Re: config(8) dumps core
on 29/04/2010 18:31 Michael Moll said the following: [snip] Assertion failed: (r != '\0' (Char present in the configuration string mustn't be equal to 0)), function kernconfdump, file /usr/src/usr.sbin/config/main.c, line 721. [snip] Any ideas on this? Yes, one idea - to verify what the message above says. You can use hd to see if you indeed have '\0' (0x00) symbol somewhere within your kernel config file. -- Andriy Gapon ___ freebsd-current@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-current To unsubscribe, send any mail to freebsd-current-unsubscr...@freebsd.org
Re: config(8) dumps core
on 30/04/2010 00:12 Michael Moll said the following: Hi, On Thu, Apr 29, 2010 at 11:33:30PM +0300, Andriy Gapon wrote: on 29/04/2010 18:31 Michael Moll said the following: You can use hd to see if you indeed have '\0' (0x00) symbol somewhere within your kernel config file. Thanks, I checked this and there are no 0x00s in the config file itself, Then that assert message is strange. Or there is something else to this situation. but a hd to /boot/kernel/kernel reveals: 09 66 77 69 70 0a 64 65 76 69 63 65 09 64 63 6f |.fwip.device.dco| 6e 73 0a 64 65 76 69 63 65 09 64 63 6f 6e 73 5f |ns.device.dcons_| 63 72 6f 6d 0a 00 00 00 00 00 00 00 00 00 00 00 |crom| 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 || This also explains why a recent config-binary worked against the old kernel... The were some commits to /src/usr.sbin/config/* in the last weeks, maybe one of them broke this. Actually I think that this doesn't mean anything. /boot/kernel/kernel is a binary, an executable, it is expected to have a fair amount of 0x00 in it. That assert was specifically about kernel _config_ file. -- Andriy Gapon ___ freebsd-current@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-current To unsubscribe, send any mail to freebsd-current-unsubscr...@freebsd.org
Re: config(8) dumps core
Hi, On Thu, Apr 29, 2010 at 11:33:30PM +0300, Andriy Gapon wrote: on 29/04/2010 18:31 Michael Moll said the following: You can use hd to see if you indeed have '\0' (0x00) symbol somewhere within your kernel config file. Thanks, I checked this and there are no 0x00s in the config file itself, but a hd to /boot/kernel/kernel reveals: 09 66 77 69 70 0a 64 65 76 69 63 65 09 64 63 6f |.fwip.device.dco| 6e 73 0a 64 65 76 69 63 65 09 64 63 6f 6e 73 5f |ns.device.dcons_| 63 72 6f 6d 0a 00 00 00 00 00 00 00 00 00 00 00 |crom| 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 || This also explains why a recent config-binary worked against the old kernel... The were some commits to /src/usr.sbin/config/* in the last weeks, maybe one of them broke this. Kind Regards -- Michael Moll ___ freebsd-current@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-current To unsubscribe, send any mail to freebsd-current-unsubscr...@freebsd.org
Re: config(8) dumps core
On Thursday 29 April 2010 22:19:45 Andriy Gapon wrote: on 30/04/2010 00:12 Michael Moll said the following: Hi, On Thu, Apr 29, 2010 at 11:33:30PM +0300, Andriy Gapon wrote: on 29/04/2010 18:31 Michael Moll said the following: You can use hd to see if you indeed have '\0' (0x00) symbol somewhere within your kernel config file. Thanks, I checked this and there are no 0x00s in the config file itself, Then that assert message is strange. Or there is something else to this situation. but a hd to /boot/kernel/kernel reveals: 09 66 77 69 70 0a 64 65 76 69 63 65 09 64 63 6f |.fwip.device.dco| 6e 73 0a 64 65 76 69 63 65 09 64 63 6f 6e 73 5f |ns.device.dcons_| 63 72 6f 6d 0a 00 00 00 00 00 00 00 00 00 00 00 |crom| 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 || This also explains why a recent config-binary worked against the old kernel... The were some commits to /src/usr.sbin/config/* in the last weeks, maybe one of them broke this. Actually I think that this doesn't mean anything. /boot/kernel/kernel is a binary, an executable, it is expected to have a fair amount of 0x00 in it. That assert was specifically about kernel _config_ file. It's expected to have some 0x00s, but hopefully not in the middle of the embedded kernel configuration file that has recently been added to GENERIC :) -- Bruce ___ freebsd-current@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-current To unsubscribe, send any mail to freebsd-current-unsubscr...@freebsd.org
