[LTS 8.6 RT] Bluetooth: L2CAP: Fix l2cap_global_chan_by_psm #93
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
CVE-2022-42896
VULN-205
Solution
The bug fix in the mainline is provided1 in two commits:
f937b758a188d6fd328a81367087eddbb2fce50f
711f8c3fb3db61897080468586b970c87c61d9e4
Of these the
711f8c3
is already applied onciqlts8_6-rt
(commit0c319402e3cb34e46cad6ed4299faffa7916b670
).(Same situation as in #41)
kABI check: omitted
Boot test: passed
boot-test.log
Kselftests: passed relative
Methodology
All kselftests used were compiled from the kernel source code, branch
ciqlts8_6-rt
(5d362bf39d935ad7de47fa83661d2965e54ae013
). Thekernel-rt-selftests-internal
package is not available for Rocky LTS 8.6 RT.Additionally the tests were ran using the installed kselftests, as with
in order to evaluate the usefulness of this feature. (The installed selftests use the same
run_kselftest.sh
-defined interface as those fromkernel-selftests-internal
,kernel-rt-selftests-internal
packages, providing high granularity of tests selection and potentially streamlining testing across different rocky versions.)Covered test groups (including those skipped during execution)
android
,bpf
,breakpoints
,capabilities
,core
,cpu-hotplug
,cpufreq
,efivarfs
,exec
,filesystems
,firmware
,fpu
,ftrace
,futex
,gpio
,intel_pstate
,ipc
,kcmp
,kvm
,lib
,livepatch
,membarrier
,memfd
,memory-hotplug
,mount
,net
,net/forwarding
,net/mptcp
,netfilter
,nsfs
,pstore
,ptrace
,rseq
,rtc
,sgx
,sigaltstack
,size
,splice
,static_keys
,sync
,sysctl
,tc-testing
,timens
,timers
,tpm2
,user
,vm
,x86
,zram
Compiled selftests
Tests stability analysis on a reference kernel
A series of 5 test runs were conducted on the reference LTS 8.6 RT kernel
ciqlts8_6-rt
(5d362bf39d935ad7de47fa83661d2965e54ae013
) of which 2 finished without issues.kselftests–source–ciqlts8_6-rt–run1.log
kselftests–source–ciqlts8_6-rt–run2.log
It was found that
bpf:test_sockmap
: Sometimes hangs the machine by blocking the kworkerbpf:test_progs
: Sometimes causes the machine to spontaneously reboot, interrupting the tests run. Potentially this also applies to thebpf:test_progs-no_alu32
test as is the case on other Rocky versions, but on these test runs it behaved properly.ipc:msgque
net:gro.sh
For the full picture of unit tests stability state refer to the column https://docs.google.com/spreadsheets/d/1tUwJ2rV57cYZXh7momPtraSjZcHDjMYHLeHA3DYWrUU/edit?pli=1&gid=0#gid=0&range=D:D
Patched kernel
A single test run was conducted on the patched kernel.
kselftests–source–ciqlts8_6-rt-CVE-2022-42896.log
Comparison
Compared to the reference kernel and omitting the unstable and flappy tests the results are identical except for the
net:reuseport_addr_any.sh
test, which the reference kernel was failing while the patched kernel was passing. An additional series ofnet
tests were conducted on the reference kernel in hope of obtaining the passing result, without success.kselftests–source–ciqlts8_6-rt–net–run1.log
kselftests–source–ciqlts8_6-rt–net–run2.log
kselftests–source–ciqlts8_6-rt–net–run3.log
Full comparison give below
However, the passing
net:reuseport_addr_any.sh
test on the reference kernel when used in the installed form (see section below) suggests that the passing result for the patched kernel is not related to the change in any way.Installed selftests
A series of 2 test runs were conducted on the reference LTS 8.6 RT kernel
ciqlts8_6-rt
(5d362bf39d935ad7de47fa83661d2965e54ae013
) with the tests proven to be unstable before omitted:bpf:test_progs
,bpf:test_sockmap
.kselftests–source-installed–ciqlts8_6-rt–run1.log
kselftests–source-installed–ciqlts8_6-rt–run2.log
It was found that the
kvm:hardware_disable_test
was flappy. Additionally, multiple tests behaved differently when run in this mode compared to the tests run withmake
.The reasons were investigated cursorily and they ranged from conditional "skipping" of some tests (eg.
bpf:test_lwt_seg6local.sh
) which showed as passing when invoked withmake
to what seems like an erroneous installation (rseq
).The full picture is provided at https://docs.google.com/spreadsheets/d/1tUwJ2rV57cYZXh7momPtraSjZcHDjMYHLeHA3DYWrUU/edit?pli=1&gid=0#gid=0&range=E:E.
These results show that the tests ran directly from kernel source and the tests ran with
run_kselftest.sh
after installation are not interchangable in a straightforward manner.Additional tests: none
Following the guidelines from the precedent #41.
Footnotes
1 GHSA-pf87-6c9q-jvm4