Patchwork libata-pmp.c patch for Silicon Image 3826 port multiplier

login
register
mail settings
Submitter Terry Suereth
Date Jan. 13, 2013, 9:58 p.m.
Message ID <CANQSiW0f4A=Y3db0-=2fdiHbGO2SigOE1ejPTYsUcZHhpEg1pQ@mail.gmail.com>
Download mbox | patch
Permalink /patch/211656/
State Not Applicable
Delegated to: David Miller
Headers show

Comments

Terry Suereth - Jan. 13, 2013, 9:58 p.m.
Hello linux-ide mailing list,

I recently purchased and equipped one of my Slackware boxes with a
Rosewill RAID enclosure (RSV-S8), which claimed to use a Sil3726 PMP
chipset, but which actually uses a Sil3826 chipset.  This appears to
be some relatively new silicon, and (presumably) a replacement for the
3726 going forward.

I had some difficulties using this hardware at first - excessively
long delays linking to the PMP at startup - but once I discovered the
chipset version issue, it was simple to fix -- libata just needs to
apply the same PMP quirks that are currently used on the 3726, to the
3826.  I've done this in a local kernel build and it's working great
so far.

The same bug/fix has been posted to the Red Hat bug tracker:
https://bugzilla.redhat.com/show_bug.cgi?id=890237 -- although the
patch noted there (in the PDF (!?) attachment) appears incomplete.

My somewhat-redundant patch, based on the current linux-stable version
from git, is pasted inline below.  It should be very simple and
unsurprising.

Best,

Terry Suereth
terry.suereth@gmail.com



--
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
Sergei Shtylyov - Jan. 14, 2013, 12:23 p.m.
Hello.

On 14-01-2013 1:58, Terry Suereth wrote:

> Hello linux-ide mailing list,

> I recently purchased and equipped one of my Slackware boxes with a
> Rosewill RAID enclosure (RSV-S8), which claimed to use a Sil3726 PMP
> chipset, but which actually uses a Sil3826 chipset.  This appears to
> be some relatively new silicon, and (presumably) a replacement for the
> 3726 going forward.

> I had some difficulties using this hardware at first - excessively
> long delays linking to the PMP at startup - but once I discovered the
> chipset version issue, it was simple to fix -- libata just needs to
> apply the same PMP quirks that are currently used on the 3726, to the
> 3826.  I've done this in a local kernel build and it's working great
> so far.

> The same bug/fix has been posted to the Red Hat bug tracker:
> https://bugzilla.redhat.com/show_bug.cgi?id=890237 -- although the
> patch noted there (in the PDF (!?) attachment) appears incomplete.

> My somewhat-redundant patch, based on the current linux-stable version
> from git, is pasted inline below.  It should be very simple and
> unsurprising.

> Best,

> Terry Suereth
> terry.suereth@gmail.com



> --- libata-pmp.c.old 2013-01-13 13:42:56.582780772 -0800
> +++ libata-pmp.c 2013-01-13 13:43:33.670964680 -0800
> @@ -289,24 +289,24 @@
>
>    /* Disable sending Early R_OK.
>    * With "cached read" HDD testing and multiple ports busy on a SATA
> - * host controller, 3726 PMP will very rarely drop a deferred
> + * host controller, 3x26 PMP will very rarely drop a deferred
>    * R_OK that was intended for the host. Symptom will be all
>    * 5 drives under test will timeout, get reset, and recover.
>    */
> - if (vendor == 0x1095 && devid == 0x3726) {
> + if (vendor == 0x1095 && (devid == 0x3726 || devid == 0x3826)) {
>    u32 reg;
>
>    err_mask = sata_pmp_read(&ap->link, PMP_GSCR_SII_POL, &reg);
>    if (err_mask) {
>    rc = -EIO;
> - reason = "failed to read Sil3726 Private Register";
> + reason = "failed to read Sil3x26 Private Register";
>    goto fail;
>    }
>    reg &= ~0x1;
>    err_mask = sata_pmp_write(&ap->link, PMP_GSCR_SII_POL, reg);
>    if (err_mask) {
>    rc = -EIO;
> - reason = "failed to write Sil3726 Private Register";
> + reason = "failed to write Sil3x26 Private Register";
>    goto fail;
>    }
>    }
> @@ -383,8 +383,8 @@
>    u16 devid = sata_pmp_gscr_devid(gscr);
>    struct ata_link *link;
>
> - if (vendor == 0x1095 && devid == 0x3726) {
> - /* sil3726 quirks */
> + if (vendor == 0x1095 && (devid == 0x3726 || devid == 0x3826)) {
> + /* sil3x26 quirks */
>    ata_for_each_link(link, ap, EDGE) {
>    /* link reports offline after LPM */
>    link->flags |= ATA_LFLAG_NO_LPM;

    Your patch is whitespace damaged, misses signoff and proper description.

MBR, Sergei

--
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
Terry Suereth - Jan. 14, 2013, 8:56 p.m.
Thanks for the tips Sergei -- I'm totally in the dark re: kernel patch
submission process.  I bothered to read the SubmittingPatches page
this morning, and will be sending a more properly-formatted patch
email/subject to the list shortly.


