* mips64 pgd_current
@ 2001-03-16 20:55 Justin Carlson
2001-03-16 22:09 ` Justin Carlson
0 siblings, 1 reply; 2+ messages in thread
From: Justin Carlson @ 2001-03-16 20:55 UTC (permalink / raw)
To: linux-mips
Doing some more code trolling here; am looking at the MP
parts of the TLB refill code in the mips64 tree, and I'm a bit
confused.
It looks to me like pgd_current[] is indexed by CP0_CONTEXT
and the resulting pointer is supposed to be the base of the
page tables for that particular process. This makes sense.
However, I see this in mmu_context.h:
#define TLBMISS_HANDLER_SETUP_PGD(pgd) \
pgd_current[smp_processor_id()] = (unsigned long)(pgd)
#define TLBMISS_HANDLER_SETUP() \
set_context((unsigned long) smp_processor_id() << (23 + 3)); \
TLBMISS_HANDLER_SETUP_PGD(swapper_pg_dir)
extern unsigned long pgd_current[];
so pgd_current[smp_processor_id()] is hard coded to be set to
swapper_pg_dir, which, insofar as I can tell, is not a per-cpu structure.
It looks to me like this will end up with pgd_current[] being an
array of NR_CPUS pointers all of which point to the same pgd table.
I know I must be misunderstanding this code. Any suggestions as to what I'm
missing?
Thanks,
Justin
^ permalink raw reply [flat|nested] 2+ messages in thread
end of thread, other threads:[~2001-03-16 22:08 UTC | newest]
Thread overview: 2+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2001-03-16 20:55 mips64 pgd_current Justin Carlson
2001-03-16 22:09 ` Justin Carlson
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox