Thread (80 messages) 80 messages, 8 authors, 2025-03-03

Re: [PATCH v2 3/2] sched/deadline: Check bandwidth overflow earlier for hotplug

From: Juri Lelli <juri.lelli@redhat.com>
Date: 2025-02-19 10:02:53
Also in: linux-tegra, lkml

On 19/02/25 10:29, Dietmar Eggemann wrote:

...
I did now. 
Thanks!
Patch-wise I have:

(1) Putting 'fair_server's __dl_server_[de|at]tach_root() under if  
    '(cpumask_test_cpu(rq->cpu, [old_rd->online|cpu_active_mask))' in 
    rq_attach_root()

    https://lkml.kernel.org/r/Z7RhNmLpOb7SLImW@jlelli-thinkpadt14gen4.remote.csb

(2) Create __dl_server_detach_root() and call it in rq_attach_root()

    https://lkml.kernel.org/r/Z4fd_6M2vhSMSR0i@jlelli-thinkpadt14gen4.remote.csb

plus debug patch:

    https://lkml.kernel.org/r/Z6M5fQB9P1_bDF7A@jlelli-thinkpadt14gen4.remote.csb

plus additional debug.
So you don't have the one with which we ignore special tasks while
rebuilding domains?

https://lore.kernel.org/all/Z6spnwykg6YSXBX_@jlelli-thinkpadt14gen4.remote.csb/ (local)

Could you please double check again against

git@github.com:jlelli/linux.git experimental/dl-debug
The suspend issue still persists.

My hunch is that it's rather an issue with having 0 CPUs left in DEF
while deactivating the last isol CPU (CPU3) so we set overflow = 1 w/o
calling __dl_overflow(). We want to account fair_server_bw=52428
against 0 CPUs. 

l B B l l l

      ^^^
      isolcpus=[3,4]


cpumask_and(mask, rd->span, cpu_active_mask)

mask = [3-5] & [0-3] = [3] -> dl_bw_cpus(3) = 1

---

dl_bw_deactivate() called cpu=5

dl_bw_deactivate() called cpu=4

dl_bw_deactivate() called cpu=3

dl_bw_cpus() cpu=6 rd->span=3-5 cpu_active_mask=0-3 cpus=1 type=DEF
                   ^^^^^^^^^^^^ ^^^^^^^^^^^^^^^^^^^
  cpumask_subset(rd->span, cpu_active_mask) is false

  for_each_cpu_and(i, rd->span, cpu_active_mask)
    cpus++                                       <-- cpus is 1 !!!

dl_bw_manage: cpu=3 cap=0 fair_server_bw=52428 total_bw=104856 dl_bw_cpus=1 type=DEF span=3-5
                                                          ^^^^^^
This still looks wrong: with a single cpu remaining we should only have
the corresponding dl server bandwidth present (unless there is some
other DL task running.

If you already had the patch ignoring sugovs bandwidth in your set, could
you please share the full dmesg?

Thanks!
Keyboard shortcuts
hback out one level
jnext message in thread
kprevious message in thread
ldrill in
Escclose help / fold thread tree
?toggle this help