[Bug ld/12772] linker failed to remove all dead code dependency properly
http://sourceware.org/bugzilla/show_bug.cgi?id=12772 --- Comment #8 from cvs-commit at gcc dot gnu.org 2011-12-07 12:16:03 UTC --- CVSROOT:/cvs/src Module name:src Changes by:amo...@sourceware.org2011-12-07 12:15:54 Modified files: bfd: ChangeLog elflink.c Log message: PR ld/12772 * elflink.c (elf_gc_sweep_symbol): Discard unmarked symbols defined in shared libraries. Patches: http://sourceware.org/cgi-bin/cvsweb.cgi/src/bfd/ChangeLog.diff?cvsroot=src&r1=1.5533&r2=1.5534 http://sourceware.org/cgi-bin/cvsweb.cgi/src/bfd/elflink.c.diff?cvsroot=src&r1=1.431&r2=1.432 -- Configure bugmail: http://sourceware.org/bugzilla/userprefs.cgi?tab=email --- You are receiving this mail because: --- You are on the CC list for the bug. ___ bug-binutils mailing list bug-binutils@gnu.org https://lists.gnu.org/mailman/listinfo/bug-binutils
[Bug ld/12772] linker failed to remove all dead code dependency properly
http://sourceware.org/bugzilla/show_bug.cgi?id=12772 Alan Modra changed: What|Removed |Added Status|ASSIGNED|RESOLVED Resolution||FIXED --- Comment #9 from Alan Modra 2011-12-07 12:17:21 UTC --- Fixed mainline -- Configure bugmail: http://sourceware.org/bugzilla/userprefs.cgi?tab=email --- You are receiving this mail because: --- You are on the CC list for the bug. ___ bug-binutils mailing list bug-binutils@gnu.org https://lists.gnu.org/mailman/listinfo/bug-binutils
[Bug binutils/13476] New: Incorrect TLS relocations created in DSOs on hppa
http://sourceware.org/bugzilla/show_bug.cgi?id=13476 Bug #: 13476 Summary: Incorrect TLS relocations created in DSOs on hppa Product: binutils Version: 2.23 (HEAD) Status: NEW Severity: normal Priority: P2 Component: binutils AssignedTo: unassig...@sourceware.org ReportedBy: sk...@netbsd.org Classification: Unclassified Created attachment 6092 --> http://sourceware.org/bugzilla/attachment.cgi?id=6092 Example code for DSO This commit http://sourceware.org/ml/binutils/2010-01/msg00594.html now means that calls to __tls_get_addr are incorrect as the relocation is no longer adjusted to be relative to the GOT. Looking at the final link object from, e.g. cc -c -O2 -fPIC -o h_tls.o h_tls.c cc -shared -o dso.so h_tls.o note the addil/ldo sequence is beyond the GOT. 04f4 : 4f4: 6b c2 3f d9 stw rp,-14(sp) 4f8: 6f c4 01 00 stw,ma r4,80(sp) 4fc: 6b c3 3f 09 stw r3,-7c(sp) 500: 08 1a 02 43 copy r26,r3 504: 2a 60 20 00 addil L%1000,r19,r1 508: 34 3a 00 70 ldo 38(r1),r26 -- Configure bugmail: http://sourceware.org/bugzilla/userprefs.cgi?tab=email --- You are receiving this mail because: --- You are on the CC list for the bug. ___ bug-binutils mailing list bug-binutils@gnu.org https://lists.gnu.org/mailman/listinfo/bug-binutils
[Bug binutils/13477] New: incorrect decoding by objdump for fsubrp on i386
http://sourceware.org/bugzilla/show_bug.cgi?id=13477 Bug #: 13477 Summary: incorrect decoding by objdump for fsubrp on i386 Product: binutils Version: 2.22 Status: NEW Severity: normal Priority: P2 Component: binutils AssignedTo: unassig...@sourceware.org ReportedBy: drspaml...@gmail.com Classification: Unclassified Created attachment 6093 --> http://sourceware.org/bugzilla/attachment.cgi?id=6093 executable demonstrating decode fail objdump on i386 decodes DE E1 to fsubp st(1),st when it should be fsubrp st(1),st probably true also for DE E2 to DE E7 An example is in the file attached. The command run was: objdump -w -d -M intel,addr32,data32 test-i386 > test-i386.disam -- Configure bugmail: http://sourceware.org/bugzilla/userprefs.cgi?tab=email --- You are receiving this mail because: --- You are on the CC list for the bug. ___ bug-binutils mailing list bug-binutils@gnu.org https://lists.gnu.org/mailman/listinfo/bug-binutils
Linux Linker / PIC and --defsym
Hello, Here is a suggestion to make --defsym / linker script defined symbols relevant in PIC binaries: Having --defsym to define a relocated symbol in a PIC library/executable is pointless (it does not serve ANY purpose other than polluting the binary). Therefore, the only relevant use of --defsym in a PIC shared library or PIC executable should be to have a *non-relocated* "absolute address". Incidentally, that's the official purpose of --defsym: "Create a global symbol in the output file, containing the absolute address given by expression." Today's solution is to patch the relocation table after the binary image has been loaded by the Linux linker. Then, --defsym becomes useful in PIC libraries/executables (with this feature, a shared library can call a function located in a stripped executable). -- Pierre Gauthier | "Either find a way, or make one." pie...@trustleap.com | (Hannibal, 247-183 BC) --- http://trustleap.ch/ --- Paradiesli 17 - CH-8842 Unteriberg, SZ - Switzerland Tel +41 (0)55 414 20 93 - Fax +41 (0)55 414 20 67 ___ bug-binutils mailing list bug-binutils@gnu.org https://lists.gnu.org/mailman/listinfo/bug-binutils
[Bug binutils/13476] Incorrect TLS relocations created in DSOs on hppa
http://sourceware.org/bugzilla/show_bug.cgi?id=13476 Mike Frysinger changed: What|Removed |Added CC||toolchain at gentoo dot org -- Configure bugmail: http://sourceware.org/bugzilla/userprefs.cgi?tab=email --- You are receiving this mail because: --- You are on the CC list for the bug. ___ bug-binutils mailing list bug-binutils@gnu.org https://lists.gnu.org/mailman/listinfo/bug-binutils
[Bug binutils/13482] New: inaccurate documentation
http://sourceware.org/bugzilla/show_bug.cgi?id=13482 Bug #: 13482 Summary: inaccurate documentation Product: binutils Version: 2.22 Status: NEW Severity: normal Priority: P2 Component: binutils AssignedTo: unassig...@sourceware.org ReportedBy: dave.paw...@gmail.com Classification: Unclassified http://sources.redhat.com/binutils/docs-2.22/as/Version.html#Version "This directive creates a .note section and places into it an ELF formatted note of type NT_VERSION. The note's name is set to string." Inaccurate. .version "Version 1.0" in arm sourcecode running arm-linux-elf-as GNU assembler (GNU Binutils) 2.22.51.2022 Copyright 2011 Free Software Foundation, Inc. Causes readelf: Warning: corrupt note found at offset 18 into core notes readelf: Warning: type: 1, namesize: 000c, descsize: It would appear that the string must not contain White space. Change to Version1.0 and the error goes DaveP -- Configure bugmail: http://sourceware.org/bugzilla/userprefs.cgi?tab=email --- You are receiving this mail because: --- You are on the CC list for the bug. ___ bug-binutils mailing list bug-binutils@gnu.org https://lists.gnu.org/mailman/listinfo/bug-binutils
[Bug binutils/13476] Incorrect TLS relocations created in DSOs on hppa
http://sourceware.org/bugzilla/show_bug.cgi?id=13476 John David Anglin changed: What|Removed |Added CC||danglin at gcc dot gnu.org --- Comment #1 from John David Anglin 2011-12-08 01:43:58 UTC --- 1) The testcase doesn't show the bug. 2) The register is still r19 (PIC register). The addil hasn't been changed to dp or r0, so the I fail to see how the commit referenced changed the situation. The first hunk doesn't apply because this is a shared link. The second hunk doesn't apply because the register isn't dp. So, we should fallthru to the original code. -- Configure bugmail: http://sourceware.org/bugzilla/userprefs.cgi?tab=email --- You are receiving this mail because: --- You are on the CC list for the bug. ___ bug-binutils mailing list bug-binutils@gnu.org https://lists.gnu.org/mailman/listinfo/bug-binutils
