On 9/26/23 12:05, Bart Van Assche wrote: > On 9/26/23 07:06, Leon Romanovsky wrote: >> On Tue, Sep 26, 2023 at 12:43:57PM +0300, Leon Romanovsky wrote: >>> >>> On Fri, 22 Sep 2023 12:32:31 -0400, Zhu Yanjun wrote: >>>> This reverts commit 9b4b7c1f9f54120940e243251e2b1407767b3381. >>>> >>>> This commit replaces tasklet with workqueue. But this results >>>> in occasionally pocess hang at the blktests test case srp/002. >>>> After the discussion in the link[1], this commit is reverted. >>>> >>>> >>>> [...] >>> >>> Applied, thanks! >>> >>> [1/1] Revert "RDMA/rxe: Add workqueue support for rxe tasks" >>> https://git.kernel.org/rdma/rdma/c/e710c390a8f860 >> >> I applied this patch, but will delay it for some time with a hope that >> fix will arrive in the near future. > > Thank you Leon. With this revert applied on top of the rdma for-next branch, I > don't see the KASAN complaint anymore that I reported yesterday. I think this > is more evidence that the KASAN complaint was caused by the RXE driver. > > Bart. Bart, I am working to try to reproduce the KASAN warning. Unfortunately, so far I am not able to see it in Ubuntu + Linus' kernel (as you described) on metal. The config file is different but copies the CONFIG_KASAN_xxx exactly as yours. With KASAN enabled it hangs on every iteration of srp/002 but without a KASAN warning. I am now building an openSuSE VM for qemu and will see if that causes the warning. Thanks for the support, Bob
