Brian King writes:
> Whenever multiple HRRQs are enabled, which is the default
> setting now, we end up seeing the following message logged
> prior to initialization of each HRRQ:
>
> Starting IOA initialization sequence
>
> This results in 16 of these messages on most adapters, which
> serves li
Brian King writes:
> A 9084 error gets logged by the ipr adapter when adapter raw mode
> gets enabled. A bunch of unformatted hex data also gets logged
> for this error, which is of little use, so let's avoid logging
> it by default in order to avoid the log getting polluted with
> useless data.
Brian King writes:
> Converts ipr to use the new EH API of libata, removing the
> last user of this API, allowing for its ultimate removal from
> the kernel.
>
> Sending this out only as an RFC for now. The patch will need
> significantly more testing, but I can now detect and initialize
> a SATA
Brian King writes:
> From: Wen Xiong
>
> Add scsi multi-queue support in ipr driver.
>
> Signed-off-by: Brian King
Patch looks good, very simple :)
Missing a sign-off from Wendy, though.
> ---
>
> drivers/scsi/ipr.c | 14 +-
> 1 file changed, 13 insertions(+), 1 deletion(-)
>
Brian King writes:
> Removes the free and pending ipr_cmd queues. This is the first patch in a
> series
> to make queuecommand lockless for the ipr driver. Creates a bitarray for each
> hardware queue so we can track whether an ipr_cmd struct is free or not. Also
> adds an atomic flag on each ip
Brian King writes:
> This patch removes all spinlocks from ipr's queuecommand path.
>
> Signed-off-by: Brian King
> ---
>
> drivers/scsi/ipr.c | 70
> +++--
> drivers/scsi/ipr.h | 52 +--
> 2 files changed,
Brian King writes:
> Export the following EH commands so that ipr can be converted to
> use libata's new EH.
>
lgtm.
--
Gabriel Krisman Bertazi
--
___
Iprdd-devel mailing li
Brian King writes:
> This converts ipr's queuecommand path to use percpu refcounts
> rather than atomics.
>
> Signed-off-by: Brian King
> ---
>
> drivers/scsi/ipr.c | 111
> ++---
> drivers/scsi/ipr.h | 53 +++--
> 2 files
Brian King writes:
> This adds a module parameter that, when set, will prevent
> ipr from attaching any SATA devices, and only attach SAS
> devices.
>
lgtm.
--
Gabriel Krisman Bertazi
--
_
Brian King writes:
> We can avoid grabbing any locks in the interrupt path for MSI-X interrupts
> and simply schedule the iopoll thread.
>
looks good.
--
Gabriel Krisman Bertazi
--
___
10 matches
Mail list logo
