Currently, libata doesn't support link powersave modes and always
disables link PS modes. However, if the link is already in powersave
mode when libata begis probing, inhibiting powersave modes by setting
IPM isn't enough. Link should be manually transited into active mode
using SPM or hardreset.
As SPM isn't avalable on all controllers and we're gonna prefer
hardreset in future anyway, force hardreset if link is in powersave
mode.
Reported-by: Paul Collins <paul@burly.ondioline.org>
Signed-off-by: Tejun Heo <htejun@gmail.com>
---
Jeff, this isn't necessary for #upstream.
drivers/ata/libata-core.c | 15 +++++++++++++++
1 file changed, 15 insertions(+)
index be95fdb..e3ddefd 100644
--- a/drivers/ata/libata-core.c
+++ b/drivers/ata/libata-core.c
@@ -3966,11 +3966,26 @@ int ata_std_prereset(struct ata_link *link, unsigned long deadline)
/* if SATA, resume link */
if (ap->flags & ATA_FLAG_SATA) {
+ u32 sstatus;
+
rc = sata_link_resume(link, timing, deadline);
/* whine about phy resume failure but proceed */
if (rc && rc != -EOPNOTSUPP)
ata_link_printk(link, KERN_WARNING, "failed to resume "
"link for reset (errno=%d)\n", rc);
+
+ /* force hardreset if the link is sleeping */
+ if (sata_scr_read(link, SCR_STATUS, &sstatus) == 0) {
+ u8 ipm = sstatus >> 8;
+
+ if (ipm == 2 || ipm == 6) {
+ ata_link_printk(link, KERN_INFO,
+ "link in %s mode, forcing hardreset\n",
+ ipm == 2 ? "partial" : "slumber");
+ ehc->i.action |= ATA_EH_HARDRESET;
+ return 0;
+ }
+ }
}
/* Wait for !BSY if the controller can wait for the first D2H
--
| Greg Kroah-Hartman | [PATCH 004/196] Chinese: add translation of SubmittingPatches |
| Alan Stern | Re: 2.6.22-rc2-mm1 |
| Satyam Sharma | Re: [PATCH 0/24] make atomic_read() behave consistently across all architectures |
| William Lee Irwin III | Re: [Announce] [patch] Modular Scheduler Core and Completely Fair Scheduler [CFS] |
git: | |
| Dale Farnsworth | Re: [PATCH 03/39] mv643xx_eth: shorten reg names |
| Jarek Poplawski | Re: HTB accuracy for high speed |
| Gerrit Renker | [PATCH 15/37] dccp: Set per-connection CCIDs via socket options |
| Jarek Poplawski | [PATCH] pkt_sched: Destroy gen estimators under rtnl_lock(). |
