[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Xen-devel] [PATCH] xen/arm: Fix rtds scheduler for arm

>>> On 02.02.15 at 13:16, <Ian.Campbell@xxxxxxxxxx> wrote:
> On Mon, 2015-02-02 at 11:40 +0000, Jan Beulich wrote:
>>  X86 calls context_saved() before
>> schedule_tail() and has no need for IRQs to be disabled after
>> __context_switch() returned.
> In fact from the comment ("...which may fault...") I think it requires
> that they are not disabled?

Correct, or else such a fault would be fatal.

>>  Otoh some roughly equivalent stuff
>> ARM does in ctxt_switch_{from,to}() is being done in
>> __context_switch() on x86 (and in the context here the restore
>> parts seem to be of most interest) - maybe the call to
>> context_saved() could be deferred on ARM until after IRQs got
>> re-enabled?
> Is it allowable to have the context_saved of prev be after (most of) the
> state of next has been restored?

I think so, but I'd clearly like this to be confirmed by e.g. George,
not the least because it's only the newer schedulers that make use
of this.


Xen-devel mailing list



Lists.xenproject.org is hosted with RackSpace, monitoring our
servers 24x7x365 and backed by RackSpace's Fanatical Support®.