NULL Pointer Dereference Affecting gfs2-kmp-default package, versions <5.14.21-150500.55.73.1


Severity

Recommended
0.0
medium
0
10

Based on SUSE Linux Enterprise Server security rating.

Threat Intelligence

EPSS
0.04% (11th percentile)

Do your applications use this vulnerable package?

In a few clicks we can analyze your entire application and see what components are vulnerable in your application, and suggest you quick fixes.

Test your applications

Snyk Learn

Learn about NULL Pointer Dereference vulnerabilities in an interactive lesson.

Start learning
  • Snyk IDSNYK-SLES155-GFS2KMPDEFAULT-7696010
  • published17 Aug 2024
  • disclosed16 Aug 2024

Introduced: 16 Aug 2024

CVE-2021-47592  (opens in a new tab)
CWE-476  (opens in a new tab)

How to fix?

Upgrade SLES:15.5 gfs2-kmp-default to version 5.14.21-150500.55.73.1 or higher.

NVD Description

Note: Versions mentioned in the description apply only to the upstream gfs2-kmp-default package and not the gfs2-kmp-default package as distributed by SLES. See How to fix? for SLES:15.5 relevant fixed versions and status.

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

net: stmmac: fix tc flower deletion for VLAN priority Rx steering

To replicate the issue:-

  1. Add 1 flower filter for VLAN Priority based frame steering:- $ IFDEVNAME=eth0 $ tc qdisc add dev $IFDEVNAME ingress $ tc qdisc add dev $IFDEVNAME root mqprio num_tc 8
    map 0 1 2 3 4 5 6 7 0 0 0 0 0 0 0 0
    queues 1@0 1@1 1@2 1@3 1@4 1@5 1@6 1@7 hw 0 $ tc filter add dev $IFDEVNAME parent ffff: protocol 802.1Q
    flower vlan_prio 0 hw_tc 0

  2. Get the 'pref' id $ tc filter show dev $IFDEVNAME ingress

  3. Delete a specific tc flower record (say pref 49151) $ tc filter del dev $IFDEVNAME parent ffff: pref 49151

From dmesg, we will observe kernel NULL pointer ooops

[ 197.170464] BUG: kernel NULL pointer dereference, address: 0000000000000000 [ 197.171367] #PF: supervisor read access in kernel mode [ 197.171367] #PF: error_code(0x0000) - not-present page [ 197.171367] PGD 0 P4D 0 [ 197.171367] Oops: 0000 [#1] PREEMPT SMP NOPTI

<snip>

[ 197.171367] RIP: 0010:tc_setup_cls+0x20b/0x4a0 [stmmac]

<snip>

[ 197.171367] Call Trace: [ 197.171367] <TASK> [ 197.171367] ? __stmmac_disable_all_queues+0xa8/0xe0 [stmmac] [ 197.171367] stmmac_setup_tc_block_cb+0x70/0x110 [stmmac] [ 197.171367] tc_setup_cb_destroy+0xb3/0x180 [ 197.171367] fl_hw_destroy_filter+0x94/0xc0 [cls_flower]

The above issue is due to previous incorrect implementation of tc_del_vlan_flow(), shown below, that uses flow_cls_offload_flow_rule() to get struct flow_rule *rule which is no longer valid for tc filter delete operation.

struct flow_rule *rule = flow_cls_offload_flow_rule(cls); struct flow_dissector *dissector = rule->match.dissector;

So, to ensure tc_del_vlan_flow() deletes the right VLAN cls record for earlier configured RX queue (configured by hw_tc) in tc_add_vlan_flow(), this patch introduces stmmac_rfs_entry as driver-side flow_cls_offload record for 'RX frame steering' tc flower, currently used for VLAN priority. The implementation has taken consideration for future extension to include other type RX frame steering such as EtherType based.

v2:

  • Clean up overly extensive backtrace and rewrite git message to better explain the kernel NULL pointer issue.

CVSS Scores

version 3.1