Port-sparc archive

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index][Old Index]

re: partially working SMP again

   > i spent a bunch of time investigating sparc SMP issues this weekend
   > and i've gotten my system to the point it turns on interrupts and
   > then loses, instead of crashing before printing the copyright or
   > while attaching cpus.
   > basic changes are:
   >     - kill union cpu_info_pg
   >     - allocate USPACE after the intstack (this is the real
   >       bug fix)
   >     - make cpus[] as array of cpu_info{} *'s.
   >     - allocate cpu_info{}'s in pmap_bootstrap via stealing
   >       pages aka pmap_bootstrap style.
   >     - we waste a page of for cpu0's cpu_info above, but for
   >       now just ignore this, we use the VA it would use, but
   >       remap the boot CPUINFO_VA physaddr to there
   Have these changes been committed?

   SS20MP fails earlier due to mutex panic:

earlier?  earlier than mine, perhaps, but not earlier than it used
to before these changes?  it used to crash in pmap_bootstrap() or
in main() for me, depending on the age.

   data fault: pc=0xf0134b80 addr=0x0 

can you find out what line pc=0xf0134b80 is?



Home | Main Index | Thread Index | Old Index