Issue 19261 in oss-fuzz: binutils:fuzz_disassemble: Timeout in fuzz_disassemble

2020-01-12 Thread sheriff… via monorail

Updates:
Labels: -restrict-view-commit

Comment #4 on issue 19261 by sheriff...@chromium.org:  
binutils:fuzz_disassemble: Timeout in fuzz_disassemble

https://bugs.chromium.org/p/oss-fuzz/issues/detail?id=19261#c4

This bug has been fixed for 30 days. It has been opened to the public.

- Your friendly Sheriffbot

--
You received this message because:
  1. You were specifically CC'd on the issue

You may adjust your notification preferences at:
https://bugs.chromium.org/hosting/settings

Reply to this email to add a comment.


[Bug ld/25374] New: linking to msvc import library results in invalid IDT entry

2020-01-12 Thread draymond at foxvalley dot net
https://sourceware.org/bugzilla/show_bug.cgi?id=25374

Bug ID: 25374
   Summary: linking to msvc import library results in invalid IDT
entry
   Product: binutils
   Version: unspecified
Status: UNCONFIRMED
  Severity: normal
  Priority: P2
 Component: ld
  Assignee: unassigned at sourceware dot org
  Reporter: draymond at foxvalley dot net
  Target Milestone: ---

I have a DLL that I built and linked with i686-w64-mingw32-gcc.  It imports
functions from two other images (an MSVC-built EXE and DLL).  I linked against
the import libraries of the MSVC images (foo.lib and bar.lib) using
'-lfoo -lbar' on the command line.  No errors were reported.  When I loaded
the MinGW DLL and called into it I experienced a crash.

The MinGW DLL was able to call functions in the MSVC DLL but when it tried to
call a function in the MSVC EXE the import library thunk pulled an invalid
address from the IAT (import address table).  It appears that the loader did
not fill in the IAT.

Examination of the .idata section in the MinGW DLL revealed that ILT and IAT
tables were present for both the MSVC DLL and EXE.  The IDT entry (which
contains pointers to the ILT and IAT) was correct for the MSVC DLL.  The IDT
entry for the MSVC EXE, however, was incorrect.  The ILT and IAT pointers
pointed to the NULL entries at the end of the tables rather than the first
entries.

For this reason 'objdump -p mingw.dll' shows the import table for the MSVC
EXE as empty.  This also explains why the loader did not fill in the IAT.

-- 
You are receiving this mail because:
You are on the CC list for the bug.


[Bug ld/25374] linking to msvc import library results in invalid IDT entry

2020-01-12 Thread martin at martin dot st
https://sourceware.org/bugzilla/show_bug.cgi?id=25374

Martin Storsjö  changed:

   What|Removed |Added

 CC||martin at martin dot st

--- Comment #1 from Martin Storsjö  ---
(In reply to Dan Raymond from comment #0)
> Examination of the .idata section in the MinGW DLL revealed that ILT and IAT
> tables were present for both the MSVC DLL and EXE.  The IDT entry (which
> contains pointers to the ILT and IAT) was correct for the MSVC DLL.  The IDT
> entry for the MSVC EXE, however, was incorrect.  The ILT and IAT pointers
> pointed to the NULL entries at the end of the tables rather than the first
> entries.

I was able to reproduce a similar issue when linking against two import
libraries for two DLLs (without involving the concept of EXEs exporting symbols
and import libraries for them).

My suspicion is that the synthesized import table chunks (when using msvc style
import libraries, where the linker needs to synthesize the import table
contents) aren't sorted correctly if there's more than one msvc style import
library involved.

-- 
You are receiving this mail because:
You are on the CC list for the bug.