public inbox for linux-ia64@vger.kernel.org
 help / color / mirror / Atom feed
* Question about backing store access rights
@ 2007-02-20 10:07 Christian Kandeler
  2007-02-20 10:42 ` Christian Kandeler
  0 siblings, 1 reply; 2+ messages in thread
From: Christian Kandeler @ 2007-02-20 10:07 UTC (permalink / raw)
  To: linux-ia64

Hello,

I hope that someone can enlighten me about how the mapping of a process' RSE 
backing store page is supposed to work. 
From what I can see, the area gets initialized in the ia64_init_addr_space() 
function, where its vm_page_prot member is set to __P011, i.e. PAGE_READONLY, 
which I assume is supposed to be changed on the first write access. However, 
according to the IASDM, RSE stores always set ISR.w to zero on an 
interruption, so the page fault handler will never know that it's dealing 
with a write access. Consequently, processes should end up in an endless loop 
of IR Data Access Rights faults. Why don't they? What am I missing here?


Thanks,
Christian

^ permalink raw reply	[flat|nested] 2+ messages in thread

* Re: Question about backing store access rights
  2007-02-20 10:07 Question about backing store access rights Christian Kandeler
@ 2007-02-20 10:42 ` Christian Kandeler
  0 siblings, 0 replies; 2+ messages in thread
From: Christian Kandeler @ 2007-02-20 10:42 UTC (permalink / raw)
  To: linux-ia64

Christian Kandeler wrote:

> [nonsense]

Please disregard. Sorry.

^ permalink raw reply	[flat|nested] 2+ messages in thread

end of thread, other threads:[~2007-02-20 10:42 UTC | newest]

Thread overview: 2+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2007-02-20 10:07 Question about backing store access rights Christian Kandeler
2007-02-20 10:42 ` Christian Kandeler

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox