise_exception eax=00000058 ebx=7d80d000 ecx=003311f0 edx=fffffd6f esi=00000040 edi=0033689c 0023:trace:seh:raise_exception ebp=00331248 esp=00331248 cs=0023 ds=002b es=002b fs=0063 gs=006b flags=00010202 0023:trace:seh:call_stack_handlers calling handler at 0x4e87c0 code=c0000005 flags=0 --- snip --- 'WCEL_InsertChar' seems to be the offender. Populating the symbol info for that function causes the crash in the debugger callback. http://source.winehq.org/git/wine.git/blob/fc38254fed5dbc07a2a61a2411bfaadc… --- snip --- 327 static void WCEL_InsertChar(WCEL_Context* ctx, WCHAR c) 328 { 329 WCHAR buffer[2]; 330 331 buffer[0] = c; 332 buffer[1] = 0; 333 WCEL_InsertString(ctx, buffer); 334 } --- snip --- During symbol evaluation of second function parameter 'c', the debugger syminfo callback corrupts parts of the caller stack hence the next for-loop iteration 'vector_length(v)' which ought to prepare 'buffer' local symbol info crashes. Source: http://source.winehq.org/git/wine.git/blob/d0665616255ae8b720e14e02072996b7… --- snip --- 909 static BOOL symt_enum_locals_helper(struct module_pair* pair, 910 const WCHAR* match, const struct sym_enum* se, 911 struct symt_function* func, const struct vector* v) 912 { 913 struct symt* lsym = NULL; 914 DWORD pc = pair->pcs->ctx_frame.InstructionOffset; 915 unsigned int i; 916 WCHAR* nameW; 917 BOOL ret; 918 919 for (i=0; i<vector_length(v); i++) 920 { 921 lsym = *(struct symt**)vector_at(v, i); 922 switch (lsym->tag) 923 { ... 933 case SymTagData: 934 nameW = symt_get_nameW(lsym); 935 ret = SymMatchStringW(nameW, match, 936 !(dbghelp_options & SYMOPT_CASE_INSENSITIVE)); 937 HeapFree(GetProcessHeap(), 0, nameW); 938 if (ret) 939 { 940 if (send_symbol(se, pair, func, lsym)) return FALSE; 941 } 942 break; ... 951 } 952 } 953 return TRUE; 954 } --- snip --- Dump of symbol info for second parameter 'c': --- snip --- Wine-dbg>p *se.sym_info {SizeOfStruct=0x58, TypeIndex=0x20a2b90, Reserved={0, 0}, info=0x20a66d4, Size=0x2, ModBase=0x7b810000, Flags=0xd0, Value=0, Address=0xffffffe0, Register=0x15, Scope=0, Tag=0x7, NameLen=0x1, MaxNameLen=0x7d0, Name=""} --- snip --- Flags: 0xD0 = SYMFLAG_LOCAL | SYMFLAG_PARAMETER | SYMFLAG_REGREL Register: 0x15 = CV_REG_ESP Disassembly of 'WCEL_InsertChar': --- snip --- $ objdump -d editline.o 00000cba <WCEL_InsertChar>: cba: 55 push %ebp cbb: 89 e5 mov %esp,%ebp cbd: 83 ec 28 sub $0x28,%esp cc0: 8b 45 0c mov 0xc(%ebp),%eax ; parameter 'c' cc3: 66 89 45 e4 mov %ax,-0x1c(%ebp) ; 'c' as stack variable cc7: 0f b7 45 e4 movzwl -0x1c(%ebp),%eax ccb: 66 89 45 f4 mov %ax,-0xc(%ebp) ; buffer[0] = c ccf: 66 c7 45 f6 00 00 movw $0x0,-0xa(%ebp) ; buffer[1] = 0 cd5: 8d 45 f4 lea -0xc(%ebp),%eax cd8: 89 44 24 04 mov %eax,0x4(%esp) cdc: 8b 45 08 mov 0x8(%ebp),%eax ; 'ctx' parameter passthru cdf: 89 04 24 mov %eax,(%esp) ce2: e8 30 fe ff ff call b17 <WCEL_InsertString> ce7: c9 leave ce8: c3 ret --- snip --- DWARF(2) symbol info: --- snip --- $ objdump -W editline.o <1><cf7>: Abbrev Number: 43 (DW_TAG_subprogram) <cf8> DW_AT_name : (indirect string, offset: 0x25b): WCEL_InsertChar <cfc> DW_AT_decl_file : 3 <cfd> DW_AT_decl_line : 327 <cff> DW_AT_prototyped : 1 <d00> DW_AT_low_pc : 0xcba <d04> DW_AT_high_pc : 0xce9 <d08> DW_AT_frame_base : 0x3b8 (location list) <d0c> DW_AT_sibling : <0xd3c> <2><d10>: Abbrev Number: 27 (DW_TAG_formal_parameter) <d11> DW_AT_name : ctx <d15> DW_AT_decl_file : 3 <d16> DW_AT_decl_line : 327 <d18> DW_AT_type : <0x847> <d1c> DW_AT_location : 2 byte block: 91 0 (DW_OP_fbreg: 0) <2><d1f>: Abbrev Number: 27 (DW_TAG_formal_parameter) <d20> DW_AT_name : c <d22> DW_AT_decl_file : 3 <d23> DW_AT_decl_line : 327 <d25> DW_AT_type : <0xf7> <d29> DW_AT_location : 2 byte block: 91 5c (DW_OP_fbreg: -36) <2><d2c>: Abbrev Number: 44 (DW_TAG_variable) <d2d> DW_AT_name : (indirect string, offset: 0x67a): buffer <d31> DW_AT_decl_file : 3 <d32> DW_AT_decl_line : 329 <d34> DW_AT_type : <0xa2a> <d38> DW_AT_location : 2 byte block: 91 6c (DW_OP_fbreg: -20) --- snip --- DW_AT_frame_base = 0x3b8 -> Location list for 0x3b8 --- snip --- $ objdump --dwarf=loc editline.o editline.o: file format elf32-i386 Contents of the .debug_loc section: ... 000003b8 00000cba 00000cbb (DW_OP_breg4 (esp): 4) 000003c4 00000cbb 00000cbd (DW_OP_breg4 (esp): 8) 000003d0 00000cbd 00000ce8 (DW_OP_breg5 (ebp): 8) 000003dc 00000ce8 00000ce9 (DW_OP_breg4 (esp): 4) 000003e8 <End of list> ... --- snip --- What might be the culprit here (see disassembly) is that 'c' is actually both: a function parameter ('DW_TAG_formal_parameter' -> 'cc0:') and a local variable ('cc3:'). The DWARF symbol info contains only one entry. Maybe that's something that confuses the debugger (function parameter handler combined with negative addressing mode which ought to apply for local variable). $ sha1sum odbg201.zip d41fe77a2801d38476f20468ab61ddce14c3abb8 odbg201.zip $ du -sh odbg201.zip 6.7M odbg201.zip $ wine --version wine-1.7.16-178-g7e874ae Regards -- 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 7560] OllyDbg 1.1 fails to debug multithreaded apps
by WineHQ Bugzilla Feb. 11, 2021

