On Wed, Nov 22, 2017 at 08:23:38PM +0100, Christoffer Dall wrote:
> Hi Dave,
>
> On Fri, Nov 17, 2017 at 04:38:53PM +, Dave Martin wrote:
> > When deciding whether to invalidate FPSIMD state cached in the cpu,
> > the backend function sve_flush_cpu_state() attempts to dereference
> > __this_cp
Hi Dave,
On Fri, Nov 17, 2017 at 04:38:53PM +, Dave Martin wrote:
> When deciding whether to invalidate FPSIMD state cached in the cpu,
> the backend function sve_flush_cpu_state() attempts to dereference
> __this_cpu_read(fpsimd_last_state). However, this is not safe:
> there is no guarantee
When deciding whether to invalidate FPSIMD state cached in the cpu,
the backend function sve_flush_cpu_state() attempts to dereference
__this_cpu_read(fpsimd_last_state). However, this is not safe:
there is no guarantee that the pointer is still valid, because the
task could have exited in the mea