CVE-2026-99766: scsi: aic79xx: Fix NULL pointer dereference in device reset
In the Linux kernel, the following vulnerability has been resolved:
scsi: aic79xx: Fix NULL pointer dereference in device reset
Since commit c67e63800446 ("scsi: aic79xx: Do not reference SCSI command when resetting device") the device reset SCB has no SCSI command, but ahdlinuxdevreset() still sets its residual, and ahdabortscbs(), which completes it, sets its status. Seen on an AIC-7902 when a failing disk triggered a device reset, which killed the SCSI error handler:
BUG: kernel NULL pointer dereference, address: 0000009c EIP: ahdlinuxdevreset.cold+0xee/0x26e
Add NULL command checks to the accessors in aic79xxosm.h. Also skip autosense for an SCB without a command and complete it with ahddone() right away: the REQUEST SENSE path can clear SCBDEVICERESET, after which ahddone() would dereference the missing command.
Affected Software
Remediation
Recommended actions to resolve this vulnerability, in priority order.
- Upgrade
Upgrade
linux/Kernelto a version that resolves this vulnerability.Fixed in 6.12.118 - Upgrade
Upgrade
linux/Kernelto a version that resolves this vulnerability.Fixed in 6.18.64 - Upgrade
Upgrade
linux/Kernelto a version that resolves this vulnerability.Fixed in 7.3.5 - Upgrade
Upgrade
debian/linuxto a version that resolves this vulnerability.Fixed in 6.1.176-1Fixed in 6.1.192-2Fixed in 7.3.5-1 - Compensating control
Add NULL command checks to the accessors in aic79xx_osm.h to prevent dereferencing a missing SCSI command.
Event History
Frequently Asked Questions
Which systems are realistically exposed to this issue?
Systems using the Linux kernel aic79xx SCSI driver with an AIC-7902 controller are known to be exposed based on the reported trigger. The failure occurs during a device-reset path, such as when a disk is failing.
What is the operational impact if this is triggered?
The NULL pointer dereference can kill the SCSI error handler during device reset. This can disrupt error recovery for the affected SCSI device.
Is an attacker required to trigger the problem?
The provided information describes a hardware or device-error trigger: a failing disk causing a device reset. It does not identify any remote, local-user, or authentication-based attack prerequisite.
What code paths does the fix change?
The fix adds NULL-command checks to aic79xx accessors, avoids autosense for reset SCBs that have no command, and completes those SCBs immediately with ahd_done(). This prevents later paths from dereferencing the missing SCSI command.