Re: config(8) dumps core
On Thu, Apr 29, 2010 at 2:41 PM, Bruce Cran br...@cran.org.uk wrote: On Thursday 29 April 2010 22:19:45 Andriy Gapon wrote: on 30/04/2010 00:12 Michael Moll said the following: Hi, On Thu, Apr 29, 2010 at 11:33:30PM +0300, Andriy Gapon wrote: on 29/04/2010 18:31 Michael Moll said the following: You can use hd to see if you indeed have '\0' (0x00) symbol somewhere within your kernel config file. Thanks, I checked this and there are no 0x00s in the config file itself, Then that assert message is strange. Or there is something else to this situation. but a hd to /boot/kernel/kernel reveals: 09 66 77 69 70 0a 64 65 76 69 63 65 09 64 63 6f |.fwip.device.dco| 6e 73 0a 64 65 76 69 63 65 09 64 63 6f 6e 73 5f |ns.device.dcons_| 63 72 6f 6d 0a 00 00 00 00 00 00 00 00 00 00 00 |crom| 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 || This also explains why a recent config-binary worked against the old kernel... The were some commits to /src/usr.sbin/config/* in the last weeks, maybe one of them broke this. Actually I think that this doesn't mean anything. /boot/kernel/kernel is a binary, an executable, it is expected to have a fair amount of 0x00 in it. That assert was specifically about kernel _config_ file. It's expected to have some 0x00s, but hopefully not in the middle of the embedded kernel configuration file that has recently been added to GENERIC :) Michael, Does a version prior to last week work, and could you please attach your kernel configuration file(s) for analysis? Thanks, -Garrett ___ freebsd-current@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-current To unsubscribe, send any mail to freebsd-current-unsubscr...@freebsd.org
Re: config(8) dumps core
On Thursday 29 April 2010 22:46:29 Garrett Cooper wrote: Does a version prior to last week work, and could you please attach your kernel configuration file(s) for analysis? I remember coming across the same error during buildkernel last week, but can't remember how, or how I resolved it. I've just built GENERIC from recent sources and can get the assert to occur by running config -x /boot/kernel/kernel. It looks like it thinks there's more data available than there is. config/main.c gets the size of the configuration by running elfdump: elfdump -c /boot/kernel/kernel | grep -A 5 kern_conf | tail -2 | cut -d ' ' -f 2 | paste - - - 100533682656 The first column is the offset, and the second is the number of bytes, but the embedded configuration only has 2649 bytes. Some of the output from config: config -x /boot/kernel/kernel options CONFIG_AUTOGENERATED ident GENERIC machine sparc64 cpu SUN4U makeoptions DEBUG=-g options USB_DEBUG options AH_SUPPORT_AR5416 options IEEE80211_SUPPORT_MESH options IEEE80211_AMPDU_AGE [...] device uath device ural device zyd device firewire device sbp device fwe device fwip device dcons device dcons_crom Assertion failed: (r != '\0' (Char present in the configuration string mustn't be equal to 0)), function kernconfdump, file /usr/src/usr.sbin/config/main.c, line 719. Abort (core dumped) -- Bruce Cran ___ freebsd-current@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-current To unsubscribe, send any mail to freebsd-current-unsubscr...@freebsd.org
Re: config(8) dumps core