Feb. 11, 2021

1 0

[Bug 20230] New: GetThreadTimes() and the Actual and Pseudo Current Thread Handles
by wine-bugs＠winehq.org Feb. 11, 2021

Feb. 11, 2021

http://bugs.winehq.org/show_bug.cgi?id=20230 Summary: GetThreadTimes() and the Actual and Pseudo Current Thread Handles Product: Wine Version: 1.1.30 Platform: PC URL: http://rh-software.com OS/Version: Linux Status: UNCONFIRMED Severity: normal Priority: P2 Component: ntdll AssignedTo: wine-bugs(a)winehq.org ReportedBy: ray(a)pobox.co.uk Some time ago I had an issue with SIV being unable to read the thread CPU time usage when using GetThreadTimes(). I tracked down to the following and then did a work-a-round. The "Current Thread" effectively has 2 current thread handles, these are the actual thread handle and the ~0 pseudo handle as returned by GetCurrentThread(). When a call is made to GetThreadTimes() this calls NtQueryInformationThread( ThreadTimes ) which contains the test "if (handle == GetCurrentThread())". As a result of this when GetThreadTimes() is called with the actual thread handle no data is returned. \WINE-1.1.30\dlls\ntdll\thread.c(1011): if (handle == GetCurrentThread()) I suspect the code should be changed to allow both the pseudo and actual thread handles to be used as a minimum, better still would, when possible, be to implement this function for any thread. While searching the source I also noticed: NTSTATUS WINAPI NtSetInformationThread( HANDLE handle, THREADINFOCLASS class, LPCVOID data, ULONG length ) { NTSTATUS status; switch(class) { case ThreadZeroTlsCell: if (handle == GetCurrentThread()) And suspect that code would benefit from a similar fix. -- 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 18

[Bug 48407] New: OllyDbg 2.x segfaults the process after attaching to it
by WineHQ Bugzilla Feb. 11, 2021

