From: steiner@sgi.com
To: akpm@linux-foundation.org, linux-kernel@vger.kernel.org
Subject: [Patch 05/25] GRU - cbe cache flush
Date: Thu, 26 Aug 2010 08:19:42 -0500 [thread overview]
Message-ID: <20100826132123.763139236@sgi.com> (raw)
In-Reply-To: 20100826131937.108920216@sgi.com
[-- Attachment #1: uv_gru_cbe_flush --]
[-- Type: text/plain, Size: 1187 bytes --]
From: Jack Steiner <steiner@sgi.com>
In some cases, for performance reasons flushing the SGI GRU cbe
needs to dirty the cacheline before doing the cacheline flush. This ensures
that the dirty cacheline actually goes back to the caching home agent.
Update a reserved field in the CBE instead of a field that might be used. This
_should_ not be required (dirtying any field should be ok) but early
silicon has an issue here.
Note that flushing the CL is not a requirement - only a performance enhancement.
Signed-off-by: Jack Steiner <steiner@sgi.com>
---
drivers/misc/sgi-gru/grufault.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
Index: linux/drivers/misc/sgi-gru/grufault.c
===================================================================
--- linux.orig/drivers/misc/sgi-gru/grufault.c 2010-06-09 08:11:40.285085853 -0500
+++ linux/drivers/misc/sgi-gru/grufault.c 2010-06-09 08:11:41.157110138 -0500
@@ -306,7 +306,7 @@ upm:
static void gru_flush_cache_cbe(struct gru_control_block_extended *cbe)
{
if (unlikely(cbe)) {
- cbe->cbrexecstatus = 0; /* make CL dirty */
+ cbe->reserved6 = 0; /* make CL dirty */
gru_flush_cache(cbe);
}
}
next prev parent reply other threads:[~2010-08-26 13:25 UTC|newest]
Thread overview: 27+ messages / expand[flat|nested] mbox.gz Atom feed top
2010-08-26 13:19 [Patch 00/25] GRU - GRU Updates - Production Driver steiner
2010-08-26 13:19 ` [Patch 01/25] GRU - delete obsolete gru instruction opcodes steiner
2010-08-26 13:19 ` [Patch 02/25] GRU - skip gru tlb purging of gru contexts:w steiner
2010-08-26 13:19 ` [Patch 03/25] GRU - update gru tlb miss statistics steiner
2010-08-26 13:19 ` [Patch 04/25] GRU - mmap gru contexts using nonlinear steiner
2010-08-26 13:19 ` steiner [this message]
2010-08-26 13:19 ` [Patch 06/25] GRU - change context stealing steiner
2010-08-26 13:19 ` [Patch 07/25] GRU - add context lock flag to gru status steiner
2010-08-26 13:19 ` [Patch 08/25] GRU - flush gru tlb when driver is loaded steiner
2010-08-26 13:19 ` [Patch 09/25] GRU - add software reserved bits to cbr definition steiner
2010-08-26 13:19 ` [Patch 10/25] GRU - eliminate gru contention on mmap_sem steiner
2010-08-26 13:19 ` [Patch 11/25] GRU - interrupt fix for processors without core 0 steiner
2010-08-26 13:19 ` [Patch 12/25] GRU - add gru hub number to context status steiner
2010-08-26 13:19 ` [Patch 13/25] GRU - delete obsolete debug code steiner
2010-08-26 13:19 ` [Patch 14/25] GRU - add polling for tlb misses steiner
2010-08-26 13:19 ` [Patch 15/25] GRU - reorder interrupt processing steiner
2010-08-26 13:19 ` [Patch 16/25] GRU - add refcnt to vdata structure steiner
2010-08-26 13:19 ` [Patch 17/25] GRU - no panic on gru malfunction steiner
2010-08-26 13:19 ` [Patch 18/25] GRU - contexts must contain cbrs steiner
2010-08-26 13:19 ` [Patch 19/25] GRU - update debug messages and comments steiner
2010-08-26 13:19 ` [Patch 20/25] GRU - add gsh information to gru dumps steiner
2010-08-26 13:19 ` [Patch 21/25] GRU - delete unused gru statistics structure steiner
2010-08-26 13:19 ` [Patch 22/25] GRU - gru api cleanup steiner
2010-08-26 13:20 ` [Patch 23/25] GRU - update driverr version steiner
2010-08-26 13:20 ` [Patch 24/25] GRU - rename gru pagesize defines steiner
2010-08-26 13:20 ` [Patch 25/25] GRU - update cbrstate definitions steiner
-- strict thread matches above, loose matches on Subject: below --
2010-07-19 21:32 [Patch 00/25] GRU - GRU Updates steiner
2010-07-19 21:32 ` [Patch 05/25] GRU - cbe cache flush steiner
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20100826132123.763139236@sgi.com \
--to=steiner@sgi.com \
--cc=akpm@linux-foundation.org \
--cc=linux-kernel@vger.kernel.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox