Bug#368863: hald generates unaligned trap errors on Alpha
I've been looking into the hald unaligned trap messages on Alpha. I can report that unaligned trap messages are still being generated by hald. However, it is not hald's fault. In my case the messages are: hald(14947): unaligned trap at 022083cc: 00011fa49391 28 3 hald(14947): unaligned trap at 022083ec: 00011fa49391 2c 1 The address of the code producing the trap is above 0x0200, and on Alpha that is where libraries are loaded. So note: ldd hald libdbus-glib-1.so.2 = /usr/lib/libdbus-glib-1.so.2 (0x02042000) libgobject-2.0.so.0 = /usr/lib/libgobject-2.0.so.0 (0x0207a000) libglib-2.0.so.0 = /usr/lib/libglib-2.0.so.0 (0x020de000) libdbus-1.so.3 = /usr/lib/libdbus-1.so.3 (0x021d2000) libm.so.6.1 = /lib/ev67/libm.so.6.1 (0x02232000) libc.so.6.1 = /lib/ev67/libc.so.6.1 (0x022ea000) libnsl.so.1.1 = /lib/ev67/libnsl.so.1.1 (0x0246c000) libselinux.so.1 = /lib/libselinux.so.1 (0x0249a000) libpcre.so.3 = /usr/lib/libpcre.so.3 (0x024cc000) /lib/ld-linux.so.2 (0x0200) libdl.so.2.1 = /lib/ev67/libdl.so.2.1 (0x02506000) and libdbus-1 has greatest address less than 0x022083cc (the address of code producing the traps). This is the likely candidate. Indeed, on running gdb with debug versions of hald and libdbus-1 I was able to identify the offending line of code and it is in the libdbus-1 library. I have opened bug #502408 on the dbus and libdbus-1 packages, and once that is resolved, hald will no longer generate unaligned traps. Cheers Michael. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#368863: [Pkg-utopia-maintainers] Bug#368863: hald generates unaligned trap errors on Alpha
Michael Cree wrote: I've been looking into the hald unaligned trap messages on Alpha. I can report that unaligned trap [..] libdbus-1 I was able to identify the offending line of code and it is in the libdbus-1 library. I have opened bug #502408 on the dbus and libdbus-1 packages, and once that is resolved, hald will no longer generate unaligned traps. Hi Michael, thanks a lot for the very detailed analysis and the dbus patch. Very much appreciated! We will include your dbus patch in the next upload. Cheers, Michael -- Why is it that all of the instruments seeking intelligent life in the universe are pointed away from Earth? signature.asc Description: OpenPGP digital signature
Bug#368863: hald generates unaligned trap errors on Alpha
tag 368863 + moreinfo thanks Hi Bob, I'm wondering if you still see the 'unaligned trap' message from HAL on an up-to-date system, since the bug you filed is rather old (May 2006). Regards, -- Sam Morris http://robots.org.uk/ PGP key id 1024D/5EA01078 3412 EA18 1277 354B 991B C869 B219 7FDB 5EA0 1078 signature.asc Description: This is a digitally signed message part
Bug#368863: hald generates unaligned trap errors on Alpha
On Sat, May 27, 2006 at 10:56:09AM -0500, Bob Tracy wrote: Sjoerd Simons wrote: Unfortunately I don't have an alpha lying around, so i can't reproduce this :). Afaik an unaligned trap doesn't do any harm, it's just a performance hit. That's pretty much my observation as well: no real harm being done that I can tell, just the warning messages that aren't really flagged as such. Note that unaligned accesses are fatal on the buildds and on all sparc machines, so you really don't want to have unaligned accesses. Also note that alpha and ia64 machines can turn off unaligned traps, so that the kernel will kill off the process like it does on sparc. Unofficial Debian policy is that unaligned accesses are not allowed. I confess to being not much good with a debugger, but I do just fine with diagnostic printf()'s (assuming the compiler doesn't help too much by moving things around). A debugger really is the best tool here. Recompile hal without optimization (use DEB_BUILD_OPTS=noopt,nostrip). Use sudo gdb /usr/lib/hal/hald-runner pid to attach. You have to be root because