Re: [PATCH blktests v2 2/2] [NOT-FOR-MERGE] just for testing

[Zhu Yanjun <yanjun.zhu@linux.dev>]

On 27.12.24 10:23, Zhu Yanjun wrote:
> On 27.12.24 06:37, Zhijian Li (Fujitsu) wrote:
>> Hi, Shin'ichiro,
>>
>> Your attached kconfig+this rnbd test triggered another BUG.
>>
>> Cced: RDMA
>>
>> Is this a known issue in RDMA/RXE communities?
> 
>  From my side, it should not be a known issue. It seems that it is 
> related with this rnbd test.
> 
> In previous tests, this problem does not appear.
> 
> I am not sure if others also find this problem or not. To me, it is my 
> first time to find this problem.

FYI, Zhijian

If I remember correctly, this problem does not occur in the previous 
kernel versions. Thus, it is very possible that this problem is 
introduced in recent kernel versions.

Because you have a test scenario that can reproduce this problem, as 
such, "git bisect" is a powerful tool that can help you to find the root 
cause.

Have a good luck.

Zhu Yanjun

> 
> Zhu Yanjun
> 
>>
>>
>> On 26/12/2024 21:17, Shinichiro Kawasaki wrote:
>>> On Dec 25, 2024 / 17:37, Li Zhijian wrote:
>>>> Hi, Shin'ichiro
>>>>
>>>> All your comments has been addressed except the success ratio one. 
>>>> Could
>>>> you help to check this patch([NOT-FOR-MERGE] just for testing) that 
>>>> can tell
>>>> where it fails at in your envrionment.
>>>>
>>>> I tested it today in my QEMU enviroment, It almost 100% success
>>>
>>> Thanks for this effort. I ran rnbd/001 with this series in my QEMU 
>>> environment.
>>> It looks still failing. Please find the 001.out.bad file generated 
>>> [X]. The
>>> kernel was v6.13-rc4 with the fix patch "RDMA/ulp: Add missing 
>>> deinit() call".
>>>
>>> I wonder what is the difference between your environment and mine. 
>>> FYI, my QEMU
>>> environment has 4 CPUs and 16GB DRAM. It runs Fedora 40. I also 
>>> attach the
>>> kernel config I used just in case you are interested in.
>>
>>
>> Due to this bug, I cannot finish rnbd/001 at all.
>>
>> However, I can reproduce your log by adding `_start_rnbd_client` 
>> before the iteration.
>> And it can be fixed by calling `_stop_rnbd_client` regardless of 
>> whether `_start_rnbd_client`
>> succeeds or not(Please feel free to give it a try when you have the 
>> opportunity).
>>
>> diff --git a/tests/rnbd/001 b/tests/rnbd/001
>> index 9c6d56e3ee98..321c4c010e78 100755
>> --- a/tests/rnbd/001
>> +++ b/tests/rnbd/001
>> @@ -26,6 +26,7 @@ test_start_stop()
>>           local loop_dev i j=0
>>           loop_dev="$(losetup -f)"
>> +       _start_rnbd_client    # this makes the _start_rnbd_client in 
>> below iteration fails
>>           for ((i=0;i<100;i++))
>>           do
>>                   if _start_rnbd_client "${loop_dev}" &>/dev/null; then
>> @@ -33,6 +34,7 @@ test_start_stop()
>>                           _stop_rnbd_client &>/dev/null && echo 
>> 'disconnect ok' || echo 'disconnect not ok'
>>                           ((j++))
>>                   else
>> +                       _stop_rnbd_client  # always stop rnbd so that 
>> we can connect again.
>>                           echo 'connect not ok'
>>                   fi
>>           done
>>
>> ===========================
>>
>> [   27.864420] run blktests rnbd/001 at 2024-12-27 13:21:37
>> [   27.888742] infiniband eth0_rxe: set active
>> [   27.889497] infiniband eth0_rxe: added eth0
>> [   27.910304] rnbd_client L599: Mapping device /dev/loop0 on session 
>> blktest, (access_mode: rw, nr_poll_queues: 0)
>> [   27.924065] rnbd_client L1190: [session=blktest] mapped 4/4 
>> default/read queues.
>> [   27.925825] rnbd_server L782: </dev/loop0@blktest>: Opened device 
>> 'loop0'
