| Summary: | Segmentation fault in grub | ||
|---|---|---|---|
| Product: | Gentoo Linux | Reporter: | Henrique Dias <hdias> |
| Component: | [OLD] Core system | Assignee: | AMD64 Project <amd64> |
| Status: | RESOLVED NEEDINFO | ||
| Severity: | critical | ||
| Priority: | High | ||
| Version: | unspecified | ||
| Hardware: | AMD64 | ||
| OS: | Linux | ||
| Whiteboard: | |||
| Package list: | Runtime testing required: | --- | |
|
Description
Henrique Dias
2005-03-29 11:22:44 UTC
Please comment the line "unset CFLAGS" in the ebuild and install grub again via FEATURES="nostrip" CFLAGS="-O -ggdb" emerge grub. Then run grub via gdb grub > run and try to reproduce the segfault. When it happens, type "backtrace" in gdb and attach the output of the complete gdb session here. # gdb grub GNU gdb 6.0 Copyright 2003 Free Software Foundation, Inc. GDB is free software, covered by the GNU General Public License, and you are welcome to change it and/or distribute copies of it under certain conditions. Type "show copying" to see the conditions. There is absolutely no warranty for GDB. Type "show warranty" for details. This GDB was configured as "x86_64-pc-linux-gnu"...Using host libthread_db library "/lib/libthread_db.so.1". (gdb) run Starting program: /sbin/grub Warning: Cannot insert breakpoint -2. Error accessing memory address 0xbc20: Input/output error. (gdb) backtrace #0 0x00000000 in ?? () (gdb) I/O error on RAM? i guess your RAM has some errors, so it's probably hardware related No, the RAM is ok. I get this message from dmesg # dmesg end_request: I/O error, dev fd0, sector 0 grub[14302]: segfault at 00000000556f2ac0 rip 00000000556f2ac0 rsp 00000000556f29ec error 15 end_request: I/O error, dev fd0, sector 0 grub[14369]: segfault at 00000000556f2ac0 rip 00000000556f2ac0 rsp 00000000556f29ec error 15 end_request: I/O error, dev fd0, sector 0 grub[14446]: segfault at 00000000556f2ac0 rip 00000000556f2ac0 rsp 00000000556f29ec error 15 I can confirm the bug - the exact same thing happens here (although no fd errors, I'm running from the hard disk). I was booting with grub-static, but now I wanted to install the "normal" grub. I'm trying to install it on a running system, but it always segfaults. The kernel logs also show a similar line: Apr 4 01:24:35 [kernel] grub[12462]: segfault at 000000005570dac0 rip 000000005570dac0 rsp 000000005570d9ec error 15 Maybe it is related to the amd64 noexec protection? I always boot the kernel with the 'noexec=on noexec32=all' parameters. Correction: there's a fd0 error after all, but it happens the moment I do 'grub> root(hd0,1)', so it doesn't seem to be related. The segfault itself happens when I run setup: grub> setup (hd0) setup (hd0) Checking if "/boot/grub/stage1" exists... yes Checking if "/boot/grub/stage2" exists... yes Checking if "/boot/grub/xfs_stage1_5" exists... yes Running "embed /boot/grub/xfs_stage1_5 (hd0)"... 25 sectors are embedded. succeeded Segmentation fault Remove grub and install grub-static. I don't have problem's with grub-static. Does this still happen if you use the "--no-floppy" argument when installing grub? Yes I get a segmentation fault with or without "--no-floppy" option. is there a fd0 in /boot/grub/device.map and is there a floppy drive in your computer ? I had the same problem, on an Intel EMT64 Xeon. I got around it by compiling grub-0.96-r2. Also, a friend of mine insists that he addded USE='xfs XFS' and that fixed it. I have no clue, all I do know is that I am using XFS, and I can see no way in which that use flag could have any effect. In either case, try 0.96-r2, and if that doesn't work, try USE="xfs XFS" with it. (In reply to comment #2) > # gdb grub > (gdb) run > Starting program: /sbin/grub > Warning: > Cannot insert breakpoint -2. > Error accessing memory address 0xbc20: Input/output error. > > (gdb) backtrace > #0 0x00000000 in ?? () ^^^ there is no debug information here. This info is useless for us. I explicitly asked you to 1. comment the line "unset CFLAGS" in your grub ebuild 2. remerge grub by running: FEATURES="nostrip" CFLAGS="-O -ggdb" emerge grub Please do this and attach the information here. I really want this one to be resolved. (In reply to comment #11) > I had the same problem, on an Intel EMT64 Xeon. I got around it by compiling > grub-0.96-r2. Also, a friend of mine insists that he addded USE='xfs XFS' and > that fixed it. I have no clue, all I do know is that I am using XFS, and I > can see no way in which that use flag could have any effect. In either case, > try 0.96-r2, and if that doesn't work, try USE="xfs XFS" with it. This is nonsense. USE flags are case insensitive. (In reply to comment #6) > The segfault itself happens when I run setup: > > grub> setup (hd0) > setup (hd0) > Checking if "/boot/grub/stage1" exists... yes > Checking if "/boot/grub/stage2" exists... yes > Checking if "/boot/grub/xfs_stage1_5" exists... yes > Running "embed /boot/grub/xfs_stage1_5 (hd0)"... 25 sectors are embedded. > succeeded > Segmentation fault > If you are using xfs on your boot partition, it is Bug 90845. Fixed in 0.96-r2 (and 0.97 upstream). Ricardo: can you reboot your system without these options? i think that's the simplest way to find out whether noexec is the problem or not. also, can you please do what Danny asked in comment 1? This will hopefully give us some useful debug information we can't fix this without debug information, sorry |
