Messages in this thread | | | Date | Fri, 28 Jan 2022 16:16:53 -0800 | From | Ira Weiny <> | Subject | Re: [PATCH V8 14/44] x86/pkeys: Introduce pks_write_pkrs() |
| |
On Fri, Jan 28, 2022 at 04:12:06PM -0800, Dave Hansen wrote: > On 1/27/22 09:54, ira.weiny@intel.com wrote: > > Writing to MSR's is inefficient. Even though the underlying > > WRMSR(MSR_IA32_PKRS) is not serializing (see below), writing to the MSR > > unnecessarily should be avoided. This is especially true when the value > > of the PKS protections is unlikely to change from the default often. > > This probably needs some context. > > The most important pks_write_pkrs() user is in the scheduler, right?
This is also used during exceptions, twice. Those are probably more important.
> > So, this is really about optimizing that scheduler code for the common > case where, even when changing threads, the PKRS value does not change. > > Can you explain a bit why you expect that to be the common case?
Yes.
Ira
| |