In the Linux kernel, the following vulnerability has been resolved:
net: kcm: Fix race condition in kcm_unattach()
syzbot found a race condition when kcmunattach(psock) and kcmrelease(kcm) are executed at the same time.
kcmunattach() is missing a check of the flag kcm->txstopped before calling queue_work().
If the kcm has a reserved psock, kcmunattach() might get executed between cancelworksync() and unreservepsock() in kcmrelease(), requeuing kcm->txwork right before kcm gets freed in kcm_done().
Remove kcm->txstopped and replace it by the less error-prone disablework_sync().