Thank you...
Will this require a complete kernel recompile?
Or is there an easier way to include this patch?
I have little to no experience with working with patches at this level.
I've been doing research on how to use this patch, and I'm close, but, I'm not
sure if I need to attack compiling a kernel, or if I can somehow just patch
libata or something like that?
Thank you,
Mike
On Thu, Oct 6, 2011 at 12:48 AM, Gwendal Grignou <gwendal@xxxxxxxxxx> wrote:
>
> I think I know what is going on. One of your disks at least is slow to
> spinup. Due to a bug/feature in silicon image disk controller and pmp,
> at bring up we can not issue a SOFT_RESET and wait for the disk to
> spinup and then continue.
> That why we set ATA_LFLAG_NO_SRST in sata_pmp_quirks().
> So what happen is we go into a function that issue identify, but we
> fail, the disk is not ready [it is spinning up], so we retry.
> 3 times.
>
> From the first hard reset: 12888.470385, to the time you got the final
> error: 12901.397305 ~ 12.9s
> In the second case, your controller can send SOFT_RESET and wait for
> the device to respond.
> Time for the disk to spinup:
> 28010.630028 - 27997.097116 ~ 13.5s
> As you can see, you are borderline with the PMP, but the controller
> did not "wait" enough in the first case.
> Given the spinup time varies with drive, age, time since last
> spin-up..., it may work one day and fail the next.
> To work around the problem, I have a patch that consist of allowing
> the silicon image control to send a reset, but if it fails, we spin
> for a fixed amount of time and retry. This is not very nice, it is a
> better design to wait for event that waiting a fixed amount of time.
> You may have to alter ATA_LFLAG_WAIT_SRST to use the first bit available.
>
> Can you try with the following patch?
>
> Thanks,
> Gwendal.
>
> diff --git a/drivers/ata/libata-eh.c b/drivers/ata/libata-eh.c
> index 228740f..b98b02d 100644
> --- a/drivers/ata/libata-eh.c
> +++ b/drivers/ata/libata-eh.c
> @@ -2798,7 +2798,14 @@ int ata_eh_reset(struct ata_link *link, int classify,
> sata_scr_read(link, SCR_STATUS, &sstatus))
> rc = -ERESTART;
>
> - if (rc == -ERESTART || try >= max_tries)
> + if (try >= max_tries)
> + goto out;
> +
> + /* Some PMP will not serve SRST until the disk is spunup,
> + * if the controller can not wait for the PMP to acknowledge the frame,
> + * wait here */
> + if (rc == -ERESTART &&
> + !((lflags & ATA_LFLAG_WAIT_SRST) && (reset == softreset)))
> goto out;
>
> now = jiffies;
> @@ -2813,6 +2820,8 @@ int ata_eh_reset(struct ata_link *link, int classify,
> delta = schedule_timeout_uninterruptible(delta);
> }
>
> + if (rc == -ERESTART)
> + goto out;
> if (try == max_tries - 1) {
> sata_down_spd_limit(link, 0);
> if (slave)
> diff --git a/drivers/ata/libata-pmp.c b/drivers/ata/libata-pmp.c
> index 00305f4..d21ad7d 100644
> --- a/drivers/ata/libata-pmp.c
> +++ b/drivers/ata/libata-pmp.c
> @@ -325,13 +351,11 @@ static void sata_pmp_quirks(struct ata_port *ap)
> if (vendor == 0x1095 && devid == 0x3726) {
> /* sil3726 quirks */
> ata_for_each_link(link, ap, EDGE) {
> - /* Class code report is unreliable and SRST
> - * times out under certain configurations.
> - */
> + /* Class code report is unreliable */
> + /* PMP does not forward SRST until the drive spins up */
> if (link->pmp < 5)
> - link->flags |= ATA_LFLAG_NO_SRST |
> - ATA_LFLAG_ASSUME_ATA;
> -
> + link->flags |= ATA_LFLAG_ASSUME_ATA |
> + ATA_LFLAG_WAIT_SRST;
> /* port 5 is for SEMB device and it doesn't like SRST */
> if (link->pmp == 5)
> link->flags |= ATA_LFLAG_NO_SRST |
> diff --git a/include/linux/libata.h b/include/linux/libata.h
> index b2f2003..3a18caa 100644
> --- a/include/linux/libata.h
> +++ b/include/linux/libata.h
> @@ -172,6 +172,7 @@ enum {
> ATA_LFLAG_NO_RETRY = (1 << 5), /* don't retry this link */
> ATA_LFLAG_DISABLED = (1 << 6), /* link is disabled */
> ATA_LFLAG_SW_ACTIVITY = (1 << 7), /* keep activity stats */
> + ATA_LFLAG_WAIT_SRST = (1 << 8), /* add delay when SRST fails */
>
> /* struct ata_port flags */
> ATA_FLAG_SLAVE_POSS = (1 << 0), /* host supports slave dev */
>
>
--
To unsubscribe from this list: send the line "unsubscribe linux-ide" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
