Re: cvs commit: src/sys/i386/isa apic_vector.s clock.c src/sys/kern kern_intr.c kern_sig.c kern_slaballoc.c kern_switch.c kern_systimer.c kern_upcall.c lwkt_msgport.c lwkt_thread.c lwkt_token.c src/sys/sys slaballoc.h

[YONETANI Tomokazu <qhwt+dragonfly-bugs@xxxxxxxxxx>]

On Wed, Feb 18, 2004 at 07:42:43AM +0900, YONETANI Tomokazu wrote:
> On Tue, Feb 17, 2004 at 10:17:37AM -0800, Matthew Dillon wrote:
> > :After this commit, the kernel built with gcc2 panics at the beginning of
> > :the boot leaving the following message on the screen(from dmesg bufffer
> > :survived across the reboot):
> > 
> >     Later commits should have fixed that issue.  The most recent commit
> >     just a day or two ago fixed common variables in modules which blew
> >     up certain loadable modules.
> > 
> >     Try using the absolute latest kernel and module set.
> 
> Updated the source tree and I have sys/conf/files.i386,v 1.16, but
> it still panics with gcc2, no problem with gcc3. The hardest part
> is that it doesn't drop to DDB after the panic.
> Oh, and some more info; if ACPI is compiled in the kernel and the
> kernel is built with gcc2, call to bus_dma_tag_create() fails:
> 
>   acpi_alloc_wakeup_handler: unable to create wake tag
> 
> this hasn't changed since a few days ago, so I believe
> sys/i386/i386/busdma_machdep.c,v 1.7 is innocent.

I'm sorry to have spammed the commits@ list, moving to bugs@.
I've found the path to the panic:
mi_startup
-> scmeminit
   -> sc_alloc_scr_buffer
      -> sc_vtb_copy
         -> sc_vtb_init, malloc failed because of M_NOWAIT flag
         -> bcopy to null pointer and panic

I'm still not sure why it failed only when the kernel is compiled with gcc2.