Re: [PATCH 1/4] zfcp: Enhance fsf status read buffer tracing

[sashiko-bot@kernel.org]

Thank you for your contribution! Sashiko AI review found 2 potential issue(s) to consider:

Pre-existing issues:
- [High] An integer underflow occurs when the hardware/firmware provides a non-zero `srb->length` smaller than the `payload` offset, leading to a massive out-of-bounds read and potential kernel panic.
- [High] Missing bounds check on hardware-provided `log_start` and `log_length` allows an out-of-bounds read and information leak.
--

--- Patch [1]: [PATCH 1/4] zfcp: Enhance fsf status read buffer tracing ---
commit 2318c7a31a1b9a647e0a90e8e4d57fd114b807d1
Author: Chinmaya Kajagar <chinmayk@linux.ibm.com>

zfcp: Enhance fsf status read buffer tracing

This commit appends the remaining SRB fields to the unsolicited status trace
records and extends the driver to get the 3-byte source ID from the reserved
field. This helps fully trace the incoming FSF status read buffer.

> diff --git a/drivers/s390/scsi/zfcp_dbf.c b/drivers/s390/scsi/zfcp_dbf.c
> index 71f625926ae19..89b859176b8bd 100644
> --- a/drivers/s390/scsi/zfcp_dbf.c
> +++ b/drivers/s390/scsi/zfcp_dbf.c

[ ... ]

> @@ -220,6 +220,13 @@ void zfcp_dbf_hba_fsf_uss(char *tag, struct zfcp_fsf_req *req)
>  	rec->u.uss.lun = srb->fcp_lun;
>  	memcpy(&rec->u.uss.queue_designator, &srb->queue_designator,
>  	       sizeof(rec->u.uss.queue_designator));
> +	rec->u.uss.length = srb->length;
> +	rec->u.uss.res1 = srb->res1;
> +	rec->u.uss.res2 = srb->res2;
> +	rec->u.uss.class = srb->class;
> +	rec->u.uss.res3 = srb->res3;
> +	rec->u.uss.s_id = ntoh24(srb->s_id);
> +	memcpy(&rec->u.uss.res4, &srb->res4, sizeof(rec->u.uss.res4));
>  
>  	/* status read buffer payload length */
>  	rec->pl_len = (!srb->length) ? 0 : srb->length -
> 			offsetof(struct fsf_status_read_buffer, payload);
> 
> 	if (rec->pl_len)
> 		zfcp_dbf_pl_write(dbf, srb->payload.data, rec->pl_len,
> 				  "fsf_uss", req->req_id, ZFCP_DBF_PAY_LEVEL);

[Severity: High]
This is a pre-existing issue, but is there a risk of an integer underflow
here if the firmware returns a non-zero srb->length that is smaller than the
payload offset?

If srb->length is less than offsetof(struct fsf_status_read_buffer, payload),
the subtraction will underflow. Because rec->pl_len is a u16, it would store a
large truncated positive integer.

Could this cause zfcp_dbf_pl_write() to perform a massive out-of-bounds read
beyond the srb buffer, potentially leading to a kernel panic from unmapped
memory access or leaking adjacent sensitive kernel heap data into the trace
ring buffer?

An identical instance of this calculation seems to exist in
zfcp_dbf_san_in_els() as well.


While reviewing this tracing code, I also noticed another pre-existing issue
in zfcp_dbf_hba_fsf_res():

drivers/s390/scsi/zfcp_dbf.c:zfcp_dbf_hba_fsf_res() {
	...
	if (q_head->log_length) {
		rec->pl_len = q_head->log_length;
		zfcp_dbf_pl_write(dbf, (char *)q_pref + q_head->log_start,
				  rec->pl_len, "fsf_res", req->req_id,
				  ZFCP_DBF_PAY_LEVEL);
	}
	...
}

[Severity: High]
This is also a pre-existing issue, but does this code adequately validate the
hardware-provided q_head->log_length and q_head->log_start values?

Since these values are read directly from the hardware-populated qtcb header,
is it possible for a malfunctioning or compromised firmware to provide
arbitrary offsets and lengths?

Without bounds checking to ensure they fall strictly within the allocated qtcb
structure, could zfcp_dbf_pl_write() perform an out-of-bounds read from the
kernel heap, potentially leaking sensitive data or causing a crash?

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20260724053934.2781087-1-niharp@linux.ibm.com?part=1