Date: Mon, 28 Jun 1999 04:50:02 +0800 From: Peter Wemm <peter@netplex.com.au> To: Dmitrij Tejblum <tejblum@arc.hq.cti.ru> Cc: Doug Rabson <dfr@nlsystems.com>, alpha@FreeBSD.ORG Subject: Re: Hmm!! Message-ID: <19990627205002.0C7CB81@overcee.netplex.com.au> In-Reply-To: Your message of "Sun, 27 Jun 1999 20:44:08 %2B0400." <199906271644.UAA00502@tejblum.dnttm.rssi.ru>
next in thread | previous in thread | raw e-mail | index | archive | help
Dmitrij Tejblum wrote:
> I once noticed that our pmap miss some calls to alpha_pal_imb(). I did
> suspect it may cause problems like the ones you see. Unfortunately, I
> did almost nothing to fix that, but still have plans :-(. At that time,
> I made this change to pmap.c (cut & pasted)
>
> @@ -2176,6 +2183,8 @@
> *pte = newpte;
> if (origpte)
> pmap_invalidate_page(pmap, va);
> + if (prot & VM_PROT_EXECUTE)
> + alpha_pal_imb();
> }
> }
>
> and probably didn't see such gremlins since that, but it may be
> coincidence. (I got a feeling that the weird things I seen in a
> nightmare, not in reality...)
>
> (I suspect a little it may be 164sx specific.)
>
> Dima
I went a bit further and it seems to have fixed my mystery problems:
Index: pmap.c
===================================================================
RCS file: /home/ncvs/src/sys/alpha/alpha/pmap.c,v
retrieving revision 1.26
diff -c -r1.26 pmap.c
*** pmap.c 1999/06/10 20:40:55 1.26
--- pmap.c 1999/06/27 20:41:52
***************
*** 728,745 ****
static void
pmap_invalidate_page(pmap_t pmap, vm_offset_t va)
{
! if (pmap_isactive(pmap))
ALPHA_TBIS(va);
! else
pmap_invalidate_asn(pmap);
}
static void
pmap_invalidate_all(pmap_t pmap)
{
! if (pmap_isactive(pmap))
ALPHA_TBIA();
! else
pmap_invalidate_asn(pmap);
}
--- 728,747 ----
static void
pmap_invalidate_page(pmap_t pmap, vm_offset_t va)
{
! if (pmap_isactive(pmap)) {
ALPHA_TBIS(va);
! alpha_pal_imb();
! } else
pmap_invalidate_asn(pmap);
}
static void
pmap_invalidate_all(pmap_t pmap)
{
! if (pmap_isactive(pmap)) {
ALPHA_TBIA();
! alpha_pal_imb();
! } else
pmap_invalidate_asn(pmap);
}
***************
*** 2176,2181 ****
--- 2178,2185 ----
*pte = newpte;
if (origpte)
pmap_invalidate_page(pmap, va);
+ if (prot & VM_PROT_EXECUTE)
+ alpha_pal_imb();
}
}
I'm pretty sure this is overkill, but it doesn't seem to hurt so far. I
have not tried your single patch though, I got out the sledgehammer to see
if I could get some sanity. I'm not sure if the alpha_pal_tbi() functions
do an implied imb as well or not. If so, then the first chunks are
irrelevant.
Cheers,
-Peter
--
Peter Wemm - peter@FreeBSD.org; peter@yahoo-inc.com; peter@netplex.com.au
To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-alpha" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?19990627205002.0C7CB81>