PATCH 1/8
This patch changes the cciss version number to 3.6.14 to reflect the following
functionality changes added by the rest of the set. They include:
Support to fire up on any HP RAID class controller
Increase nr_cmds to 512 for most controllers by adding it to the product table
PCI subsystem ID fix fix was pulled
Disable DMA prefetch for the P600 on IPF platforms
Change from 512 to 2048 sector_size for performance
Fix in cciss_open for consistency
Remove the no longer used revalidate_allvol function
Please consider this for inclusion.
Thanks,
mikem
Signed-off-by: Mike Miller <[email protected]>
cciss.c | 7 ++++---
1 files changed, 4 insertions(+), 3 deletions(-)
--------------------------------------------------------------------------------
diff -urNp linux-2.6.orig/drivers/block/cciss.c linux-2.6/drivers/block/cciss.c
--- linux-2.6.orig/drivers/block/cciss.c 2006-10-27 11:16:05.000000000 -0500
+++ linux-2.6/drivers/block/cciss.c 2006-10-31 14:01:04.000000000 -0600
@@ -47,14 +47,15 @@
#include <linux/completion.h>
#define CCISS_DRIVER_VERSION(maj,min,submin) ((maj<<16)|(min<<8)|(submin))
-#define DRIVER_NAME "HP CISS Driver (v 3.6.10)"
-#define DRIVER_VERSION CCISS_DRIVER_VERSION(3,6,10)
+#define DRIVER_NAME "HP CISS Driver (v 3.6.14)"
+#define DRIVER_VERSION CCISS_DRIVER_VERSION(3,6,14)
/* Embedded module documentation macros - see modules.h */
MODULE_AUTHOR("Hewlett-Packard Company");
-MODULE_DESCRIPTION("Driver for HP Controller SA5xxx SA6xxx version 3.6.10");
+MODULE_DESCRIPTION("Driver for HP Controller SA5xxx SA6xxx version 3.6.14");
MODULE_SUPPORTED_DEVICE("HP SA5i SA5i+ SA532 SA5300 SA5312 SA641 SA642 SA6400"
" SA6i P600 P800 P400 P400i E200 E200i E500");
+MODULE_VERSION("3.6.14");
MODULE_LICENSE("GPL");
#include "cciss_cmd.h"
On Wed, Nov 01 2006, Mike Miller (OS Dev) wrote:
>
> PATCH 1/8
>
> This patch changes the cciss version number to 3.6.14 to reflect the following
> functionality changes added by the rest of the set. They include:
Mike, only some of your patches appeared to go out, both in personal
mail and on the list.
Comment on this patch - normally you'd change the version number last,
not as the first patch.
--
Jens Axboe
On Thu, Nov 02, 2006 at 03:10:45PM +0100, Jens Axboe wrote:
> On Wed, Nov 01 2006, Mike Miller (OS Dev) wrote:
> >
> > PATCH 1/8
> >
> > This patch changes the cciss version number to 3.6.14 to reflect the following
> > functionality changes added by the rest of the set. They include:
>
> Mike, only some of your patches appeared to go out, both in personal
> mail and on the list.
I ran into some last minute issues so I stopped at 5. Still re-testing
the others.
This snippet seems to tbe the culprit.
+ if (blk_queue_stopped(h->gendisk[curr_queue]->queue) ||
+ blk_queue_plugged(h->gendisk[curr_queue]->queue))
+ blk_start_queue(h->gendisk[curr_queue]->queue);
We're testing to see if the queue is stopped or plugged so we don't
try to start am already running queue. Without the blk_queue_plugged
test it hangs every time. We added blk_queue_plugged and the first tests
seem to run ok. Then at the last minute something broke. Does this look
ok to you?
>
> Comment on this patch - normally you'd change the version number last,
> not as the first patch.
Yep, put the cart before horse I suppose. Maybe I shouldn't have a
version number at all. :)
mikem
On Thu, 2006-11-02 at 08:46 -0600, Mike Miller (OS Dev) wrote:
> On Thu, Nov 02, 2006 at 03:10:45PM +0100, Jens Axboe wrote:
> > On Wed, Nov 01 2006, Mike Miller (OS Dev) wrote:
> > >
> > > PATCH 1/8
> > >
> > > This patch changes the cciss version number to 3.6.14 to reflect the following
> > > functionality changes added by the rest of the set. They include:
> >
> > Mike, only some of your patches appeared to go out, both in personal
> > mail and on the list.
>
> I ran into some last minute issues so I stopped at 5. Still re-testing
> the others.
> This snippet seems to tbe the culprit.
> + if (blk_queue_stopped(h->gendisk[curr_queue]->queue) ||
> + blk_queue_plugged(h->gendisk[curr_queue]->queue))
> + blk_start_queue(h->gendisk[curr_queue]->queue);
>
> We're testing to see if the queue is stopped or plugged so we don't
> try to start am already running queue. Without the blk_queue_plugged
> test it hangs every time. We added blk_queue_plugged and the first tests
> seem to run ok. Then at the last minute something broke. Does this look
> ok to you?
it looks like a design mistake to me if a device driver needs to care
about a queue being plugged at all....
On Thu, Nov 02, 2006 at 04:05:14PM +0100, Arjan van de Ven wrote:
> > This snippet seems to tbe the culprit.
> > + if (blk_queue_stopped(h->gendisk[curr_queue]->queue) ||
> > + blk_queue_plugged(h->gendisk[curr_queue]->queue))
> > + blk_start_queue(h->gendisk[curr_queue]->queue);
> >
> > We're testing to see if the queue is stopped or plugged so we don't
> > try to start am already running queue. Without the blk_queue_plugged
> > test it hangs every time. We added blk_queue_plugged and the first tests
> > seem to run ok. Then at the last minute something broke. Does this look
> > ok to you?
>
>
> it looks like a design mistake to me if a device driver needs to care
> about a queue being plugged at all....
We getting hw soon that will support up to 1024 logical volumes. We thought
the test would actually save time with many volumes. Maybe not.
mikem
On Thu, Nov 02 2006, Arjan van de Ven wrote:
> On Thu, 2006-11-02 at 08:46 -0600, Mike Miller (OS Dev) wrote:
> > On Thu, Nov 02, 2006 at 03:10:45PM +0100, Jens Axboe wrote:
> > > On Wed, Nov 01 2006, Mike Miller (OS Dev) wrote:
> > > >
> > > > PATCH 1/8
> > > >
> > > > This patch changes the cciss version number to 3.6.14 to reflect the following
> > > > functionality changes added by the rest of the set. They include:
> > >
> > > Mike, only some of your patches appeared to go out, both in personal
> > > mail and on the list.
> >
> > I ran into some last minute issues so I stopped at 5. Still re-testing
> > the others.
> > This snippet seems to tbe the culprit.
> > + if (blk_queue_stopped(h->gendisk[curr_queue]->queue) ||
> > + blk_queue_plugged(h->gendisk[curr_queue]->queue))
> > + blk_start_queue(h->gendisk[curr_queue]->queue);
> >
> > We're testing to see if the queue is stopped or plugged so we don't
> > try to start am already running queue. Without the blk_queue_plugged
> > test it hangs every time. We added blk_queue_plugged and the first tests
> > seem to run ok. Then at the last minute something broke. Does this look
> > ok to you?
>
>
> it looks like a design mistake to me if a device driver needs to care
> about a queue being plugged at all....
The queue can't be plugged and stopped at the same time anyway, so the
code is confusing. But yes I agree, that should go.
--
Jens Axboe