In the Linux kernel, the following vulnerability has been resolved:
scsi: qla2xxx: Avoid double completion in async IOCB timeout
qla2x00_async_iocb_timeout() tries to abort a timed-out async IOCB. When
qla24xx_async_abort_cmd() fails, both the SRB_LOGIN_CMD path and the
SRB_CTRL_VP/default path scan outstanding_cmds[] for the SRB and then
call sp->done(sp, QLA_FUNCTION_TIMEOUT) unconditionally, without checking
whether the SRB was actually found and removed.
If the response ISR completes the same handle first, it removes the SRB
under qp_lock_ptr and runs sp->done() -> complete(sp->comp). The
submitter qla24xx_control_vp() wakes from wait_for_completion(), clears
sp->comp, drops its reference and returns, reclaiming the on-stack
completion. The timer reference keeps the SRB alive across the timeout
handler, but not the submitter's stack. The timeout then issues a second
sp->done() -> qla_ctrlvp_sp_done(), which evaluates "if (sp->comp)
complete(sp->comp)"; with the pointer loaded before the submitter's NULL
store, complete() writes into the freed stack frame, a use-after-free.
Track whether this path removed the SRB from outstanding_cmds and only
call sp->done() when it did, so the command is completed exactly once by
whichever path owns it. This mirrors the sp_found guard already used in
qla24xx_abort_iocb_timeout().
CVSS Vector: CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:H/I:H/A:H
CVSS Score: 9.8
AV:N - qla2x00_async_iocb_timeout() runs on fabric login, CT/SNS, and target-mode NACK IOCBs; a remote FC/FCoE/FCIP peer delays then completes those exchanges so the response ISR races the timer. That is a SAN-network path, consistent with other qla2xxx FC-peer memory-corruption scores.
AC:L - The peer controls both sides of the race by withholding the login/CT/NACK completion until the async timer fires, then posting it so the ISR removes the SRB while timeout still calls sp->done(). Abort failure is also attacker-driven via GFP_ATOMIC SRB exhaustion or ISP-abort from injected link errors.
PR:N - Fabric login, name-server CT, and target-mode notify-ack run in driver ISR/DPC context with no Linux credential or capability check. FC-SP/EDIF is opt-in and rare, so a fabric peer needs no account on the victim host.
UI:N - Discovery, PLOGI, and NACK IOCBs are posted automatically on link-up, RSCN, and inbound ELS. The async timer and response ISR then run without any victim mount, open, or administrator action.
S:U - Double completion corrupts host-kernel SRB and completion state inside the qla2xxx driver. The impact stays in that kernel authority and is not a VM escape, IOMMU bypass, or other cross-boundary breakout.
C:H - The extra sp->done() drops the SRB INIT kref twice so the timeout path later frees a mempool SRB still in use, a use-after-free that can disclose reused heap contents. The CTRL_VP path also complete()s a reclaimed on-stack completion.
I:H - The same mempool SRB use-after-free and complete() write into a freed stack frame are kernel write primitives. Reuse of the freed SRB can hijack control flow rather than producing only a clean timeout.
A:H - Double completion of a live SRB is a use-after-free that oopses or panics the host, as in the original switch-CT crash in qla2x00_chk_ms_status from the response ISR, taking down storage on the adapter.
| Attack Vector |
Network |
Scope |
Unchanged |
| Attack Complexity |
Low |
Confidentiality Impact |
High |
| Privileges Required |
None |
Integrity Impact |
High |
| User Interaction |
None |
Availability Impact |
High |
AV:N - qla2x00_async_iocb_timeout() runs on fabric login, CT/SNS, and target-mode NACK IOCBs; a remote FC/FCoE/FCIP peer delays then completes those exchanges so the response ISR races the timer. That is a SAN-network path, consistent with other qla2xxx FC-peer memory-corruption scores.
AC:L - The peer controls both sides of the race by withholding the login/CT/NACK completion until the async timer fires, then posting it so the ISR removes the SRB while timeout still calls sp->done(). Abort failure is also attacker-driven via GFP_ATOMIC SRB exhaustion or ISP-abort from injected link errors.
PR:N - Fabric login, name-server CT, and target-mode notify-ack run in driver ISR/DPC context with no Linux credential or capability check. FC-SP/EDIF is opt-in and rare, so a fabric peer needs no account on the victim host.
UI:N - Discovery, PLOGI, and NACK IOCBs are posted automatically on link-up, RSCN, and inbound ELS. The async timer and response ISR then run without any victim mount, open, or administrator action.
S:U - Double completion corrupts host-kernel SRB and completion state inside the qla2xxx driver. The impact stays in that kernel authority and is not a VM escape, IOMMU bypass, or other cross-boundary breakout.
C:H - The extra sp->done() drops the SRB INIT kref twice so the timeout path later frees a mempool SRB still in use, a use-after-free that can disclose reused heap contents. The CTRL_VP path also complete()s a reclaimed on-stack completion.
I:H - The same mempool SRB use-after-free and complete() write into a freed stack frame are kernel write primitives. Reuse of the freed SRB can hijack control flow rather than producing only a clean timeout.
A:H - Double completion of a live SRB is a use-after-free that oopses or panics the host, as in the original switch-CT crash in qla2x00_chk_ms_status from the response ISR, taking down storage on the adapter.
CVSS 3.1