[PATCH v2] scsi: introduce short_inquiry flag for broken host adapters

[PATCH v2] scsi: introduce short_inquiry flag for broken host adapters

[Vitaly Kuznetsov]

Some host adapters (e.g. Hyper-V storvsc) are known for not respecting the
SPC-2/3/4 requirement for 'INQUIRY data (see table ...) shall contain at
least 36 bytes'. As a result we get tons on 'scsi 0:7:1:1: scsi scan:
INQUIRY result too short (5), using 36' messages on console. This can be
problematic for slow consoles. Introduce short_inquiry host template flag
to avoid printing error messages for such adapters.

Signed-off-by: Vitaly Kuznetsov <vkuznets@redhat.com>
---
Changes since v1:
- This is a successor of previously sent "scsi_scan: move 'INQUIRY result
  too short' message to debug level" patch. Instead of moving the message
  to debug level for all adapters introduce a special 'short_inquiry' flag
  for host template [inspired by James Bottomley].
---
 drivers/scsi/scsi_scan.c   | 7 ++++---
 drivers/scsi/storvsc_drv.c | 1 +
 include/scsi/scsi_host.h   | 6 ++++++
 3 files changed, 11 insertions(+), 3 deletions(-)

diff --git a/drivers/scsi/scsi_scan.c b/drivers/scsi/scsi_scan.c
index f9f3f82..f1d00a0 100644
--- a/drivers/scsi/scsi_scan.c
+++ b/drivers/scsi/scsi_scan.c
@@ -701,9 +701,10 @@ static int scsi_probe_lun(struct scsi_device *sdev, unsigned char *inq_result,
 	 * strings.
 	 */
 	if (sdev->inquiry_len < 36) {
-		sdev_printk(KERN_INFO, sdev,
-			    "scsi scan: INQUIRY result too short (%d),"
-			    " using 36\n", sdev->inquiry_len);
+		if (!sdev->host->hostt->short_inquiry)
+			sdev_printk(KERN_INFO, sdev,
+				    "scsi scan: INQUIRY result too short (%d),"
+				    " using 36\n", sdev->inquiry_len);
 		sdev->inquiry_len = 36;
 	}
 
diff --git a/drivers/scsi/storvsc_drv.c b/drivers/scsi/storvsc_drv.c
index 3c6584f..f3b4d0f 100644
--- a/drivers/scsi/storvsc_drv.c
+++ b/drivers/scsi/storvsc_drv.c
@@ -1711,6 +1711,7 @@ static struct scsi_host_template scsi_driver = {
 	/* Make sure we dont get a sg segment crosses a page boundary */
 	.dma_boundary =		PAGE_SIZE-1,
 	.no_write_same =	1,
+	.short_inquiry =	1,
 };
 
 enum {
diff --git a/include/scsi/scsi_host.h b/include/scsi/scsi_host.h
index e113c75..7b022af 100644
--- a/include/scsi/scsi_host.h
+++ b/include/scsi/scsi_host.h
@@ -454,6 +454,12 @@ struct scsi_host_template {
 	unsigned no_async_abort:1;
 
 	/*
+	 * True if this host adapter returns short (<36 bytes) responses to
+	 * some INQUIRY requests.
+	 */
+	unsigned short_inquiry:1;
+
+	/*
 	 * Countdown for host blocking with no commands outstanding.
 	 */
 	unsigned int max_host_blocked;
-- 
2.4.3

Re: [PATCH v2] scsi: introduce short_inquiry flag for broken host adapters

[Vitaly Kuznetsov]

Vitaly Kuznetsov <vkuznets@redhat.com> writes:

> Some host adapters (e.g. Hyper-V storvsc) are known for not respecting the
> SPC-2/3/4 requirement for 'INQUIRY data (see table ...) shall contain at
> least 36 bytes'. As a result we get tons on 'scsi 0:7:1:1: scsi scan:
> INQUIRY result too short (5), using 36' messages on console. This can be
> problematic for slow consoles. Introduce short_inquiry host template flag
> to avoid printing error messages for such adapters.
>
> Signed-off-by: Vitaly Kuznetsov <vkuznets@redhat.com>
> ---
> Changes since v1:
> - This is a successor of previously sent "scsi_scan: move 'INQUIRY result
>   too short' message to debug level" patch. Instead of moving the message
>   to debug level for all adapters introduce a special 'short_inquiry' flag
>   for host template [inspired by James Bottomley].

James,

sorry for the ping but can you please let me know your opinion? This is
not a 'cosmetic fix', serial port on Hyper-V is extremely slow and users
get softlockups just because we output too much. Here is a freshly
booted guest with SCSI and FC adapters connected:

# dmesg | grep -c INQUIRY
2076

(my other pernding '[PATCH] scsi_scan: don't dump trace when
scsi_prep_async_scan() is called twice' is related to the same issue).

See also: https://lkml.org/lkml/2015/9/6/119

Thanks,

[...]

