Message ID | 20250303124324.3563605-1-ming.lei@redhat.com (mailing list archive) |
---|---|
Headers | show |
Series | selftests: ublk: bug fixes & consolidation | expand |
On Mon, Mar 3, 2025 at 8:43 PM Ming Lei <ming.lei@redhat.com> wrote: > > Hello Jens and guys, > > This patchset fixes several issues(1, 2, 4) and consolidate & improve > the tests in the following ways: > > - support shellcheck and fixes all warning > > - misc cleanup > > - improve cleanup code path(module load/unload, cleanup temp files) > > - help to reuse the same test source code and scripts for other > projects(liburing[1], blktest, ...) > > - add two stress tests for covering IO workloads vs. removing device & > killing ublk server, given buffer lifetime is one big thing for ublk-zc > > > [1] https://github.com/ming1/liburing/commits/ublk-zc > > - just need one line change for overriding skip_code, libring uses 77 and > kselftests takes 4 Hi Jens, Can you merge this patchset if you are fine? Thanks,
On 3/10/25 9:09 AM, Ming Lei wrote: > On Mon, Mar 3, 2025 at 8:43?PM Ming Lei <ming.lei@redhat.com> wrote: >> >> Hello Jens and guys, >> >> This patchset fixes several issues(1, 2, 4) and consolidate & improve >> the tests in the following ways: >> >> - support shellcheck and fixes all warning >> >> - misc cleanup >> >> - improve cleanup code path(module load/unload, cleanup temp files) >> >> - help to reuse the same test source code and scripts for other >> projects(liburing[1], blktest, ...) >> >> - add two stress tests for covering IO workloads vs. removing device & >> killing ublk server, given buffer lifetime is one big thing for ublk-zc >> >> >> [1] https://github.com/ming1/liburing/commits/ublk-zc >> >> - just need one line change for overriding skip_code, libring uses 77 and >> kselftests takes 4 > > Hi Jens, > > Can you merge this patchset if you are fine? Yep sorry, was pondering how best to get it staged. Should go into block, but depends on the other bits that I staged for io_uring. So I'll just put it there, not a big deal.
On Mon, 03 Mar 2025 20:43:10 +0800, Ming Lei wrote: > This patchset fixes several issues(1, 2, 4) and consolidate & improve > the tests in the following ways: > > - support shellcheck and fixes all warning > > - misc cleanup > > [...] Applied, thanks! [01/11] selftests: ublk: make ublk_stop_io_daemon() more reliable commit: 9894e0eaae980df1ed3f2e86a487fe4c8ef1ab46 [02/11] selftests: ublk: fix build failure commit: 9d80f48c5e08b2e003e506c6e5326a35a652ea2f [03/11] selftests: ublk: add --foreground command line commit: 2ecdcdfee58c028c15ed00b691104249370db075 [04/11] selftests: ublk: fix parsing '-a' argument commit: cf2132935639813a0b88e55074e6e52a4b82f26a [05/11] selftests: ublk: support shellcheck and fix all warning commit: 30aab83035048c70e09ff058a73e8428de9bd103 [06/11] selftests: ublk: don't pass ${dev_id} to _cleanup_test() commit: 8da9f88fee59fe5aa99014a2621b07347edd5780 [07/11] selftests: ublk: move zero copy feature check into _add_ublk_dev() commit: b95b47eaa8d7c8b595d93397d1b85f1559c2d220 [08/11] selftests: ublk: load/unload ublk_drv when preparing & cleaning up tests commit: 9e71305495d1b79f96729b8d77d4d823a6bd998a [09/11] selftests: ublk: add one stress test for covering IO vs. removing device commit: 6f3004e78b59e98a903e20e2240ae77e76dfde77 [10/11] selftests: ublk: add stress test for covering IO vs. killing ublk server commit: 4fcd5b5a6dff71cf82212dd208dc1765ca8a8088 [11/11] selftests: ublk: improve test usability commit: 22c880f446a149f5ee11260690a34d4b3f95c221 Best regards,
On Mon, Mar 10, 2025 at 09:17:56AM -0600, Jens Axboe wrote: > On 3/10/25 9:09 AM, Ming Lei wrote: > > On Mon, Mar 3, 2025 at 8:43?PM Ming Lei <ming.lei@redhat.com> wrote: > >> > >> Hello Jens and guys, > >> > >> This patchset fixes several issues(1, 2, 4) and consolidate & improve > >> the tests in the following ways: > >> > >> - support shellcheck and fixes all warning > >> > >> - misc cleanup > >> > >> - improve cleanup code path(module load/unload, cleanup temp files) > >> > >> - help to reuse the same test source code and scripts for other > >> projects(liburing[1], blktest, ...) > >> > >> - add two stress tests for covering IO workloads vs. removing device & > >> killing ublk server, given buffer lifetime is one big thing for ublk-zc > >> > >> > >> [1] https://github.com/ming1/liburing/commits/ublk-zc > >> > >> - just need one line change for overriding skip_code, libring uses 77 and > >> kselftests takes 4 > > > > Hi Jens, > > > > Can you merge this patchset if you are fine? > > Yep sorry, was pondering how best to get it staged. Should go into > block, but depends on the other bits that I staged for io_uring. So I'll > just put it there, not a big deal. Thanks for pulling it in! BTW, the test behavior depends on block too, otherwise it may fail because ublk zc actually depends on the fix of "ublk: complete command synchronously on error". So if anyone wants to try the test, please do it against next tree. Thanks, Ming
On 3/10/25 10:35 PM, Ming Lei wrote: > On Mon, Mar 10, 2025 at 09:17:56AM -0600, Jens Axboe wrote: >> On 3/10/25 9:09 AM, Ming Lei wrote: >>> On Mon, Mar 3, 2025 at 8:43?PM Ming Lei <ming.lei@redhat.com> wrote: >>>> >>>> Hello Jens and guys, >>>> >>>> This patchset fixes several issues(1, 2, 4) and consolidate & improve >>>> the tests in the following ways: >>>> >>>> - support shellcheck and fixes all warning >>>> >>>> - misc cleanup >>>> >>>> - improve cleanup code path(module load/unload, cleanup temp files) >>>> >>>> - help to reuse the same test source code and scripts for other >>>> projects(liburing[1], blktest, ...) >>>> >>>> - add two stress tests for covering IO workloads vs. removing device & >>>> killing ublk server, given buffer lifetime is one big thing for ublk-zc >>>> >>>> >>>> [1] https://github.com/ming1/liburing/commits/ublk-zc >>>> >>>> - just need one line change for overriding skip_code, libring uses 77 and >>>> kselftests takes 4 >>> >>> Hi Jens, >>> >>> Can you merge this patchset if you are fine? >> >> Yep sorry, was pondering how best to get it staged. Should go into >> block, but depends on the other bits that I staged for io_uring. So I'll >> just put it there, not a big deal. > > Thanks for pulling it in! > > BTW, the test behavior depends on block too, otherwise it may fail > because ublk zc actually depends on the fix of "ublk: complete command > synchronously on error". > > So if anyone wants to try the test, please do it against next tree. Indeed - not a huge deal, as they will go into the main tree at roughly the same time anyway. But good to note.