public inbox for linux-mtd@lists.infradead.org
 help / color / mirror / Atom feed
* Processor Specific Files
@ 2000-08-23 21:59 mark.langsdorf
  2000-08-24  7:52 ` David Woodhouse
  0 siblings, 1 reply; 2+ messages in thread
From: mark.langsdorf @ 2000-08-23 21:59 UTC (permalink / raw)
  To: mtd

I'm using an Elan Sc520 with 16 megs of AMD CFI compliant 
flash on it.  The flash is arranged into two banks, each 
built from 4 interleaved 8-bit flash chips.  Each bank is 
separately mapped into physically memory using a Programmable 
Address Region (PAR), and the BIOS sets these up and makes 
them write-cacheable by default.

My problem is that the cfi_probe() code fails if the PARS 
are write-cacheable, since the return value of an earlier 
probe is cached, and then returned for the 4x8 probe.  It's 
possible to reprogram the PARs, but I'm not sure where to 
do it.  Should I put it in physmap.c, at the start of the
initialization routine?  Someplace else?  Create a specific
module for my processor and do it there?  What's the best
way to handle this?

Mark Langsdorf
Advanced Micro Devices, Inc             Tel: 512.602.3756
5204 E. Ben White Blvd. M/S 590         Fax: 512.602.5051
Austin, TX 78741                        mark.langsdorf@amd.com


To unsubscribe, send "unsubscribe mtd" to majordomo@infradead.org

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

* Re: Processor Specific Files
  2000-08-23 21:59 Processor Specific Files mark.langsdorf
@ 2000-08-24  7:52 ` David Woodhouse
  0 siblings, 0 replies; 2+ messages in thread
From: David Woodhouse @ 2000-08-24  7:52 UTC (permalink / raw)
  To: mark.langsdorf; +Cc: mtd


mark.langsdorf@amd.com said:
>  My problem is that the cfi_probe() code fails if the PARS  are
> write-cacheable, since the return value of an earlier  probe is
> cached, and then returned for the 4x8 probe.

Why not just fix the BIOS? Making flash regions write-cacheable is just 
wrong, surely?

It would possibly be useful to have some generic code in the physmap.c map 
driver which attempts to turn of caching for the region it's using - is 
there a portable way to call arch-specific routines to do that?

--
dwmw2




To unsubscribe, send "unsubscribe mtd" to majordomo@infradead.org

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

end of thread, other threads:[~2000-08-24  7:53 UTC | newest]

Thread overview: 2+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2000-08-23 21:59 Processor Specific Files mark.langsdorf
2000-08-24  7:52 ` David Woodhouse

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