supposed bug in assembler "as"
Hello, I encountered a problem which seems to be a bug in the assembler 'as': Redefining a symbol with the '.set' directive in connection with the location counter does not work correctly. I am using: GNU assembler 2.14 20030612, configured for `i586-pc-msdosdjgpp', getting to DOS from Windows 98. I hope that the condensed little program I append (also in the file 'astest.S'), while isolating the supposed bug, contains all necessary information. Greetings from Germany, B. Strowitzki #- start of sample program .section .data .p2align 4,,15 somedata: .int 1,2,3,4,5,6,7,8,9,0 /* dummy values */ .set goodval_1, (.-somedata)/4 /* calculate number of entries */ .set badval, (.-somedata)/4 startval: .int badval /* retaining badval for later printing */ /* in real program to be stored into a "C"-defined struct producing an entry like "datablock.length" */ moredata: .rept 1024 .byte 32 .endr .set goodval_2, (.-moredata) /* same procedure again */ .set badval, (.-moredata) /* redefining badval, ok */ midval: .int badval evenmoredata: .space 200 .set goodval_3, (.-evenmoredata) .set badval, (.-evenmoredata) /* third definition: doesn't work! 'as' seems to get confused */ # /* shortcut with self-diagnosis by 'as' */ # /* gives an extra error message, but works */ #.if (goodval_3-badval) #.print "difference encountered" #.err #.endif breakval: .int badval lastdata: .space 1200, -1 .set goodval_4, (.-lastdata)/4 .set badval, (.-lastdata)/4 /* doesn't work also */ # further strange behavior .set dotval_good_1, . .set dotval_bad, . firstdot: .int dotval_bad /* no problem so far */ .space 12 /* dummy */ .set dotval_good_2, . .set dotval_bad, . /* corrupts the value already stored in 'firstdot' */ nextdot: .int dotval_bad format: .asciz " %i. right val: %i, badval: %i\12\15" .section .text .p2align 4,,15 .globl _main _main: pushl %ebp movl %esp,%ebp movl $startval,%eax pushl (%eax) pushl $goodval_1 pushl $1 pushl $format call _printf/* emits equal numbers, ok */ movl $midval,%eax pushl (%eax) pushl $goodval_2 pushl $2 pushl $format call _printf/* emits equal numbers, ok */ movl $breakval,%eax pushl (%eax) pushl $goodval_3 pushl $3 pushl $format call _printf /* emits two different numbers */ pushl $badval pushl $goodval_4 pushl $4 pushl $format call _printf /* different numbers again */ pushl $10 /* newline */ call _putch movl $firstdot,%eax pushl (%eax) pushl $dotval_good_1 pushl $5 pushl $format call _printf /* emits two different numbers with last value assigned to 'dotval_bad' */ movl $nextdot,%eax pushl (%eax) pushl $dotval_good_2 pushl $6 pushl $format call _printf /* emits equal numbers, ok */ xorl %eax,%eax /* exit code zero */ movl %ebp,%esp popl %ebp ret _ Der WEB.DE SmartSurfer hilft bis zu 70% Ihrer Onlinekosten zu sparen! http://smartsurfer.web.de/?mc=100071&distributionid=0066 astest.S Description: Binary data ___ bug-binutils mailing list bug-binutils@gnu.org http://lists.gnu.org/mailman/listinfo/bug-binutils
[Bug ld/5276] New: relax is not yet perfect. fails linking large c++
Last time I tried rebuilding firefox 2.0.0.8 rpm from Fedora on my Alpha it failed with some relocation problems like this: ... relocation truncated to fit: GPREL16 against symbol ... (Build log: http://buildsys.zero42.at/koji/getfile?taskID=53055&name=build.log) I know this error very well, because I've seen it many times when using relaxed binutils, especially when linking large c++ stuff. So to prove this, I today tried to build firefox 2.0.0.9 again and a similar problem occured. After the failed build I added -Wl,--no-relax to my LDFLAGS. Well, you can expect what happened: It finished building. It - of course - took longer to build, but worked... Do you need any input? I'm willing to give ssh access to the build machine if that helps! -- Summary: relax is not yet perfect. fails linking large c++ Product: binutils Version: 2.17 Status: NEW Severity: normal Priority: P2 Component: ld AssignedTo: unassigned at sources dot redhat dot com ReportedBy: oliver at linux-kernel dot at CC: bug-binutils at gnu dot org GCC build triplet: alpha-redhat-linux GCC host triplet: alpha-redhat-linux GCC target triplet: alpha-redhat-linux http://sourceware.org/bugzilla/show_bug.cgi?id=5276 --- 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
Re: supposed bug in assembler "as"
