All of lore.kernel.org
 help / color / mirror / Atom feed
From: Anuj Gupta <anuj20.g@samsung.com>
To: Anuj gupta <anuj1072538@gmail.com>
Cc: "Martin K. Petersen" <martin.petersen@oracle.com>,
	Christoph Hellwig <hch@lst.de>, M Nikhil <nikh1092@linux.ibm.com>,
	linux-block@vger.kernel.org, dm-devel@lists.linux.dev,
	linux-raid@vger.kernel.org, nvdimm@lists.linux.dev,
	linux-scsi@vger.kernel.org, hare@suse.de,
	steffen Maier <maier@linux.ibm.com>,
	Benjamin Block <bblock@linux.ibm.com>,
	Nihar Panda <niharp@linux.ibm.com>
Subject: Re: Change in reported values of some block integrity sysfs attributes
Date: Fri, 21 Feb 2025 17:37:29 +0530	[thread overview]
Message-ID: <20250221120729.GA5233@green245> (raw)
In-Reply-To: <CACzX3AvbM4qG+ZOWJoCTNMMgSz8gMjoRcQ10_HJbMyi0Nv9qvQ@mail.gmail.com>

[-- Attachment #1: Type: text/plain, Size: 2310 bytes --]

On Fri, Feb 21, 2025 at 04:07:55PM +0530, Anuj gupta wrote:
> > I don't see any change in what's reported with block/for-next in a
> > regular SCSI HBA/disk setup. Will have to look at whether there is a
> > stacking issue wrt. multipathing.
> 
> Hi Martin, Christoph,
> 
> It seems this change in behaviour is not limited to SCSI only. As Nikhil
> mentioned an earlier commit
> [9f4aa46f2a74 ("block: invert the BLK_INTEGRITY_{GENERATE,VERIFY} flags")]
> causes this change in behaviour. On my setup with a NVMe drive not formatted
> with PI, I see that:
> 
> Without this commit:
> Value reported by read_verify and write_generate sysfs entries is 0.
> 
> With this commit:
> Value reported by read_verify and write_generate sysfs entries is 1.
> 
> Diving a bit deeper, both these flags got inverted due to this commit.
> But during init (in nvme_init_integrity) these values get initialized to 0,
> inturn setting the sysfs entries to 1. In order to fix this, the driver has to
> initialize both these flags to 1 in nvme_init_integrity if PI is not supported.
> That way, the value in sysfs for these entries would become 0 again. Tried this
> approach in my setup, and I am able to see the right values now. Then something
> like this would also need to be done for SCSI too.
> 

I tried to make it work for SCSI too. However my testing is limited as I
don't have a SCSI device. With scsi_debug I see this currently:

# modprobe scsi_debug dev_size_mb=128 dix=0 dif=0
# cat /sys/block/sda/integrity/write_generate
1
# cat /sys/block/sda/integrity/read_verify
1
# cat /sys/class/scsi_host/host0/prot_capabilities
0

To fix this, I added this. Nikhil can you try below patch? Martin, can
you please take a look as well.

After this patch, with the same scsi_debug device, I see sysfs entries
populated as 0.

diff --git a/drivers/scsi/sd_dif.c b/drivers/scsi/sd_dif.c
index ae6ce6f5d622..be2cd06f500b 100644
--- a/drivers/scsi/sd_dif.c
+++ b/drivers/scsi/sd_dif.c
@@ -40,8 +40,10 @@ void sd_dif_config_host(struct scsi_disk *sdkp, struct queue_limits *lim)
 		dif = 0; dix = 1;
 	}
 
-	if (!dix)
+	if (!dix) {
+		bi->flags |= BLK_INTEGRITY_NOGENERATE | BLK_INTEGRITY_NOVERIFY;
 		return;
+	}
 
 	/* Enable DMA of protection information */
 	if (scsi_host_get_guard(sdkp->device->host) & SHOST_DIX_GUARD_IP)

[-- Attachment #2: Type: text/plain, Size: 0 bytes --]



  parent reply	other threads:[~2025-02-21 12:16 UTC|newest]

Thread overview: 10+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2024-12-13  7:16 Change in reported values of some block integrity sysfs attributes M Nikhil
2024-12-13 14:33 ` Christoph Hellwig
2024-12-17 12:46   ` M Nikhil
2024-12-13 22:30 ` Martin K. Petersen
2024-12-17 12:39   ` M Nikhil
2025-02-21  6:01     ` M Nikhil
2025-02-21 10:37   ` Anuj gupta
2025-02-21 10:35     ` Anuj Gupta
2025-02-21 12:07     ` Anuj Gupta [this message]
2025-02-25  1:07       ` Martin K. Petersen

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20250221120729.GA5233@green245 \
    --to=anuj20.g@samsung.com \
    --cc=anuj1072538@gmail.com \
    --cc=bblock@linux.ibm.com \
    --cc=dm-devel@lists.linux.dev \
    --cc=hare@suse.de \
    --cc=hch@lst.de \
    --cc=linux-block@vger.kernel.org \
    --cc=linux-raid@vger.kernel.org \
    --cc=linux-scsi@vger.kernel.org \
    --cc=maier@linux.ibm.com \
    --cc=martin.petersen@oracle.com \
    --cc=niharp@linux.ibm.com \
    --cc=nikh1092@linux.ibm.com \
    --cc=nvdimm@lists.linux.dev \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.