[Bug binutils/13476] Incorrect TLS relocations created in DSOs on hppa
http://sourceware.org/bugzilla/show_bug.cgi?id=13476 --- Comment #2 from John David Anglin 2011-12-08 02:08:56 UTC --- Looking a bit more, I think this break is questionable: /* Now try to make things easy for the dynamic linker. */ break; -- Configure bugmail: http://sourceware.org/bugzilla/userprefs.cgi?tab=email --- You are receiving this mail because: --- You are on the CC list for the bug. ___ bug-binutils mailing list bug-binutils@gnu.org https://lists.gnu.org/mailman/listinfo/bug-binutils
[Bug binutils/13476] Incorrect TLS relocations created in DSOs on hppa
http://sourceware.org/bugzilla/show_bug.cgi?id=13476 --- Comment #3 from Nick Hudson 2011-12-08 07:43:50 UTC --- To hopefully show the bug clearer here's some more output. $ cc -c -O2 -fPIC -o h_tls.o h_tls.c $ $ ../prefix/bin/hppa--netbsd-ld -v GNU ld (GNU Binutils) 2.22.51.20111207 $ $ ../prefix/bin/hppa--netbsd-ld --eh-frame-hdr -shared -o dso.so -L /usr/lib /usr/lib/crti.o /usr/lib/crtbeginS.o h_tls.o -lgcc -lc -lgcc /usr/lib/crtendS.o /usr/lib/crtn.o $ $ nm dso.so | grep test 04e8 T testf $ $ objdump -D dso.so --start-address=0x4e8 --stop-address=0x504 dso.so: file format elf32-hppa-netbsd Disassembly of section .text: 04e8 : 4e8: 6b c2 3f d9 stw rp,-14(sp) 4ec: 6f c4 01 00 stw,ma r4,80(sp) 4f0: 6b c3 3f 09 stw r3,-7c(sp) 4f4: 08 1a 02 43 copy r26,r3 4f8: 2a 60 20 00 addil L%1000,r19,r1 ! r1 = gp + 0x1000 4fc: 34 3a 00 70 ldo 38(r1),r26 ! r26 = gp + 0x1038 500: e8 5f 1d 1d b,l 394 <_init+0x40>,rp $ $ # On elf32-hppa-netbsd gp is always set to point to the GOT. GOT + 0x1038 is $ # well outside the got $ # $ objdump -h dso.so dso.so: file format elf32-hppa-netbsd Sections: Idx Name Size VMA LMA File off Algn 0 .hash 004c 00b4 00b4 00b4 2**2 CONTENTS, ALLOC, LOAD, READONLY, DATA 1 .dynsym 00e0 0100 0100 0100 2**2 CONTENTS, ALLOC, LOAD, READONLY, DATA 2 .dynstr 0083 01e0 01e0 01e0 2**0 CONTENTS, ALLOC, LOAD, READONLY, DATA 3 .rela.dyn 00cc 0264 0264 0264 2**2 CONTENTS, ALLOC, LOAD, READONLY, DATA 4 .rela.plt 0024 0330 0330 0330 2**2 CONTENTS, ALLOC, LOAD, READONLY, DATA 5 .init 0030 0354 0354 0354 2**0 CONTENTS, ALLOC, LOAD, READONLY, CODE 6 .text 0384 0384 0384 0384 2**2 CONTENTS, ALLOC, LOAD, READONLY, CODE 7 .fini 0028 0708 0708 0708 2**0 CONTENTS, ALLOC, LOAD, READONLY, CODE 8 .note.netbsd.ident 0018 0730 0730 0730 2**2 CONTENTS, ALLOC, LOAD, READONLY, DATA 9 .note.netbsd.pax 0014 0748 0748 0748 2**2 CONTENTS, ALLOC, LOAD, READONLY, DATA 10 .PARISC.unwind 00b0 075c 075c 075c 2**2 CONTENTS, ALLOC, LOAD, READONLY, DATA 11 .eh_frame 0004 080c 080c 080c 2**2 CONTENTS, ALLOC, LOAD, READONLY, DATA 12 .ctors0008 1000 1000 1000 2**2 CONTENTS, ALLOC, LOAD, DATA 13 .dtors0008 1008 1008 1008 2**2 CONTENTS, ALLOC, LOAD, DATA 14 .jcr 0004 1010 1010 1010 2**2 CONTENTS, ALLOC, LOAD, DATA 15 .data.rel.ro 0008 1014 1014 1014 2**2 CONTENTS, ALLOC, LOAD, DATA 16 .dynamic 00a8 101c 101c 101c 2**2 CONTENTS, ALLOC, LOAD, DATA 17 .data 0004 10c4 10c4 10c4 2**2 CONTENTS, ALLOC, LOAD, DATA 18 .plt 0034 10c8 10c8 10c8 2**2 CONTENTS, ALLOC, LOAD, CODE 19 .got 0040 10fc 10fc 10fc 2**2 CONTENTS, ALLOC, LOAD, DATA 20 .bss 0014 113c 113c 113c 2**2 ALLOC 21 .comment 0042 113c 2**0 CONTENTS, READONLY 22 .debug_aranges 0020 117e 2**0 CONTENTS, READONLY, DEBUGGING 23 .debug_pubnames 0039 119e 2**0 CONTENTS, READONLY, DEBUGGING 24 .debug_info 01d9 11d7 2**0 CONTENTS, READONLY, DEBUGGING 25 .debug_abbrev 0120 13b0 2**0 CONTENTS, RE