Feb. 11, 2021

https://bugs.winehq.org/show_bug.cgi?id=48407 Bug ID: 48407 Summary: OllyDbg 2.x segfaults the process after attaching to it Product: Wine Version: 5.0-rc3 Hardware: x86 URL: http://www.ollydbg.de/odbg201.zip OS: Linux Status: NEW Keywords: regression Severity: normal Priority: P2 Component: -unknown Assignee: wine-bugs(a)winehq.org Reporter: gabrielopcode(a)gmail.com CC: jacek(a)codeweavers.com Regression SHA1: 4ee629a3bafb1408a4e567908fef949837a39f10 Distribution: --- Since commit 4ee629a3bafb1408a4e567908fef949837a39f10, OllyDbg will hang after attaching to a process while the process itself will crash, and the message `Segmentation fault' is printed in the terminal where the process is launched from (not OllyDbg). How to reproduce after downloading OllyDbg (link provided in report) using a 32-bit prefix: 1) Launch a simple window app, such as `winemine' 2) From another terminal, launch OllyDbg in same prefix. Optionally: To speed up the attaching in OllyDbg, go to Options->Options->Analysis. In `Automatic Module Analysis' set it to `Off'. 3) In OllyDbg, go to File->Attach and select the process (winemine). Wait a few seconds until modules are processed, then the process will segfault and OllyDbg will hang. I tried to debug this to no avail, it's very unfamiliar territory for me, so it's a bit over my head. Reverting that commit on current wine git is not easy and I don't know how to do it, since the break_process and related functions have been removed from the wineserver at some point. So unfortunately I don't know where to start to fix this 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.

1 8

[Bug 23431] New: provide a stub for ntdll.NtSystemDebugControl() (Ollydbg 2.x)
by wine-bugs＠winehq.org Feb. 11, 2021

Feb. 11, 2021

http://bugs.winehq.org/show_bug.cgi?id=23431 Summary: provide a stub for ntdll.NtSystemDebugControl() (Ollydbg 2.x) Product: Wine Version: 1.2-rc5 Platform: x86 URL: http://www.ollydbg.de/version2.html OS/Version: Linux Status: NEW Severity: normal Priority: P2 Component: ntdll AssignedTo: wine-bugs(a)winehq.org ReportedBy: focht(a)gmx.net Hello, a simple one... Ollydbg 2.x uses NtSystemDebugControl( DebugSysReadVirtual, ...) to read kernel memory (driver/OS address space) for informational purposes. Failure to do so doesn't cause harm. Although this approach doesn’t require any additional kernel driver (only the SeDebugPrivilege privilege on the running process) it's only useful for Windows XP systems. It's disallowed since Windows Vista/Windows 7. Not really useful in Wine but it obviously needs a stub. --- snip --- wine: Call from 0x7bc3d2d0 to unimplemented function ntdll.dll.NtSystemDebugControl, aborting First chance exception: unimplemented function ntdll.dll.NtSystemDebugControl called in 32-bit code (0x7bc3d2d0). ... =>0 0x7bc3d2d0 __wine_spec_unimplemented_stub+0x50(module="ntdll.dll", function="NtSystemDebugControl") [/opt/wine/wine-build/dlls/ntdll/../../../wine-git/dlls/ntdll/exception.c:268] in ntdll (0x00334434) 1 0x7bc1f895 __wine_stub_NtSystemDebugControl+0x28() in ntdll (0x00334434) 2 0x004467ae in ollydbg (+0x467ad) (0x00334460) 3 0x0044698b in ollydbg (+0x4698a) (0x00334484) --- snip --- Dump params and return STATUS_NOT_IMPLEMENTED to keep it running ;-) Regards -- 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 6

[Bug 27222] New: dbghelp.dll.SymEnumSourceFilesW needed (ollydbg2)
by wine-bugs＠winehq.org Feb. 11, 2021

Feb. 11, 2021

http://bugs.winehq.org/show_bug.cgi?id=27222 Summary: dbghelp.dll.SymEnumSourceFilesW needed (ollydbg2) Product: Wine Version: 1.3.20 Platform: x86 OS/Version: Linux Status: NEW Severity: normal Priority: P2 Component: dbghelp AssignedTo: wine-bugs(a)winehq.org ReportedBy: focht(a)gmx.net Hello, while attaching to a target using JIT (replacing winedbg), the following crash occurred: --- snip --- wine: Call from 0x7b838aa3 to unimplemented function dbghelp.dll.SymEnumSourceFilesW, aborting --- snip --- A stub keeps it from crashing (though it should not be hard to implement a wide character version). +tid,+dbghelp: --- snip --- ... fixme:dbghelp:SymEnumSourceFilesW 0x58 400000 (null) 0x4c06d8 0x400000: stub! fixme:dbghelp:SymEnumSourceFilesW 0x58 68350000 (null) 0x4c06d8 0x68350000: stub! fixme:dbghelp:SymEnumSourceFilesW 0x58 68490000 (null) 0x4c06d8 0x68490000: stub! fixme:dbghelp:SymEnumSourceFilesW 0x58 68530000 (null) 0x4c06d8 0x68530000: stub! fixme:dbghelp:SymEnumSourceFilesW 0x58 68590000 (null) 0x4c06d8 0x68590000: stub! fixme:dbghelp:SymEnumSourceFilesW 0x58 685b0000 (null) 0x4c06d8 0x685b0000: stub! fixme:dbghelp:SymEnumSourceFilesW 0x58 685f0000 (null) 0x4c06d8 0x685f0000: stub! ... --- snip --- Regards -- 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 8

