https://llvm.org/bugs/show_bug.cgi?id=26968
Bug ID: 26968
Summary: i386 lld produces incorrect fatal error "SHF_MERGE
section size must be a multiple of sh_entsize" linking
FreeBSD/i386 userland
Product: lld
Version: unspecified
Hardware: PC
URL: http://reviews.llvm.org/D18222
OS: FreeBSD
Status: NEW
Severity: normal
Priority: P
Component: All Bugs
Assignee: unassignedb...@nondot.org
Reporter: ema...@freebsd.org
CC: llvm-bugs@lists.llvm.org
Classification: Unclassified
lld considers sh_entsize == 0 a fatal error in a SHF_MERGE section.
if (!EntSize || Sec.sh_size % EntSize)
fatal("SHF_MERGE section size must be a multiple of sh_entsize");
In a FreeBSD/i386 buildworld many objects have SHF_MERGE with sh_entsize == 0,
and this shouldn't be a fatal error. At a minimum we can just return false.
For reference here is an example readelf -S from one of the objects:
/tank/emaste/obj/i386.i386/tank/emaste/src/freebsd/tmp/usr/lib/crt1.o
There are 14 section headers, starting at offset 0x68c:
Section Headers:
[Nr] Name Type Addr Off Size ES Flg Lk Inf
Al
[ 0] NULL 00000000 000000 000000 00 0 0
0
[ 1] .eh_frame PROGBITS 00000000 000034 000034 00 A 0 0
4
[ 2] .note.tag NOTE 00000000 000068 000030 00 A 0 0
4
[ 3] .rodata PROGBITS 00000000 000098 000001 00 AMS 0 0
1
[ 4] .text PROGBITS 00000000 0000a0 0001aa 00 AX 0 0
16
[ 5] .data PROGBITS 00000000 00024c 000004 00 WA 0 0
4
[ 6] .bss NOBITS 00000000 000250 000004 00 WA 0 0
4
[ 7] .rel.text REL 00000000 000250 0000b8 08 I 12 4
4
[ 8] .comment PROGBITS 00000000 000308 00005d 00 MS 0 0
1
[ 9] .rel.eh_frame REL 00000000 000368 000008 08 I 12 1
4
[10] .rel.data REL 00000000 000370 000008 08 I 12 5
4
[11] .shstrtab STRTAB 00000000 000548 00005d 00 0 0
1
[12] .symtab SYMTAB 00000000 000378 0001d0 10 13 21
4
[13] .strtab STRTAB 00000000 0005a5 0000e5 00 0 0
1
As an aside, it would be useful to have the file name in the error message too.
--
You are receiving this mail because:
You are on the CC list for the bug.
_______________________________________________
llvm-bugs mailing list
llvm-bugs@lists.llvm.org
http://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-bugs
