| Message ID | 20230510070207.1501-1-yangx.jy@fujitsu.com (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
| Series | [blktests] common: Replace _have_module() with _have_driver() | expand |
On 5/10/23 00:02, Xiao Yang wrote: > nvmeof-mp and srp test groups can be executed with built-in > scsi_dh_alua, scsi_dh_emc and scsi_dh_rdac drivers. > > Signed-off-by: Xiao Yang <yangx.jy@fujitsu.com> > --- > we might remove nvmeof-mp, so not sure if that part is needed, let's wait for others to reply .. -ck
On 2023/5/10 15:35, Chaitanya Kulkarni wrote: > we might remove nvmeof-mp, so not sure if that part is > needed, let's wait for others to reply .. Hi CK, Thanks for your feedback. Could you tell me why nvmeof-mp will be removed? Is there any URL about the decision? Best Regards, Xiao Yang
On 5/10/23 00:55, Xiao Yang (Fujitsu) wrote: > On 2023/5/10 15:35, Chaitanya Kulkarni wrote: >> we might remove nvmeof-mp, so not sure if that part is >> needed, let's wait for others to reply .. > Hi CK, > > Thanks for your feedback. > Could you tell me why nvmeof-mp will be removed? Is there any URL about > the decision? > > Best Regards, > Xiao Yang We talked about this in the LSFMM, I hope to see lwn article covering blktests session .. -ck
On Wed, May 10, 2023 at 08:00:12AM +0000, Chaitanya Kulkarni wrote: > On 5/10/23 00:55, Xiao Yang (Fujitsu) wrote: > > On 2023/5/10 15:35, Chaitanya Kulkarni wrote: > >> we might remove nvmeof-mp, so not sure if that part is > >> needed, let's wait for others to reply .. > > Hi CK, > > > > Thanks for your feedback. > > Could you tell me why nvmeof-mp will be removed? Is there any URL about > > the decision? > > > > Best Regards, > > Xiao Yang > > We talked about this in the LSFMM, I hope to see lwn article covering > blktests session .. I think a couple of those test might make sense to be migrated to the nvme suite. I am thinking about the test with simulate_network_failure_loop.
On May 10, 2023 / 18:35, Daniel Wagner wrote: > On Wed, May 10, 2023 at 08:00:12AM +0000, Chaitanya Kulkarni wrote: > > On 5/10/23 00:55, Xiao Yang (Fujitsu) wrote: > > > On 2023/5/10 15:35, Chaitanya Kulkarni wrote: > > >> we might remove nvmeof-mp, so not sure if that part is > > >> needed, let's wait for others to reply .. > > > Hi CK, > > > > > > Thanks for your feedback. > > > Could you tell me why nvmeof-mp will be removed? Is there any URL about > > > the decision? > > > > > > Best Regards, > > > Xiao Yang > > > > We talked about this in the LSFMM, I hope to see lwn article covering > > blktests session .. Yes, the conclusion at LSFMM was to remove nvmeof-mp test group, since dm- multipath for NVME devices is no longer recommended. Xiao, could you repost your patch to cover only srp? > > I think a couple of those test might make sense to be migrated to the nvme > suite. I am thinking about the test with simulate_network_failure_loop. It does not sound reasonable to keep dm-multipath tests in nvme group. I guess you mean to reuse the simulate_network_failure_loop not to test dm-multipath but to test nvme fabrics transports, right?
On Thu, May 18, 2023 at 04:13:51AM +0000, Shinichiro Kawasaki wrote: > > I think a couple of those test might make sense to be migrated to the nvme > > suite. I am thinking about the test with simulate_network_failure_loop. > > It does not sound reasonable to keep dm-multipath tests in nvme group. I guess > you mean to reuse the simulate_network_failure_loop not to test dm-multipath > but to test nvme fabrics transports, right? Correct. I think it would be good to carry over the network failure testing for the fabrics transport test in the nvme group. I can look into this next week.
On 2023/5/18 12:13, Shinichiro Kawasaki wrote: > Yes, the conclusion at LSFMM was to remove nvmeof-mp test group, since dm- > multipath for NVME devices is no longer recommended. > > Xiao, could you repost your patch to cover only srp? Hi Kawasaki-san, Sorry, I overlooked the comment. I have sent the v2 patch. https://lore.kernel.org/linux-block/20230524030643.404546-1-yangx.jy@fujitsu.com/T/#u Best Regards, Xiao Yang
On 2023/5/18 12:13, Shinichiro Kawasaki wrote: > Yes, the conclusion at LSFMM was to remove nvmeof-mp test group, since dm- > multipath for NVME devices is no longer recommended. > > Xiao, could you repost your patch to cover only srp? Hi Kawasaki-san, Sorry, I overlooked the comment. I have sent v2 patch. https://lore.kernel.org/linux-block/20230524030643.404546-1-yangx.jy@fujitsu.com/T/#u Best Regards, Xiao Yang
diff --git a/tests/nvmeof-mp/rc b/tests/nvmeof-mp/rc index 4f4d518..0771497 100755 --- a/tests/nvmeof-mp/rc +++ b/tests/nvmeof-mp/rc @@ -26,6 +26,10 @@ group_requires() { _have_configfs || return + _have_driver scsi_dh_alua + _have_driver scsi_dh_emc + _have_driver scsi_dh_rdac + _have_module dm_multipath _have_module dm_queue_length _have_module dm_service_time @@ -36,9 +40,6 @@ group_requires() { _have_module nvme-rdma _have_module nvmet-rdma _have_module rdma_rxe - _have_module scsi_dh_alua - _have_module scsi_dh_emc - _have_module scsi_dh_rdac for p in mkfs.ext4 mkfs.xfs multipath multipathd pidof rdma fio; do _have_program "$p" || return diff --git a/tests/srp/rc b/tests/srp/rc index fb3a113..252fd0c 100755 --- a/tests/srp/rc +++ b/tests/srp/rc @@ -37,6 +37,9 @@ group_requires() { fi _have_driver sd_mod _have_driver sg + _have_driver scsi_dh_alua + _have_driver scsi_dh_emc + _have_driver scsi_dh_rdac _have_module dm_multipath _have_module dm_queue_length @@ -50,9 +53,6 @@ group_requires() { _have_module rdma_cm _have_module rdma_rxe _have_module scsi_debug - _have_module scsi_dh_alua - _have_module scsi_dh_emc - _have_module scsi_dh_rdac _have_module target_core_iblock _have_module target_core_mod
nvmeof-mp and srp test groups can be executed with built-in scsi_dh_alua, scsi_dh_emc and scsi_dh_rdac drivers. Signed-off-by: Xiao Yang <yangx.jy@fujitsu.com> --- tests/nvmeof-mp/rc | 7 ++++--- tests/srp/rc | 6 +++--- 2 files changed, 7 insertions(+), 6 deletions(-)