[Bug 37940] New: Invalid file buffering behavior causes data corruption
https://bugs.winehq.org/show_bug.cgi?id=37940 Bug ID: 37940 Summary: Invalid file buffering behavior causes data corruption Product: Wine Version: 1.7.34 Hardware: x86 OS: Linux Status: UNCONFIRMED Severity: major Priority: P2 Component: msvcrt Assignee: wine-bugs(a)winehq.org Reporter: ho-rr-or(a)mail.ru Distribution: --- Created attachment 50543 --> https://bugs.winehq.org/attachment.cgi?id=50543 Compiled binary to reproduce msvcrt file routines sometimes corrupt file data by invalid buffer flushing. Reproduced by following code (compiled binary attached): char data[68200]; char buf[sizeof(data)]; for (int i = 0; i < sizeof(data); i++) data[i] = rand(); FILE* f = _wfsopen(L"test.bin", L"wb+", 64); rewind(f); fwrite(data, 1, 68100, f); // Following calls possible corrupt buffer rewind(f); fread(buf, 1, 40000, f); fread(buf, 1, 40000, f); // Following call causes wrong buffer flushing fwrite(&data[68100], 1, 100, f); fclose(f); f = fopen("test.bin", "rb"); fread(buf, 1, sizeof(buf), f); fclose(f); printf("Memcmp: %d\n", memcmp(data, buf, sizeof(buf))); Windows output: Memcmp: 0 Wine output: Memcmp: -1 Windows file size: 68200 Wine file size: 72296 Bug causes at least cl.exe in Microsoft Visual Studio 2008/2010. Reproduced when cl.exe generates precompiled header. test.h: #pragma once #include "Windows.h" #include <iostream> test.cpp: #include "test.h" int main() { std::cout << "WORKS!" << std::endl; return 0; } Command to reproduce: rm test.pch; wine cl.exe /Yctest.h test.cpp -- 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=37940 Mikhail Bystryantsev <ho-rr-or(a)mail.ru> changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |ho-rr-or(a)mail.ru -- 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=37940 Austin English <austinenglish(a)gmail.com> changed: What |Removed |Added ---------------------------------------------------------------------------- Keywords| |download, source, testcase -- 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=37940 Piotr Caban <piotr.caban(a)gmail.com> changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |piotr.caban(a)gmail.com --- Comment #1 from Piotr Caban <piotr.caban(a)gmail.com> --- The attached test works now (commit ba0d1586ff53221e449f2ed885a059232dd4bea8). Please retest. -- 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=37940 --- Comment #2 from Mikhail Bystryantsev <ho-rr-or(a)mail.ru> --- Attached test now works fine, but the cl.exe still does not. I'll try to check that it is still due to this bug and provide more test cases. -- 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=37940 --- Comment #3 from Mikhail Bystryantsev <ho-rr-or(a)mail.ru> --- Created attachment 50553 --> https://bugs.winehq.org/attachment.cgi?id=50553 Second binary for reproduce There is new binary for reproduce. I reproduced a nearly the same operations that performs cl.exe. char data[68100 + 65536]; char buf[sizeof(data)]; for (int i = 0; i < sizeof(data); i++) data[i] = rand(); FILE* f = _wfsopen(L"test.bin", L"wb+", 64); rewind(f); fwrite(data, 1, 68100, f); // Following calls possible corrupt buffer rewind(f); fread(buf, 1, 32512, f); fread(buf, 1, 4096, f); fread(buf, 1, 28672, f); fread(buf, 1, 4096, f); fread(buf, 1, 29948, f); // Following call causes wrong buffer flushing fwrite(&data[68100], 1, 65536, f); fclose(f); f = fopen("test.bin", "rb"); fread(buf, 1, sizeof(buf), f); fclose(f); printf("Memcmp: %d\n", memcmp(data, buf, sizeof(buf))); -- 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=37940 --- Comment #4 from Piotr Caban <piotr.caban(a)gmail.com> --- The attached test is working again. Could you please give it a try? (commit 90b68f99648a0a0be73673fd99e140e18bfb5c35) -- 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=37940 --- Comment #5 from Mikhail Bystryantsev <ho-rr-or(a)mail.ru> --- Now cl.exe works fine. Nice work! -- 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=37940 Piotr Caban <piotr.caban(a)gmail.com> changed: What |Removed |Added ---------------------------------------------------------------------------- Fixed by SHA1| |90b68f99648a0a0be73673fd99e | |140e18bfb5c35 Status|UNCONFIRMED |RESOLVED Resolution|--- |FIXED --- Comment #6 from Piotr Caban <piotr.caban(a)gmail.com> --- Marking as fixed. -- 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=37940 Alexandre Julliard <julliard(a)winehq.org> changed: What |Removed |Added ---------------------------------------------------------------------------- Status|RESOLVED |CLOSED --- Comment #7 from Alexandre Julliard <julliard(a)winehq.org> --- Closing bugs fixed in 1.7.36. -- 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