On 11.02.20 11:37, Dario Faggioli wrote:
On Mon, 2020-02-10 at 16:39 +0100, Juergen Gross wrote:
sched_init_pdata() is used nowhere, it can be removed. Same applies
to
the .init_pdata hook of the per-scheduler interface.
Right, and that appear to be the case since
f855dd962523b6cb47a92037bdd28b1485141abe ("sched: add minimalistic idle
scheduler for free cpus"), which removed all call sites.
And that is because switching to a scheduler always happens via
switch_sched from the idle scheduler, and it's there that we do all the
initializations, right?
This change is obviously doing the right thing, removing code that is
never called! :-)
Can we, though:
- add a mention to the commit above and a quick explanation of things
in the changelog?
Okay.
- update the following comments too:
1) in cpu_schedule_callback()
"* This happens by calling the deinit_pdata and free_pdata hooks, in this
* order. If no per-pCPU memory was allocated, there is no need to
* provide an implementation of free_pdata. deinit_pdata may, however,
* be necessary/useful in this case too (e.g., it can undo something done
* on scheduler wide data structure during init_pdata). Both deinit_pdata
* and free_pdata are called during CPU_DEAD."
2) schedule_cpu_add()
"* - a valid instance of per-CPU scheduler specific data, as it is
* allocated by sched_alloc_pdata(). Note that we do not want to
* initialize it yet (i.e., we are not calling sched_init_pdata()).
* That will be done by the target scheduler, in sched_switch_sched(),
* in proper ordering and with locking."
Oh, I missed those. Will modify the comments.
Juergen
_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xenproject.org
https://lists.xenproject.org/mailman/listinfo/xen-devel
