DLPX-93331 stbtrace invocations crash on 24.04

[palash-gandhi]

Problem

stbtrace invocations fail with:
1.

delphix@pg-ntp:~$ sudo python3 /usr/bin/stbtrace io -a op -s avgLatency,latency,count,throughput
/virtual/main.c:45:33: error: incomplete definition of type 'struct request'
   45 |     struct gendisk *diskp = reqp->rq_disk;
      |                             ~~~~^
include/linux/blkdev.h:32:8: note: forward declaration of 'struct request'
   32 | struct request;
...
...

In this kernel commit, the struct was moved from linux/blkdev.h to linux/blk-mq.h.

2. After fixing this, I hit this error:

/virtual/main.c:46:35: error: no member named 'rq_disk' in 'struct request'

In this kernel commit, the rq_disk field of the struct was removed and replaced with q->disk.

3. After fixing this, I hit this error:

[2025-02-11T09:02:46,006][ERROR][common.TracingScriptExecutor#readLine:50][Thread-11][] There were 1 errors in the last 10 minutes for script [io]. The first <= 50 unique error lines seen were:
cannot attach kprobe, Invalid argument
[2025-02-11T09:02:46,160][WARN][analytics.impl.datacollector.TracingDataCollector#reachedEof:205][Thread-10][] Tracing process for script [io] died without collecting any data. This likely means that the script failed to compile.


delphix@dlpx-palashgandhi-os-upgrade-qar-161994-766f97e2:~$ sudo /usr/bin/stbtrace io -a op -s avgLatency,latency,count,throughput
cannot attach kprobe, Invalid argument
Failed to attach BPF program b'disk_io_done' to kprobe b'blk_account_io_done', it's not traceable (either non-existing, inlined, or marked as "notrace")

According to iovisor/bcc#5124 and the associated fix, this should be replaced with blk_mq_end_request

Solution

1. Include linux/blk-mq.h in our scripts.
2. Replace the use of rq_disk with q->disk
3. Replace blk_account_io_done with blk_mq_end_request

Testing Done

pre_checkin: https://selfservice-jenkins.eng-tools-prd.aws.delphixcloud.com/job/blackbox-self-service/162274/console

11:49:56  test_create_domain - SUCCESS (0:04:37.032303)

storage which will run the analytics_positive suite: https://selfservice-jenkins.eng-tools-prd.aws.delphixcloud.com/job/blackbox-self-service/162276/console

os_tests which run the performance_diagnostics_positive suite to test estat: https://selfservice-jenkins.eng-tools-prd.aws.delphixcloud.com/job/blackbox-self-service/162278/console