[Bug gas/10193] gas produces different .o files with the same input .s file and cmd line arguments on enfs
--- Additional Comments From chenyang dot cq at gmail dot com 2009-05-26 14:03 --- Thank u Alan for the directions. -- http://sourceware.org/bugzilla/show_bug.cgi?id=10193 --- You are receiving this mail because: --- You are on the CC list for the bug, or are watching someone who is. ___ bug-binutils mailing list bug-binutils@gnu.org http://lists.gnu.org/mailman/listinfo/bug-binutils
[Bug gas/10193] gas produces different .o files with the same input .s file and cmd line arguments on enfs
--- Additional Comments From amodra at bigpond dot net dot au 2009-05-26 00:19 --- Because that would be more work in BFD if we wanted to catch all possible cases of padding, slowing down all file writes. You could of course patch your own binutils sources. The place that skipped writing for your testcase was bfd/elf.c:assign_file_positions_except_relocs, here: /* Place the section headers. */ off = align_file_position (off, 1 << bed->s->log_file_align); You will find that the same thing will happen when ld generates executable files, so there will be more places to patch. -- http://sourceware.org/bugzilla/show_bug.cgi?id=10193 --- You are receiving this mail because: --- You are on the CC list for the bug, or are watching someone who is. ___ bug-binutils mailing list bug-binutils@gnu.org http://lists.gnu.org/mailman/listinfo/bug-binutils
[Bug gas/10193] gas produces different .o files with the same input .s file and cmd line arguments on enfs
--- Additional Comments From chenyang dot cq at gmail dot com 2009-05-25 08:22 --- >You have a filesystem problem. gas isn't writing these bytes, and a sane >filesystem would set them to zeros. So, not a gas bug. Why not explicitly set them to zeros so that gas can work well with those insane network file systems like the one (BWFS http://english.cas.ac.cn/english/news/detailnewsb.asp?InfoNo=26092) I'm using. -- http://sourceware.org/bugzilla/show_bug.cgi?id=10193 --- You are receiving this mail because: --- You are on the CC list for the bug, or are watching someone who is. ___ bug-binutils mailing list bug-binutils@gnu.org http://lists.gnu.org/mailman/listinfo/bug-binutils
[Bug gas/10193] gas produces different .o files with the same input .s file and cmd line arguments on enfs
--- Additional Comments From amodra at bigpond dot net dot au 2009-05-24 06:47 --- $ cmp -l obj[12].o 66022 40 325 66023 325 375 66024 373 312 66022 == 0x101e6. Looking at part of the section table, I see [23] .shstrtab STRTAB 0+0 010103 e2 00 0 0 1 readelf also tells me Start of section headers: 66024 (bytes into file) cmp starts counting bytes from 1, readelf from 0, so the bytes that differ are after the end of .shstrtab and the beginning of the section header table. You have a filesystem problem. gas isn't writing these bytes, and a sane filesystem would set them to zeros. So, not a gas bug. -- What|Removed |Added Status|NEW |RESOLVED Resolution||INVALID http://sourceware.org/bugzilla/show_bug.cgi?id=10193 --- You are receiving this mail because: --- You are on the CC list for the bug, or are watching someone who is. ___ bug-binutils mailing list bug-binutils@gnu.org http://lists.gnu.org/mailman/listinfo/bug-binutils
[Bug gas/10193] gas produces different .o files with the same input .s file and cmd line arguments on enfs
--- Additional Comments From chenyang dot cq at gmail dot com 2009-05-23 13:55 --- Created an attachment (id=3960) --> (http://sourceware.org/bugzilla/attachment.cgi?id=3960&action=view) testtree.s obj1.o obj2.o the assembler file and two different result object files. -- http://sourceware.org/bugzilla/show_bug.cgi?id=10193 --- You are receiving this mail because: --- You are on the CC list for the bug, or are watching someone who is. ___ bug-binutils mailing list bug-binutils@gnu.org http://lists.gnu.org/mailman/listinfo/bug-binutils