Hi, On Thu, Apr 29, 2010 at 11:30:06PM +0100, Bruce Cran wrote: /boot/kernel/kernel. It looks like it thinks there's more data available than there is. config/main.c gets the size of the configuration by running elfdump: elfdump -c /boot/kernel/kernel | grep -A 5 kern_conf | tail -2 | cut -d ' ' -f 2 | paste - - - 100533682656 The first column is the offset, and the second is the number of bytes, but the embedded configuration only has 2649 bytes. OK, that explains that with all the related commits backup out (207265-206664) the resulting config-binary still fails. I don't have time today to build the kernel with the different revisions to hunt the problem down to one commit... imp@: The last commits to usr.sbin/config/* might have raised this problem, could you have a look at it? Kind Regards -- Michael Moll ___ freebsd-current@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-current To unsubscribe, send any mail to freebsd-current-unsubscr...@freebsd.org
Re: config(8) dumps core
In message: 20100429224938.gi90...@darkthrone.kvedulv.de Michael Moll kved...@kvedulv.de writes: : Hi, : : On Thu, Apr 29, 2010 at 11:30:06PM +0100, Bruce Cran wrote: : /boot/kernel/kernel. It looks like it thinks there's more data available : than there is. : : config/main.c gets the size of the configuration by running elfdump: : : elfdump -c /boot/kernel/kernel | grep -A 5 kern_conf | tail -2 | cut -d ' ' : -f 2 | paste - - - : : 100533682656 : : The first column is the offset, and the second is the number of bytes, but the : embedded configuration only has 2649 bytes. : : OK, that explains that with all the related commits backup out : (207265-206664) the resulting config-binary still fails. I don't have : time today to build the kernel with the different revisions to hunt the : problem down to one commit... : : imp@: The last commits to usr.sbin/config/* might have raised this : problem, could you have a look at it? Do you have a small, reproducible sequence of events that will recreate this problem? I've seen no problems at all in my testing. I assume this is in -current? Warner ___ freebsd-current@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-current To unsubscribe, send any mail to freebsd-current-unsubscr...@freebsd.org
Re: config(8) dumps core
On Thu, Apr 29, 2010 at 05:07:00PM -0600, M. Warner Losh wrote: Do you have a small, reproducible sequence of events that will recreate this problem? I've seen no problems at all in my testing. I assume this is in -current? Yes, -CURRENT. Compile a kernel on sparc64 (didn't try this yet on other archs) and run config -x on the resulting file. Kind Regards -- Michael Moll ___ freebsd-current@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-current To unsubscribe, send any mail to freebsd-current-unsubscr...@freebsd.org
Re: config(8) dumps core
In message: 2010043016.ga90...@darkthrone.kvedulv.de Michael Moll kved...@kvedulv.de writes: : On Thu, Apr 29, 2010 at 05:07:00PM -0600, M. Warner Losh wrote: : Do you have a small, reproducible sequence of events that will : recreate this problem? I've seen no problems at all in my testing. : I assume this is in -current? : : Yes, -CURRENT. Compile a kernel on sparc64 (didn't try this yet on other : archs) and run config -x on the resulting file. Can you please enclose the exact config file that you are using? Thanks. Warner ___ freebsd-current@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-current To unsubscribe, send any mail to freebsd-current-unsubscr...@freebsd.org
Re: config(8) dumps core
On Thursday 29 April 2010 23:49:38 Michael Moll wrote: OK, that explains that with all the related commits backup out (207265-206664) the resulting config-binary still fails. I don't have time today to build the kernel with the different revisions to hunt the problem down to one commit... imp@: The last commits to usr.sbin/config/* might have raised this problem, could you have a look at it? It appears it was r188214 that caused the breakage: in versions prior to that, it outputs 6 trailing NULs and exits normally on a modern kernel. If it only broke recently then I guess it was kernel, not changes to the config app, that broke it. -- Bruce Cran ___ freebsd-current@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-current To unsubscribe, send any mail to freebsd-current-unsubscr...@freebsd.org