http://bugs.winehq.org/show_bug.cgi?id=14797
Juan Lang juan_lang@yahoo.com changed:
What |Removed |Added ---------------------------------------------------------------------------- Component|secur32 |-unknown
--- Comment #17 from Juan Lang juan_lang@yahoo.com 2009-08-07 18:14:45 --- Based on the logs, and the fact that this is using Peter Gutman's cryptlib, I'm going to say this isn't a secur32 bug. The log is mostly lines like the following:
fixme:ntdll:NtQuerySystemInformation info_class SYSTEM_PERFORMANCE_INFORMATION fixme:ntdll:NtQuerySystemInformation (0x00000004,0x1407fc0,0x00008000,0x7ec6b8f4) stub fixme:ntdll:NtQuerySystemInformation (0x00000006,0x1407fc0,0x00008000,0x7ec6b8f4) stub fixme:ntdll:NtQuerySystemInformation (0x00000007,0x1407fc0,0x00008000,0x7ec6b8f4) stub fixme:ntdll:NtQuerySystemInformation info_class SYSTEM_PROCESSOR_PERFORMANCE_INFORMATION fixme:ntdll:NtQuerySystemInformation (0x00000009,0x1407fc0,0x00008000,0x7ec6b8f4) stub fixme:ntdll:NtQuerySystemInformation (0x0000000a,0x1407fc0,0x00008000,0x7ec6b8f4) stub fixme:ntdll:NtQuerySystemInformation (0x0000000c,0x1407fc0,0x00008000,0x7ec6b8f4) stub fixme:ntdll:NtQuerySystemInformation (0x0000000d,0x1407fc0,0x00008000,0x7ec6b8f4) stub
That is, it's calling NtQuerySystemInformation with an incrementing value of SystemInformationClass, the first parameter. I'll guess that cryptlib is attempting to fill an entropy pool with random data. The code that's likely being executed is something like that here: http://src.gnu-darwin.org/ports/security/cryptlib/work/random/win32.c
I'm guessing we don't handle enough of these queries to provide enough entropy for cryptlib, and therefore cryptlib refused to initialize.