Hey, ping here :)
Any thoughts or suggestions about the proposed patch series?
On 13/01/2022 13:42, Vasiliy Ulyanov wrote:
v2 of
https://listman.redhat.com/archives/libvir-list/2022-January/msg00008.html
The v2 series introduces the new approach following the previous review
comments. It adds a function for pidfile lock validation (i.e. if its
locked by the expected process) and adds the new check to
virPidFileReadPathIfAlive. The daemonization and pidfile handling for
swtpm command are now handled by libvirt.
The fix to qemu_vhost_user_gpu is currently done by visual code
checking. Not really sure how to test e2e as I dont have a quick
reproducer atm :(
Note: I wasn't sure about the refactoring of virPidFileReadPathIfAlive
(i.e. whether to remove the binary path from there). For now decided to
introduce the new lock check only if the path is not provided. But I am
open for suggestions on how to better fit the new check.
Vasiliy Ulyanov (4):
virfile: Add virFileGetLockOwner function
virpidfile: Refactor virPidFileReadPathIfAlive
qemu_tpm: Get swtpm pid without binary validation
qemu: gpu: Get pid without binary validation
src/libvirt_private.syms | 1 +
src/qemu/qemu_tpm.c | 26 +++++++++-----------
src/qemu/qemu_vhost_user_gpu.c | 9 +++----
src/util/virfile.c | 45 ++++++++++++++++++++++++++++++++++
src/util/virfile.h | 2 ++
src/util/virpidfile.c | 20 +++++++++++++++
6 files changed, 82 insertions(+), 21 deletions(-)
--
Vasily Ulyanov <vulyanov(a)suse.de>
Software Engineer, SUSE Labs Core