hald-runner is running with privileges. Then, look up the address that triggered the bug, and type disassemble address. It will give you a function name as part of the disassembly, and you should report that function to the bug. Then, we'll have the problem narrowed down to one function. If you CC me on that, I will try to investigate the problem. I'll retrieve a copy of the source package and start playing around with it. If I find out anything useful, I'll let you know. On the assumption (most certainly valid) that you know more about this than I do, what am I looking for? Suboptimally chosen pointer/variable sizes? Weird structure layouts? Questionable (implicit) typecasts? All of these are possible. Unfortunately, there isn't one that stands out much more than the others, although explicit typecasts are common, especially from a less-aligned type (such as char) to a more-aligned type (such as int or long). -- brian m. carlson / brian with sandals: Houston, Texas, US +1 713 440 7475 | http://crustytoothpaste.ath.cx/~bmc | My opinion only a typesetting engine: http://crustytoothpaste.ath.cx/~bmc/code/thwack OpenPGP: RSA v4 4096b 88AC E9B2 9196 305B A994 7552 F1BA 225C 0223 B187 signature.asc Description: Digital signature
Bug#368863: hald generates unaligned trap errors on Alpha
On Thu, Dec 13, 2007 at 06:59:53PM +0100, Sjoerd Simons wrote: On Thu, Dec 13, 2007 at 05:48:40PM +, brian m. carlson wrote: Note that unaligned accesses are fatal on the buildds and on all sparc machines, so you really don't want to have unaligned accesses. Also note that alpha and ia64 machines can turn off unaligned traps, so that the kernel will kill off the process like it does on sparc. Unofficial Debian policy is that unaligned accesses are not allowed. Do you still actually see unaligned traps with recent hal? We fixed some of them some time ago. Also hal seems to work just fine on my SPARC, which suggest these problems have been fixed. I don't have an alpha, and I don't run hal on my Ultra 5. I've never seen any problems, but I was trying to see if the submitter could solve the problems and at least tell us where the bug was. (If I could see them, I would have debugged them as far as I could.) It is possible that an unaligned access only occurs on a subset of platforms, just because of the data layout in the binary. It's just my experience that when sparc tends to have unaligned access problems, they tend to be crippling. If they're gone, then great. Just close the bug, and ignore my email. :-) -- brian m. carlson / brian with sandals: Houston, Texas, US +1 713 440 7475 | http://crustytoothpaste.ath.cx/~bmc | My opinion only a typesetting engine: http://crustytoothpaste.ath.cx/~bmc/code/thwack OpenPGP: RSA v4 4096b 88AC E9B2 9196 305B A994 7552 F1BA 225C 0223 B187 signature.asc Description: Digital signature
Bug#368863: hald generates unaligned trap errors on Alpha
On Thu, Dec 13, 2007 at 05:48:40PM +, brian m. carlson wrote: On Sat, May 27, 2006 at 10:56:09AM -0500, Bob Tracy wrote: Sjoerd Simons wrote: Unfortunately I don't have an alpha lying around, so i can't reproduce this :). Afaik an unaligned trap doesn't do any harm, it's just a performance hit. That's pretty much my observation as well: no real harm being done that I can tell, just the warning messages that aren't really flagged as such. Note that unaligned accesses are fatal on the buildds and on all sparc machines, so you really don't want to have unaligned accesses. Also note that alpha and ia64 machines can turn off unaligned traps, so that the kernel will kill off the process like it does on sparc. Unofficial Debian policy is that unaligned accesses are not allowed. Do you still actually see unaligned traps with recent hal? We fixed some of them some time ago. Also hal seems to work just fine on my SPARC, which suggest these problems have been fixed. I have no problem in looking for more, but i need some hints where to look :) Sjoerd -- There is more to life than increasing its speed. -- Mahatma Gandhi -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#368863: hald generates unaligned trap errors on Alpha