On Mon, Jan 14, 2013 at 4:23 AM, Sergei Shtylyov <sshtylyov@mvista.com> wrote:
> Hello.
>
>
> On 14-01-2013 1:58, Terry Suereth wrote:
>
>> Hello linux-ide mailing list,
>
>
>> I recently purchased and equipped one of my Slackware boxes with a
>> Rosewill RAID enclosure (RSV-S8), which claimed to use a Sil3726 PMP
>> chipset, but which actually uses a Sil3826 chipset.  This appears to
>> be some relatively new silicon, and (presumably) a replacement for the
>> 3726 going forward.
>
>
>> I had some difficulties using this hardware at first - excessively
>> long delays linking to the PMP at startup - but once I discovered the
>> chipset version issue, it was simple to fix -- libata just needs to
>> apply the same PMP quirks that are currently used on the 3726, to the
>> 3826.  I've done this in a local kernel build and it's working great
>> so far.
>
>
>> The same bug/fix has been posted to the Red Hat bug tracker:
>> https://bugzilla.redhat.com/show_bug.cgi?id=890237 -- although the
>> patch noted there (in the PDF (!?) attachment) appears incomplete.
>
>
>> My somewhat-redundant patch, based on the current linux-stable version
>> from git, is pasted inline below.  It should be very simple and
>> unsurprising.
>
>
>> Best,
>
>
>> Terry Suereth
>> terry.suereth@gmail.com
>
>
>
>
>> --- libata-pmp.c.old 2013-01-13 13:42:56.582780772 -0800
>> +++ libata-pmp.c 2013-01-13 13:43:33.670964680 -0800
>> @@ -289,24 +289,24 @@
>>
>>    /* Disable sending Early R_OK.
>>    * With "cached read" HDD testing and multiple ports busy on a SATA
>> - * host controller, 3726 PMP will very rarely drop a deferred
>> + * host controller, 3x26 PMP will very rarely drop a deferred
>>    * R_OK that was intended for the host. Symptom will be all
>>    * 5 drives under test will timeout, get reset, and recover.
>>    */
>> - if (vendor == 0x1095 && devid == 0x3726) {
>> + if (vendor == 0x1095 && (devid == 0x3726 || devid == 0x3826)) {
>>    u32 reg;
>>
>>    err_mask = sata_pmp_read(&ap->link, PMP_GSCR_SII_POL, &reg);
>>    if (err_mask) {
>>    rc = -EIO;
>> - reason = "failed to read Sil3726 Private Register";
>> + reason = "failed to read Sil3x26 Private Register";
>>    goto fail;
>>    }
>>    reg &= ~0x1;
>>    err_mask = sata_pmp_write(&ap->link, PMP_GSCR_SII_POL, reg);
>>    if (err_mask) {
>>    rc = -EIO;
>> - reason = "failed to write Sil3726 Private Register";
>> + reason = "failed to write Sil3x26 Private Register";
>>    goto fail;
>>    }
>>    }
>> @@ -383,8 +383,8 @@
>>    u16 devid = sata_pmp_gscr_devid(gscr);
>>    struct ata_link *link;
>>
>> - if (vendor == 0x1095 && devid == 0x3726) {
>> - /* sil3726 quirks */
>> + if (vendor == 0x1095 && (devid == 0x3726 || devid == 0x3826)) {
>> + /* sil3x26 quirks */
>>    ata_for_each_link(link, ap, EDGE) {
>>    /* link reports offline after LPM */
>>    link->flags |= ATA_LFLAG_NO_LPM;
>
>
>    Your patch is whitespace damaged, misses signoff and proper description.
>
> MBR, Sergei
>
--
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

--- libata-pmp.c.old 2013-01-13 13:42:56.582780772 -0800
+++ libata-pmp.c 2013-01-13 13:43:33.670964680 -0800
@@ -289,24 +289,24 @@ 

  /* Disable sending Early R_OK.
  * With "cached read" HDD testing and multiple ports busy on a SATA
- * host controller, 3726 PMP will very rarely drop a deferred
+ * host controller, 3x26 PMP will very rarely drop a deferred
  * R_OK that was intended for the host. Symptom will be all
  * 5 drives under test will timeout, get reset, and recover.
  */
- if (vendor == 0x1095 && devid == 0x3726) {
+ if (vendor == 0x1095 && (devid == 0x3726 || devid == 0x3826)) {
  u32 reg;

  err_mask = sata_pmp_read(&ap->link, PMP_GSCR_SII_POL, &reg);
  if (err_mask) {
  rc = -EIO;
- reason = "failed to read Sil3726 Private Register";
+ reason = "failed to read Sil3x26 Private Register";
  goto fail;
  }
  reg &= ~0x1;
  err_mask = sata_pmp_write(&ap->link, PMP_GSCR_SII_POL, reg);
  if (err_mask) {
  rc = -EIO;
- reason = "failed to write Sil3726 Private Register";
+ reason = "failed to write Sil3x26 Private Register";
  goto fail;
  }
  }
@@ -383,8 +383,8 @@ 
  u16 devid = sata_pmp_gscr_devid(gscr);
  struct ata_link *link;

- if (vendor == 0x1095 && devid == 0x3726) {
- /* sil3726 quirks */
+ if (vendor == 0x1095 && (devid == 0x3726 || devid == 0x3826)) {
+ /* sil3x26 quirks */
  ata_for_each_link(link, ap, EDGE) {
  /* link reports offline after LPM */
  link->flags |= ATA_LFLAG_NO_LPM;