CVE-2019-3896

A double-free can happen in idr_remove_all() in lib/idr.c in the Linux kernel 2.6 branch. An unprivileged local attacker can use this flaw for a privilege escalation or for a system crash and a denial of service (DoS).

The SUSE Linux Enterprise 11 SP4 kernel was updated to receive various security and bugfixes.

The following security bugs were fixed:

• CVE-2019-3459: A heap address information leak while using L2CAP_GET_CONF_OPT was discovered in the Linux kernel(bnc#1120758).
• CVE-2019-3460: A heap data infoleak in multiple locations including L2CAP_PARSE_CONF_RSP was found in the Linux kernel before (bnc#1120758).
• CVE-2019-3896: A double-free could happen in idr_remove_all() in lib/idr.c in the Linux kernel. An unprivileged local attacker could use this flaw for a privilege escalation or for a system crash and a denial of service (DoS) (bnc#1138943).
• CVE-2018-5390: The Linux kernel could be forced to make very expensive calls to tcp_collapse_ofo_queue() and tcp_prune_ofo_queue() for every incoming packet which could lead to a denial of service (bnc#1102340).
• CVE-2018-20836: An issue was discovered in the Linux kernel There was a race condition in smp_task_timedout() and smp_task_done() in drivers/scsi/libsas/sas_expander.c, leading to a use-after-free (bnc#1134395).
• CVE-2019-12614: An issue was discovered in dlpar_parse_cc_property in arch/powerpc/platforms/pseries/dlpar.c in the Linux kernel There was an unchecked kstrdup of prop-name, which might allow an attacker to cause a denial of service (NULL pointer dereference and system crash) (bnc#1137194).
• CVE-2019-3846: A flaw that allowed an attacker to corrupt memory and possibly escalate privileges was found in the mwifiex kernel module while connecting to a malicious wireless network (bnc#1136424 1136446).

The following non-security bugs were fixed:

• KEYS: do not let add_key() update an uninstantiated key (bnc#1063416).
• fnic: Fix to cleanup aborted IO to avoid device being offlined by mid-layer (bsc#1134835).
• signal: give SEND_SIG_FORCED more power to beat SIGNAL_UNKILLABLE (bsc#1135650).
• signal: oom_kill_task: use SEND_SIG_FORCED instead of force_sig() (bsc#1135650).
• tcp: a regression in the previous fix for the TCP SACK issue was fixed (bnc#1139751)