diff mbox series

[2/2] scsi: target: iscsi: don't warn of R/W when no data

Message ID 437f863520874ee386b6882ef749bf8d988839ca.1701540918.git.lduncan@suse.com
State New
Headers show
Series scsi: target: iscsi: Fix two protocol issues | expand

Commit Message

Lee Duncan Dec. 7, 2023, 5:42 p.m. UTC 
From: Lee Duncan <lduncan@suse.com>

The LIO target code has a warning about setting the
read and/or write header bits with a PDU that has zero
transfer length, even though the code mentions that the
SPEC (RFC 3720) allows this, and that some initiators
set these bits. But in practice such initiators end up
flooding the logs with thousands of warning messages for
IO that is allowed.

So change this to a debug message, and clean up the wording
just a little bit while at it.

Signed-off-by: Lee Duncan <lduncan@suse.com>
Reviewed-by: David Bond <dbond@suse.com>
---
 drivers/target/iscsi/iscsi_target.c | 5 +++--
 1 file changed, 3 insertions(+), 2 deletions(-)

Comments

Lee Duncan March 9, 2024, 6:05 p.m. UTC | #1 
I didn't see any objections to this patch.

Are there any issues?

On Sat, Dec 2, 2023 at 10:45 AM lduncan@suse.com <lduncan@suse.com> wrote:
>
> From: Lee Duncan <lduncan@suse.com>
>
> The LIO target code has a warning about setting the
> read and/or write header bits with a PDU that has zero
> transfer length, even though the code mentions that the
> SPEC (RFC 3720) allows this, and that some initiators
> set these bits. But in practice such initiators end up
> flooding the logs with thousands of warning messages for
> IO that is allowed.
>
> So change this to a debug message, and clean up the wording
> just a little bit while at it.
>
> Signed-off-by: Lee Duncan <lduncan@suse.com>
> Reviewed-by: David Bond <dbond@suse.com>
> ---
>  drivers/target/iscsi/iscsi_target.c | 5 +++--
>  1 file changed, 3 insertions(+), 2 deletions(-)
>
> diff --git a/drivers/target/iscsi/iscsi_target.c b/drivers/target/iscsi/iscsi_target.c
> index f246e5015868..c82dc2cd08b3 100644
> --- a/drivers/target/iscsi/iscsi_target.c
> +++ b/drivers/target/iscsi/iscsi_target.c
> @@ -1039,9 +1039,10 @@ int iscsit_setup_scsi_cmd(struct iscsit_conn *conn, struct iscsit_cmd *cmd,
>                 hdr->flags &= ~ISCSI_FLAG_CMD_READ;
>                 hdr->flags &= ~ISCSI_FLAG_CMD_WRITE;
>
> -               pr_warn("ISCSI_FLAG_CMD_READ or ISCSI_FLAG_CMD_WRITE"
> +               pr_debug("ISCSI_FLAG_CMD_READ or ISCSI_FLAG_CMD_WRITE"
>                         " set when Expected Data Transfer Length is 0 for"
> -                       " CDB: 0x%02x, Fixing up flags\n", hdr->cdb[0]);
> +                       " CDB: 0x%02x, cleared READ/WRITE flag(s)\n",
> +                       hdr->cdb[0]);
>         }
>
>         if (!(hdr->flags & ISCSI_FLAG_CMD_READ) &&
> --
> 2.43.0
>
Mike Christie March 11, 2024, 3:59 p.m. UTC | #2 
On 12/7/23 11:42 AM, lduncan@suse.com wrote:
> From: Lee Duncan <lduncan@suse.com>
> 
> The LIO target code has a warning about setting the
> read and/or write header bits with a PDU that has zero
> transfer length, even though the code mentions that the
> SPEC (RFC 3720) allows this, and that some initiators
> set these bits. But in practice such initiators end up
> flooding the logs with thousands of warning messages for
> IO that is allowed.
> 

I've never seen us hit this. What initiator is doing this and what is 
the command they are sending?

Is it also related to the first patch? Is some initiator sending 
something like a TUR with the immediate bit set during some sort of 
stall/timeout?
diff mbox series

Patch

diff --git a/drivers/target/iscsi/iscsi_target.c b/drivers/target/iscsi/iscsi_target.c
index f246e5015868..c82dc2cd08b3 100644
--- a/drivers/target/iscsi/iscsi_target.c
+++ b/drivers/target/iscsi/iscsi_target.c
@@ -1039,9 +1039,10 @@  int iscsit_setup_scsi_cmd(struct iscsit_conn *conn, struct iscsit_cmd *cmd,
 		hdr->flags &= ~ISCSI_FLAG_CMD_READ;
 		hdr->flags &= ~ISCSI_FLAG_CMD_WRITE;
 
-		pr_warn("ISCSI_FLAG_CMD_READ or ISCSI_FLAG_CMD_WRITE"
+		pr_debug("ISCSI_FLAG_CMD_READ or ISCSI_FLAG_CMD_WRITE"
 			" set when Expected Data Transfer Length is 0 for"
-			" CDB: 0x%02x, Fixing up flags\n", hdr->cdb[0]);
+			" CDB: 0x%02x, cleared READ/WRITE flag(s)\n",
+			hdr->cdb[0]);
 	}
 
 	if (!(hdr->flags & ISCSI_FLAG_CMD_READ) &&