>> [   27.927554] rnbd_client L1612: </dev/loop0@blktest> map_device: 
>> Device mapped as rnbd0 (nsectors: 0, logical_block_size: 512, 
>> physical_block_size: 512, max_write_zeroes_sectors: 0, 
>> max_discard_sectors: 0, discard_granularity: 51
>> 2, discard_alignment: 0, secure_discard: 0, max_segments: 128, 
>> max_hw_sectors: 248, wc: 0, fua: 0)
>> [   27.938295] rnbd_client L323: </dev/loop0@blktest> Unmapping 
>> device, option: normal.
>> [   27.962570] rnbd_server L238: </dev/loop0@blktest>: Device closed
>> [   27.967500] BUG: kernel NULL pointer dereference, address: 
>> 0000000000000000
>> [   27.967500] BUG: kernel NULL pointer dereference, address: 
>> 0000000000000000                                                                                                                                       13:21:38 [11/9189]
>> [   27.976554] #PF: supervisor read access in kernel mode
>> [   27.984926] #PF: error_code(0x0000) - not-present page
>> [   27.989126] PGD 0 P4D 0
>> [   27.991067] Oops: Oops: 0000 [#1] PREEMPT SMP PTI
>> [   27.993226] CPU: 3 UID: 0 PID: 304 Comm: kworker/u20:2 Not tainted 
>> 6.13.0-rc3+ #1
>> [   27.996697] Hardware name: QEMU Standard PC (Q35 + ICH9, 2009), 
>> BIOS rel-1.16.3-0-ga6ed6b701f0a-prebuilt.qemu.org 04/01/2014
>> [   27.999333] Workqueue: rxe_wq do_work [rdma_rxe]
>> [   28.000309] RIP: 0010:memcpy_orig+0xd5/0x140
>> [   28.001304] Code: 16 f8 4c 89 07 4c 89 4f 08 4c 89 54 17 f0 4c 89 
>> 5c 17 f8 c3 cc cc cc cc 66 66 2e 0f 1f 84 00 00 00 00 00 66 90 83 fa 
>> 08 72 1b <4c> 8b 06 4c 8b 4c 16 f8 4c 89 07 4c 89 4c 17 f8 c3 cc cc cc 
>> cc 66
>> [   28.004932] RSP: 0018:ffffb934c0643cc0 EFLAGS: 00010246
>> [   28.005845] RAX: ffff976bc1e12d5a RBX: 0000000000000000 RCX: 
>> 0000000000000000
>> [   28.007090] RDX: 0000000000000008 RSI: 0000000000000000 RDI: 
>> ffff976bc1e12d5a
>> [   28.008380] RBP: ffff976bc1e12d5a R08: 0000000000000001 R09: 
>> 0000000000000001
>> [   28.009639] R10: 0000000000000005 R11: 0000000000000000 R12: 
>> 0000000080000000
>> [   28.010836] R13: 0000000000000008 R14: 0000000000000008 R15: 
>> 0000000000000008
>> [   28.011948] FS:  0000000000000000(0000) GS:ffff976f2fd80000(0000) 
>> knlGS:0000000000000000
>> [   28.013335] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
>> [   28.014275] CR2: 0000000000000000 CR3: 00000001837da002 CR4: 
>> 00000000001706f0
>> [   28.015424] DR0: 0000000000000000 DR1: 0000000000000000 DR2: 
>> 0000000000000000
>> [   28.016598] DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 
>> 0000000000000400
>> [   28.017728] Call Trace:
>> [   28.018114]  <TASK>
>> [   28.018453]  ? __die_body.cold+0x19/0x27
>> [   28.019167]  ? page_fault_oops+0x15a/0x2d0
>> [   28.019861]  ? search_module_extables+0x19/0x60
>> [   28.020617]  ? search_bpf_extables+0x5f/0x80
>> [   28.021611]  ? exc_page_fault+0x7e/0x180
>> [   28.022488]  ? asm_exc_page_fault+0x26/0x30
>> [   28.023547]  ? memcpy_orig+0xd5/0x140
>> [   28.024396]  rxe_mr_copy+0x1c3/0x200 [rdma_rxe]
>> [   28.025476]  ? rxe_pool_get_index+0x4b/0x80 [rdma_rxe]
>> [   28.026612]  copy_data+0xa5/0x230 [rdma_rxe]
>> [   28.027611]  rxe_requester+0xd9b/0xf70 [rdma_rxe]
>> [   28.028727]  ? finish_task_switch.isra.0+0x99/0x2e0
>> [   28.029878]  rxe_sender+0x13/0x40 [rdma_rxe]
>> [   28.030920]  do_task+0x68/0x1e0 [rdma_rxe]
>> [   28.031893]  process_one_work+0x177/0x330
>> [   28.032854]  worker_thread+0x252/0x390
>> [   28.033748]  ? __pfx_worker_thread+0x10/0x10
>> [   28.034665]  kthread+0xd2/0x100
>> [   28.035382]  ? __pfx_kthread+0x10/0x10
>> [   28.036252]  ret_from_fork+0x34/0x50
>> [   28.037220]  ? __pfx_kthread+0x10/0x10