[Bug 27802] New: ollydbg2 crashes on unimpl dbghelp.dll.SymEnumSourceLinesW stub
by wine-bugs＠winehq.org Feb. 11, 2021

Feb. 11, 2021

http://bugs.winehq.org/show_bug.cgi?id=27802 Summary: ollydbg2 crashes on unimpl dbghelp.dll.SymEnumSourceLinesW stub Product: Wine Version: 1.3.24 Platform: x86 OS/Version: Linux Status: NEW Severity: normal Priority: P2 Component: dbghelp AssignedTo: wine-bugs(a)winehq.org ReportedBy: focht(a)gmx.net Hello, ollydbg2 crashes when trying to display source: --- snip --- ... wine: Call from 0x7b8384bb to unimplemented function dbghelp.dll.SymEnumSourceLinesW, aborting --- snip --- Stop at process main (dlls along with their debug syms loaded). "View" -> "Source Files" -> double click any source file $ wine --version wine-1.3.24-174-g4b4dd30 Regards -- 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 7

[Bug 39672] New: Folder names not clickable
by wine-bugs＠winehq.org Feb. 11, 2021

Feb. 11, 2021

https://bugs.winehq.org/show_bug.cgi?id=39672 Bug ID: 39672 Summary: Folder names not clickable Product: Wine Version: 1.7.55 Hardware: x86-64 OS: Linux Status: UNCONFIRMED Severity: minor Priority: P2 Component: -unknown Assignee: wine-bugs(a)winehq.org Reporter: strife(a)home.se Distribution: --- Created attachment 52892 --> https://bugs.winehq.org/attachment.cgi?id=52892 clicking multiple folder texts When opening a file, in Rhino3d, nothing happens when you click the names of the folders, no files shown or updated. In order to show files you must click the folder icon for anything to update. Worked in 1.7.4 -- 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 6

[Bug 39635] New: Hungarian Excel Viewer 2007: "Open" dialog on Win8 does not refresh correctly
by wine-bugs＠winehq.org Feb. 11, 2021

Feb. 11, 2021

https://bugs.winehq.org/show_bug.cgi?id=39635 Bug ID: 39635 Summary: Hungarian Excel Viewer 2007: "Open" dialog on Win8 does not refresh correctly Product: Wine Version: 1.8-rc1 Hardware: x86 OS: Linux Status: NEW Severity: normal Priority: P2 Component: -unknown Assignee: wine-bugs(a)winehq.org Reporter: wylda(a)volny.cz Distribution: --- Continuation of bug 33736. When you set Win version to 8 and then you want to open an XLS file in a folder, you notice, that the existing file (C:\Users\text.xls) is not listed, so there is nothing to open. The problem is, that the folder content does not refresh, even if you right click and select refresh from context menu. The trick is, that in File dialog you can't click on "User" link, but you have to click on the folder-icon standing on the left. It has nothing to do, with filtering (*.xls etc.). For example you can click on _icon_ of "Program files", then all the folders are listed. When you then click on "Users" _link_, it still shows content of "Program Files" instead of "Users". -- 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 6

[Bug 45920] New: Crossfire EU - program error at start
by wine-bugs＠winehq.org Feb. 10, 2021

Feb. 10, 2021

https://bugs.winehq.org/show_bug.cgi?id=45920 Bug ID: 45920 Summary: Crossfire EU - program error at start Product: Wine Version: unspecified Hardware: x86 OS: Linux Status: UNCONFIRMED Severity: major Priority: P2 Component: -unknown Assignee: wine-bugs(a)winehq.org Reporter: luca.finizio.mgbx(a)hotmail.it Distribution: --- Created attachment 62447 --> https://bugs.winehq.org/attachment.cgi?id=62447 detailed information When I run the program I get a Program Error. You can download the program here: http://crossfire.gamerage.com/EN/Download/Client/ -- 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 7