Hi Daniela, Redefining a symbol with the '.set' directive in connection with the location counter does not work correctly. I am using: GNU assembler 2.14 This is your problem. Version 2.14 is an old version which was known to have problems in this area. The current release is 2.18. Please upgrade to this release and you should find that everything works. Cheers Nick ___ bug-binutils mailing list bug-binutils@gnu.org http://lists.gnu.org/mailman/listinfo/bug-binutils
[Bug binutils/5278] New: objdump SEGFAULT in _bfd_pe_print_private_bfd_data_common
Several Win32 PE files, which appear to be UPX compressed, with broken/tweeked headers, cause SEGFAULTs (sig 11) when run as "objdump -x foo.exe" Example backtrace: The Export Tables (interpreted UPX1 section contents) Export Flagseb34e215 Time/Date stamp ebd57074 Major/Minor -21660/-29690 Program received signal SIGSEGV, Segmentation fault. 0x2ab3ac7a2cb0 in strlen () from /lib/tls/libc.so.6 (gdb) bt #0 0x2ab3ac7a2cb0 in strlen () from /lib/tls/libc.so.6 #1 0x2ab3ac775c8d in vfprintf () from /lib/tls/libc.so.6 #2 0x2ab3ac77b588 in fprintf () from /lib/tls/libc.so.6 #3 0x2ab3ac5ba698 in _bfd_pe_print_private_bfd_data_common () from /usr/lib64/binutils/x86_64-pc-linux-gnu/2.18/libbfd-2.18.so #4 0x2ab3ac5b3d89 in pe_print_private_bfd_data () from /usr/lib64/binutils/x86_64-pc-linux-gnu/2.18/libbfd-2.18.so #5 0x0040769a in dump_bfd () #6 0x00407f39 in display_bfd () #7 0x0040807d in display_file () #8 0x00408633 in main () (gdb) -- Summary: objdump SEGFAULT in _bfd_pe_print_private_bfd_data_common Product: binutils Version: 2.18 Status: NEW Severity: critical Priority: P2 Component: binutils AssignedTo: unassigned at sources dot redhat dot com ReportedBy: julia at fireeye dot com CC: bug-binutils at gnu dot org http://sourceware.org/bugzilla/show_bug.cgi?id=5278 --- 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 binutils/5279] New: objdump SEGFAULT in bfd_getl16
In certain Win32 PE (i386) EXE files (which might be packed), objdump segfaults due to something in the PE headers. I can provide an example .EXE file to reproduce this. Reproduction: "objdump -x foo.exe" Example backtrace: [...] 191009 0 RegQueryValueA 191009 19101b 0 RegSetValueA 19101b 19102b 0 RegSetValueExA 19102b 00190014 0019015c 03e7 00190f21 00190334 DLL Name: KERNEL32.dll vma: Hint/Ord Member-Name Bound-To Program received signal SIGSEGV, Segmentation fault. 0x2ab55058b150 in bfd_getl16 () from /usr/lib64/binutils/x86_64-pc-linux-gnu/2.18/libbfd-2.18.so (gdb) bt #0 0x2ab55058b150 in bfd_getl16 () from /usr/lib64/binutils/x86_64-pc-linux-gnu/2.18/libbfd-2.18.so #1 0x2ab5505e00fa in _bfd_pe_print_private_bfd_data_common () from /usr/lib64/binutils/x86_64-pc-linux-gnu/2.18/libbfd-2.18.so #2 0x2ab5505d8d89 in pe_print_private_bfd_data () from /usr/lib64/binutils/x86_64-pc-linux-gnu/2.18/libbfd-2.18.so #3 0x0040769a in dump_bfd () #4 0x00407f39 in display_bfd () #5 0x0040807d in display_file () #6 0x00408633 in main () (gdb) -- Summary: objdump SEGFAULT in bfd_getl16 Product: binutils Version: 2.18 Status: NEW Severity: normal Priority: P2 Component: binutils AssignedTo: unassigned at sources dot redhat dot com ReportedBy: julia at fireeye dot com CC: bug-binutils at gnu dot org http://sourceware.org/bugzilla/show_bug.cgi?id=5279 --- 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 ld/5276] relax is not yet perfect. fails linking large c++
--- Additional Comments From hjl at lucon dot org 2007-11-07 04:51 --- Can you try binutils 2.18? -- What|Removed |Added CC||hjl at lucon dot org Status|NEW |WAITING http://sourceware.org/bugzilla/show_bug.cgi?id=5276 --- 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 binutils/5279] objdump SEGFAULT in bfd_getl16
--- Additional Comments From hjl at lucon dot org 2007-11-07 04:53 --- Please provide testcases. -- What|Removed |Added CC||hjl at lucon dot org Status|NEW |WAITING http://sourceware.org/bugzilla/show_bug.cgi?id=5279 --- 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 binutils/5278] objdump SEGFAULT in _bfd_pe_print_private_bfd_data_common
--- Additional Comments From hjl at lucon dot org 2007-11-07 04:54 --- Please provide testcase -- What|Removed |Added CC||hjl at lucon dot org http://sourceware.org/bugzilla/show_bug.cgi?id=5278 --- 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