-- 
  Vitaly

Re: [PATCH v2] scsi: introduce short_inquiry flag for broken host adapters

[James Bottomley]

On Tue, 2015-09-15 at 16:27 +0200, Vitaly Kuznetsov wrote:
> Vitaly Kuznetsov <vkuznets@redhat.com> writes:
> 
> > Some host adapters (e.g. Hyper-V storvsc) are known for not respecting the
> > SPC-2/3/4 requirement for 'INQUIRY data (see table ...) shall contain at
> > least 36 bytes'. As a result we get tons on 'scsi 0:7:1:1: scsi scan:
> > INQUIRY result too short (5), using 36' messages on console. This can be
> > problematic for slow consoles. Introduce short_inquiry host template flag
> > to avoid printing error messages for such adapters.
> >
> > Signed-off-by: Vitaly Kuznetsov <vkuznets@redhat.com>
> > ---
> > Changes since v1:
> > - This is a successor of previously sent "scsi_scan: move 'INQUIRY result
> >   too short' message to debug level" patch. Instead of moving the message
> >   to debug level for all adapters introduce a special 'short_inquiry' flag
> >   for host template [inspired by James Bottomley].
> 
> James,
> 
> sorry for the ping but can you please let me know your opinion? This is
> not a 'cosmetic fix', serial port on Hyper-V is extremely slow and users
> get softlockups just because we output too much. Here is a freshly
> booted guest with SCSI and FC adapters connected:
> 
> # dmesg | grep -c INQUIRY
> 2076

It needs a review ... preferably from someone who validates that we
don't get any adverse impact from the missing inquiry data

> (my other pernding '[PATCH] scsi_scan: don't dump trace when
> scsi_prep_async_scan() is called twice' is related to the same issue).

So does this.

James


> See also: https://lkml.org/lkml/2015/9/6/119
> 
> Thanks,
> 
> [...]
> 

RE: [PATCH v2] scsi: introduce short_inquiry flag for broken host adapters

[KY Srinivasan]

> -----Original Message-----
> From: James Bottomley [mailto:James.Bottomley@HansenPartnership.com]
> Sent: Tuesday, September 15, 2015 11:24 AM
> To: Vitaly Kuznetsov <vkuznets@redhat.com>
> Cc: linux-scsi@vger.kernel.org; linux-kernel@vger.kernel.org; KY Srinivasan
> <kys@microsoft.com>; Long Li <longli@microsoft.com>; Dexuan Cui
> <decui@microsoft.com>
> Subject: Re: [PATCH v2] scsi: introduce short_inquiry flag for broken host
> adapters
> 
> On Tue, 2015-09-15 at 16:27 +0200, Vitaly Kuznetsov wrote:
> > Vitaly Kuznetsov <vkuznets@redhat.com> writes:
> >
> > > Some host adapters (e.g. Hyper-V storvsc) are known for not respecting
> the
> > > SPC-2/3/4 requirement for 'INQUIRY data (see table ...) shall contain at
> > > least 36 bytes'. As a result we get tons on 'scsi 0:7:1:1: scsi scan:
> > > INQUIRY result too short (5), using 36' messages on console. This can be
> > > problematic for slow consoles. Introduce short_inquiry host template flag
> > > to avoid printing error messages for such adapters.
> > >
> > > Signed-off-by: Vitaly Kuznetsov <vkuznets@redhat.com>
> > > ---
> > > Changes since v1:
> > > - This is a successor of previously sent "scsi_scan: move 'INQUIRY result
> > >   too short' message to debug level" patch. Instead of moving the
> message
> > >   to debug level for all adapters introduce a special 'short_inquiry' flag
> > >   for host template [inspired by James Bottomley].
> >
> > James,
> >
> > sorry for the ping but can you please let me know your opinion? This is
> > not a 'cosmetic fix', serial port on Hyper-V is extremely slow and users
> > get softlockups just because we output too much. Here is a freshly
> > booted guest with SCSI and FC adapters connected:
> >
> > # dmesg | grep -c INQUIRY
> > 2076
> 
> It needs a review ... preferably from someone who validates that we
> don't get any adverse impact from the missing inquiry data
> 
> > (my other pernding '[PATCH] scsi_scan: don't dump trace when
> > scsi_prep_async_scan() is called twice' is related to the same issue).
> 
> So does this.

James,

I will review the code and test this as well.

Regards,

K. Y
> 
> James
> 
> 
> > See also:
> https://na01.safelinks.protection.outlook.com/?url=https%3a%2f%2flkml.or
> g%2flkml%2f2015%2f9%2f6%2f119&data=01%7c01%7ckys%40microsoft.com
> %7c7a6e9ccabb2f43c0a01108d2bdfad826%7c72f988bf86f141af91ab2d7cd011
> db47%7c1&sdata=t3IGLMiG2sU4VJfmGiYKGMy22d3IxHbGQZop63nrV7Y%3d
> >
> > Thanks,
> >
> > [...]
> >
> 
>