>> [   28.038072]  ret_from_fork_asm+0x1a/0x30
>> [   28.038991]  </TASK>
>> [   28.039543] Modules linked in: loop rnbd_client rtrs_client 
>> rnbd_server rtrs_server rtrs_core rdma_cm iw_cm ib_cm rdma_rxe 
>> ib_uverbs ib_core ip6_udp_tunnel udp_tunnel rfkill intel_rapl_msr 
>> intel_rapl_common kmem rapl cxl_mem iTC
>> O_wdt intel_pmc_bxt cxl_pmem dax_hmem iTCO_vendor_support device_dax 
>> cxl_acpi cxl_pci cxl_port joydev qxl cxl_core pcspkr drm_ttm_helper 
>> lpc_ich ttm i2c_i801 virtio_balloon i2c_smbus nd_pmem nd_btt dax_pmem 
>> einj ip_tables crct10dif
>> _pclmul crc32_pclmul crc32c_intel polyval_clmulni polyval_generic 
>> ghash_clmulni_intel sha512_ssse3 sha256_ssse3 sha1_ssse3 virtiofs fuse 
>> virtio_net nfit virtio_console net_failover libnvdimm serio_raw 
>> virtio_blk failover qemu_fw_cf
>> g dm_multipath sunrpc
>> [   28.051034] CR2: 0000000000000000
>> [   28.052072] ---[ end trace 0000000000000000 ]---
>> [   28.053099] RIP: 0010:memcpy_orig+0xd5/0x140
>> [   28.054188] Code: 16 f8 4c 89 07 4c 89 4f 08 4c 89 54 17 f0 4c 89 
>> 5c 17 f8 c3 cc cc cc cc 66 66 2e 0f 1f 84 00 00 00 00 00 66 90 83 fa 
>> 08 72 1b <4c> 8b 06 4c 8b 4c 16 f8 4c 89 07 4c 89 4c 17 f8 c3 cc cc cc 
>> cc 66
>> [   28.058290] RSP: 0018:ffffb934c0643cc0 EFLAGS: 00010246
>> [   28.059514] RAX: ffff976bc1e12d5a RBX: 0000000000000000 RCX: 
>> 0000000000000000
>> [   28.061194] RDX: 0000000000000008 RSI: 0000000000000000 RDI: 
>> ffff976bc1e12d5a
>> [   28.062588] RBP: ffff976bc1e12d5a R08: 0000000000000001 R09: 
>> 0000000000000001
>>
>>
>>
>>
>>>
>>>
>>> [X]
>>>
>>> 001.out.bad
>>> ----------------------------------------------------------------------------
>>> Running rnbd/001
>>> connect ok
>>> disconnect not ok
>>> connect not ok
>>> connect not ok
>>> connect not ok
>>> connect not ok
>>> connect not ok
>>> connect not ok
>>> connect not ok
>>> connect not ok
>>> connect not ok
>>> connect not ok
>>> connect not ok
>>> connect not ok
>>> connect not ok
>>> connect not ok
>>> connect not ok
>>> connect not ok
>>> connect not ok
>>> connect not ok
>>> connect not ok
>>> connect not ok
>>> connect not ok
>>> connect not ok
>>> connect not ok
>>> connect not ok
>>> connect not ok
>>> connect not ok
>>> connect not ok
>>> connect not ok
>>> connect not ok
>>> connect not ok
>>> connect not ok
>>> connect not ok
>>> connect not ok
>>> connect not ok
>>> connect not ok
>>> connect not ok
>>> connect not ok
>>> connect not ok
>>> connect not ok
>>> connect not ok
>>> connect not ok
>>> connect not ok
>>> connect not ok
>>> connect not ok
>>> connect not ok
>>> connect not ok
>>> connect not ok
>>> connect not ok
>>> connect not ok
>>> connect not ok
>>> connect not ok
>>> connect not ok
>>> connect not ok
>>> connect not ok
>>> connect not ok
>>> connect not ok
>>> connect not ok
>>> connect not ok
>>> connect not ok
>>> connect not ok
>>> connect not ok
>>> connect not ok
>>> connect not ok
>>> connect not ok
>>> connect not ok
>>> connect not ok
>>> connect not ok
>>> connect not ok
>>> connect not ok
>>> connect not ok
>>> connect not ok
>>> connect not ok
>>> connect not ok
>>> connect not ok
>>> connect not ok
>>> connect not ok
>>> connect not ok
>>> connect not ok
>>> connect not ok
>>> connect not ok
>>> connect not ok
>>> connect not ok
>>> connect not ok
>>> connect not ok
>>> connect not ok
>>> connect not ok
>>> connect not ok
>>> connect not ok
>>> connect not ok
>>> connect not ok
>>> connect not ok
>>> connect not ok
>>> connect not ok
>>> connect not ok
>>> connect not ok
>>> connect not ok
>>> connect not ok
>>> connect not ok
>>> connect not ok
>>> Failed: 1/100
>>> Test complete
> 

-- 
Best Regards,
Yanjun.Zhu