CVE-2026-90414

In the Linux kernel, the following vulnerability has been resolved:

IB/isert: reject PDUs declaring more data than was received

isertrecvdone() hands each received PDU to the opcode handlers without

ever looking at wc->byte_len, the number of bytes the HCA actually placed

in the receive descriptor. The handlers then copy that many bytes - the

data-segment length the initiator declared in the BHS

(ntoh24(hdr->dlength), via the derived unsoldatalen / immdatalen) -

out of the fixed-size descriptor:

  iserthandleiscsi_dataout():

        sgcopyfrombuffer(sgstart, sgnents, isertgetdata(rxdesc),

                            unsoldatalen);

  iserthandlescsi_cmd():

        sgcopyfrombuffer(cmd->secmd.tdatasg, sg_nents,

                            isertgetdata(rxdesc), immdata_len);

Because the declared length is never checked against wc->byte_len, an

initiator can declare a data segment larger than the bytes it actually

sent (and larger than the descriptor) and cause an out-of-bounds read of

the receive buffer.

Nothing upstream of isert closes this door:

• _iscsitcheckdataouthdr() bounds the inbound payload against

    conn_ops->MaxXmitDataSegmentLength (MXDSL) - a transmit parameter,

    used here for the inbound check.

• iscsisetconnection_parameters() sets

    ops->MaxXmitDataSegmentLength = ops->TargetRecvDataSegmentLength;

    and TARGETRECVDATASEGMENTLENGTH is absent from the min()-clamp list in

    iscsicheckacceptor_state(), so the value the initiator declares is

    adopted verbatim (type range 512..16777215). The initiator effectively

    raises its own ceiling.

• isert never clamps the negotiated value to its own fixed receive

    descriptor (ISERRXSIZE, 9216 bytes), so the target core's bound and

    the descriptor size are unrelated.

The immdatalen == data_len path is more than an over-read: it aliases

the receive descriptor via sgsetbuf() and passes it to the backend as

the data source for the SCSI WRITE, so an over-declared length causes heap

contents past the descriptor to be written through the backend to the

backing store. The backend is the victim of the oversized scatterlist

isert hands it, not the cause; no read-back of the written bytes was

demonstrated.

Trigger: after login completes (full feature phase), an initiator that has

declared a large TargetRecvDataSegmentLength and a FirstBurstLength that

permits unsolicited/immediate data sends a PDU whose declared data-segment

length exceeds what was received. With KASAN:

  BUG: KASAN: slab-out-of-bounds in sgcopybuffer+0x150/0x1c0

  Read of size 4096 at addr ffff888109720800 by task kworker/1:0H/25

  Workqueue: ib-comp-wq ibcqpoll_work

  Call Trace:

   sgcopybuffer+0x150/0x1c0

   isertrecvdone+0xba6/0x2390

   _ibprocess_cq+0xe1/0x390

   ibcqpoll_work+0x46/0x150

isertrecvdone+0xba6 resolves to iserthandleiscsi_dataout()

(ibisert.c:1160), inlined through isertrx_opcode().

Validate wc->bytelen against the framing in isertrecv_done() before the

PDU reaches any handler, and reinstate the connection if it is short.

Because the test compares without subtracting the header length, it also

rejects PDUs shorter than the iSER and iSCSI headers, which would otherwise

be parsed out of stale descriptor contents. The login handler rejects PDUs

shorter than ISERHEADERSLEN (commit 29e7b925ae6d ("IB/isert: Reject login

PDUs shorter than ISERHEADERSLEN")) but does not bound the declared

length either; that is fixed in the next patch. The data handlers had no

length check at all.

isert reads the data segment from a fixed offset: isertgetdata()

returns the iSER header plus ISERHEADERSLEN and makes no adjustment for

an AHS.  The bytes the handlers touch are therefore exactly

[ISERHEADERSLEN, ISERHEADERSLEN + dlength), and comparing that sum

against wc->byte_len bounds precisely the region that is read.  An AHS

term would only make the test stricter without bounding anything furth

---truncated---