Re: port-amd64/49643: 7.99.4/amd64 xrestr()/fxsavel panic

[David Laight <david%l8s.co.uk@localhost>]

The following reply was made to PR port-amd64/49643; it has been noted by GNATS.

From: David Laight <david%l8s.co.uk@localhost>
To: gnats-bugs%NetBSD.org@localhost
Cc: 
Subject: Re: port-amd64/49643: 7.99.4/amd64 xrestr()/fxsavel panic
Date: Tue, 10 Feb 2015 22:00:20 +0000

 On Sat, Feb 07, 2015 at 05:10:11PM +0000, Christoph Badura wrote:
 > The following reply was made to PR port-amd64/49643; it has been noted by GNATS.
 > 
 >  The tail of msgbuf and and backtrace from crash(8), pid 1680.1 is Xorg:
 >  
 >  fatal protection fault in supervisor mode
 >  trap type 4 code 0 rip ffffffff8028967a cs 8 rflags 13046 cr2 7f7fffffff28 ileve
 >  l 8 rsp fffffe810f6adec8
 >  curlwp 0xfffffe823971a680 pid 1680.1 lowest kstack 0xfffffe810f6aa2c0
 >  
 >  dumping to dev 0,1 (offset=217031, size=2070023):
 >  dump
 >  crash> bt
 >  _KERNEL_OPT_NVGA_RASTERCONSOLE() at 0
 >  _KERNEL_OPT_IPFILTER_COMPAT() at _KERNEL_OPT_IPFILTER_COMPAT+0x7
 >  db_reboot_cmd() at db_reboot_cmd
 >  db_command() at db_command+0xf0
 >  db_command_loop() at db_command_loop+0xd0
 >  db_trap() at db_trap+0xe9
 >  kdb_trap() at kdb_trap+0xe5
 >  trap() at trap+0x1b3
 >  --- trap (number 4) ---
 >  xrstor() at xrstor+0xa
 
 Not entirely useful - looks like crash fails to follow the
 stack out of the xrstor() stub.
 OTOH I'd guess there is only one call and that is from the 'fp registers
 not present' handler.
 
 A real guess is that the 'wakeup' sequence does not execute the code
 that enables xsave/xrstor instructions.
 I don't have a source tree handy, but it should be easy enough to
 add a diagnostic check before the xrstor.
 
 	David
 
 -- 
 David Laight: david%l8s.co.uk@localhost