Patchwork [v2,1/3] ahci: add AHCI_HFLAG_DELAY_ENGINE host flag

login
register
mail settings
Submitter Brian Norris
Date Feb. 21, 2012, 6:38 p.m.
Message ID <1329849524-23758-2-git-send-email-computersforpeace@gmail.com>
Download mbox | patch
Permalink /patch/142348/
State Not Applicable
Delegated to: David Miller
Headers show

Comments

Brian Norris - Feb. 21, 2012, 6:38 p.m.
The following commit was intended to fix problems with specific AHCI
controller(s) that would become bricks if the AHCI specification was not
followed strictly (that is, if ahci_start_engine() was called while the
controller was in the wrong state):

    commit 7faa33da9b7add01db9f1ad92c6a5d9145e940a7
    ahci: start engine only during soft/hard resets

However, some devices currently have issues with that fix, so we must
implement a flag that delays the ahci_start_engine() call only for specific
controllers.

This commit simply introduces the flag, without enabling it in any driver.

Note that even when AHCI_HFLAG_DELAY_ENGINE is not enabled, this patch does
not constitue a full revert to commit 7faa33da; there is still a change in
behavior to the ahci_port_suspend() failure path.

Signed-off-by: Brian Norris <computersforpeace@gmail.com>
---
 drivers/ata/ahci.h    |    3 +++
 drivers/ata/libahci.c |    5 +++++
 2 files changed, 8 insertions(+), 0 deletions(-)

--
To unsubscribe from this list: send the line "unsubscribe linux-ide" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Jeff Garzik - March 13, 2012, 8:36 p.m.
On 02/21/2012 01:38 PM, Brian Norris wrote:
> The following commit was intended to fix problems with specific AHCI
> controller(s) that would become bricks if the AHCI specification was not
> followed strictly (that is, if ahci_start_engine() was called while the
> controller was in the wrong state):
>
>      commit 7faa33da9b7add01db9f1ad92c6a5d9145e940a7
>      ahci: start engine only during soft/hard resets
>
> However, some devices currently have issues with that fix, so we must
> implement a flag that delays the ahci_start_engine() call only for specific
> controllers.
>
> This commit simply introduces the flag, without enabling it in any driver.
>
> Note that even when AHCI_HFLAG_DELAY_ENGINE is not enabled, this patch does
> not constitue a full revert to commit 7faa33da; there is still a change in
> behavior to the ahci_port_suspend() failure path.
>
> Signed-off-by: Brian Norris<computersforpeace@gmail.com>

applied 1-3



--
To unsubscribe from this list: send the line "unsubscribe linux-ide" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Patch

diff --git a/drivers/ata/ahci.h b/drivers/ata/ahci.h
index b175000..feb127e 100644
--- a/drivers/ata/ahci.h
+++ b/drivers/ata/ahci.h
@@ -210,6 +210,9 @@  enum {
 	AHCI_HFLAG_NO_SNTF		= (1 << 12), /* no sntf */
 	AHCI_HFLAG_NO_FPDMA_AA		= (1 << 13), /* no FPDMA AA */
 	AHCI_HFLAG_YES_FBS		= (1 << 14), /* force FBS cap on */
+	AHCI_HFLAG_DELAY_ENGINE		= (1 << 15), /* do not start engine on
+						        port start (wait until
+						        error-handling stage) */
 
 	/* ap->flags bits */
 
diff --git a/drivers/ata/libahci.c b/drivers/ata/libahci.c
index a72bfd0..f9eaa82 100644
--- a/drivers/ata/libahci.c
+++ b/drivers/ata/libahci.c
@@ -737,6 +737,7 @@  static void ahci_power_down(struct ata_port *ap)
 
 static void ahci_start_port(struct ata_port *ap)
 {
+	struct ahci_host_priv *hpriv = ap->host->private_data;
 	struct ahci_port_priv *pp = ap->private_data;
 	struct ata_link *link;
 	struct ahci_em_priv *emp;
@@ -746,6 +747,10 @@  static void ahci_start_port(struct ata_port *ap)
 	/* enable FIS reception */
 	ahci_start_fis_rx(ap);
 
+	/* enable DMA */
+	if (!(hpriv->flags & AHCI_HFLAG_DELAY_ENGINE))
+		ahci_start_engine(ap);
+
 	/* turn on LEDs */
 	if (ap->flags & ATA_FLAG_EM) {
 		ata_for_each_link(link, ap, EDGE) {