Re: 4.8-current fails with uvm_fault in em(4)
On Tue, Aug 03, 2010 at 11:12:01PM +0100, Stuart Henderson wrote: Do you know what the last working kernel was? I have an archive of old kernels if you want to track it down without the cvs up -D / compile dance, mail me offlist. broken: OpenBSD 4.7-current (GENERIC) #104: Thu Jul 8 16:47:03 MDT 2010 works: OpenBSD 4.7-current (GENERIC) #98: Wed Jul 7 18:25:12 MDT 2010 I don't see how it could be an em problem as there are no commits in this window and all the commits either side of it are chip specific to other chips.
Re: 4.8-current fails with uvm_fault in em(4)
When I revert my tree back to the working snapshot time ( with cvs -d anon...@anoncvs.openbsd.org:/cvs -q up -PAd -D 2010-07-07 18:25:12 MDT ) the resulting kernel doesn't work. The only difference to the snapshot kernel is the compiler. In the mentioned time window i386 moved to gcc4: B B B CVSROOT:B B B /cvs B B B Module name:B B B src B B B Changes by:B B B kette...@cvs.openbsd.orgb B B 2010/07/08 14:25:40 B B B Modified files: B B B B B B share/mkB B B B B B : bsd.own.mk B B B B B B lib/libcB B B B B B : shlib_version B B B Log message: B B B Switch hppa, i386 and powerpc to gcc4.B Bump libc major to handle ABI corner B B B cases fixed in gcc4. The kernel stops working at em_init_eeprom_params+0x16:B movl 0x4(%eax),%edx so maybe a similar fix like below needs to be done in a second place ... B B B CVSROOT:B B B /cvs B B B Module name:B B B src B B B Changes by:B B B kette...@cvs.openbsd.orgb B B 2010/07/09 12:44:10 B B B Modified files: B B B B B B sys/arch/i386/include: profile.h B B B Log message: B B B We need to preserve %eax, %ecx and %edx since GCC 4.x might use these B B B registers in the prologue to realign the stack. B B B ok marco@, deraadt@, guenther@ ... but that's just guessing. Regards, Holger ;-se Jonathan Gray j...@goblin.cx hat am 4. August 2010 um 12:17 geschrieben: On Tue, Aug 03, 2010 at 11:12:01PM +0100, Stuart Henderson wrote: Do you know what the last working kernel was? I have an archive of old kernels if you want to track it down without the cvs up -D / compile dance, mail me offlist. broken: OpenBSD 4.7-current (GENERIC) #104: Thu JulB 8 16:47:03 MDT 2010 works: OpenBSD 4.7-current (GENERIC) #98: Wed JulB 7 18:25:12 MDT 2010 I don't see how it could be an em problem as there are no commits in this window and all the commits either side of it are chip specific to other chips.
Re: 4.8-current fails with uvm_fault in em(4)
Sorry for the last (unreadable) mail. Same text, second try: When reverting the tree back to the working snapshot time ( with cvs -q up -PAd -D 2010-07-07 18:25:12 MDT ) the resulting kernel doesn't work. The only difference to the snapshot kernel is the compiler. In the mentioned time window i386 moved to gcc4: B CVSROOT:B B B /cvs B Module name:B B B src B Changes by:B B B kette...@cvs.openbsd.orgb B B 2010/07/08 14:25:40 B Modified files: B B B B B share/mkB B B B B B : bsd.own.mk B B B B B lib/libcB B B B B B : shlib_version B Log message: B Switch hppa, i386 and powerpc to gcc4.B Bump libc major to handle ABI corner B cases fixed in gcc4. The kernel stops working at em_init_eeprom_params+0x16:B movl 0x4(%eax),%edx so maybe a similar fix like below needs to be done in a second place as well... B CVSROOT:B B B /cvs B Module name:B B B src B Changes by:B B B kette...@cvs.openbsd.orgb B B 2010/07/09 12:44:10 B Modified files: B B B B B sys/arch/i386/include: profile.h B Log message: B We need to preserve %eax, %ecx and %edx since GCC 4.x might use these B registers in the prologue to realign the stack. B ok marco@, deraadt@, guenther@ ... but that's just guessing. Regards, Holger ;-se