On Thu, May 25, 2006 at 09:57:52AM -0500, Bob Tracy wrote: Package: hal Version: 0.5.7-1 At boot time, and whenever I load/unload a hardware module, hald generates unaligned trap error messages on my Alpha system. For the boot time case, here is the corresponding syslog output with a few lines on each side of the error for context: May 24 08:07:49 smirkin kernel: hald-runner(2760): unaligned trap at 021ca7bc: 00011f57d8d1 28 4 May 24 08:07:49 smirkin kernel: hald-runner(2760): unaligned trap at 021ca7e0: 00011f57d8d1 2c 1 May 24 08:07:49 smirkin kernel: hald-runner(2760): unaligned trap at 021ca7bc: 00011f57d8d1 28 4 May 24 08:07:49 smirkin kernel: hald-runner(2760): unaligned trap at 021ca7e0: 00011f57d8d1 2c 1 Unfortunately I don't have an alpha lying around, so i can't reproduce this :). Afaik an unaligned trap doesn't do any harm, it's just a performance hit. I can fix this for you, but i need you to pinpoint what exactly causes these traps. Which requires some debugging on your part. Sorry that i can't help more at this point.. Sjoerd -- No amount of careful planning will ever replace dumb luck. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#368863: hald generates unaligned trap errors on Alpha
Sjoerd Simons wrote: Unfortunately I don't have an alpha lying around, so i can't reproduce this :). Afaik an unaligned trap doesn't do any harm, it's just a performance hit. That's pretty much my observation as well: no real harm being done that I can tell, just the warning messages that aren't really flagged as such. I can fix this for you, but i need you to pinpoint what exactly causes these traps. Which requires some debugging on your part. I confess to being not much good with a debugger, but I do just fine with diagnostic printf()'s (assuming the compiler doesn't help too much by moving things around). I'll retrieve a copy of the source package and start playing around with it. If I find out anything useful, I'll let you know. On the assumption (most certainly valid) that you know more about this than I do, what am I looking for? Suboptimally chosen pointer/variable sizes? Weird structure layouts? Questionable (implicit) typecasts? Thanks for your help!! -- --- Bob Tracy WTO + WIPO = DMCA? http://www.anti-dmca.org [EMAIL PROTECTED] --- -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#368863: hald generates unaligned trap errors on Alpha
Package: hal Version: 0.5.7-1 At boot time, and whenever I load/unload a hardware module, hald generates unaligned trap error messages on my Alpha system. For the boot time case, here is the corresponding syslog output with a few lines on each side of the error for context: ... May 24 08:07:45 smirkin kernel: parport0: PC-style at 0x3bc [PCSPP,TRISTATE] May 24 08:07:45 smirkin kernel: lp0: using parport0 (polling). May 24 08:07:45 smirkin kernel: eth0: no IPv6 routers present May 24 08:07:49 smirkin kernel: hald-runner(2760): unaligned trap at 021ca7bc: 00011f57d8d1 28 4 May 24 08:07:49 smirkin kernel: hald-runner(2760): unaligned trap at 021ca7e0: 00011f57d8d1 2c 1 May 24 08:07:49 smirkin kernel: hald-runner(2760): unaligned trap at 021ca7bc: 00011f57d8d1 28 4 May 24 08:07:49 smirkin kernel: hald-runner(2760): unaligned trap at 021ca7e0: 00011f57d8d1 2c 1 May 24 08:07:54 smirkin kernel: NET: Registered protocol family 5 May 24 08:07:54 smirkin atalkd[2852]: restart (2.0.3) ... uname -a: Linux smirkin 2.6.17-rc4 #1 Mon May 15 11:28:37 CDT 2006 alpha GNU/Linux C library: libc6.1 package version 2.3.6-7 System: 433au (miata) -- --- Bob Tracy WTO + WIPO = DMCA? http://www.anti-dmca.org [EMAIL PROTECTED] --- -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]