On Fri, Mar 04, 2016 at 03:09:00PM +0000, Lars Persson wrote:
> > 1 mars 2016 kl. 03:38 Paul Burton <paul.burton@xxxxxxxxxx>:
> >
> > The flush_kernel_dcache_page function was previously essentially a nop.
> > This is incorrect for MIPS, where if a page has been modified & either
> > it aliases or it's executable & the icache doesn't fill from dcache then
> > the content needs to be written back from dcache to the next level of
> > the cache hierarchy (which is shared with the icache).
> >
> > Implement this by simply calling flush_dcache_page, treating this
> > kmapped cache flush function (flush_kernel_dcache_page) exactly the same
> > as its non-kmapped counterpart (flush_dcache_page).
> >
> > Signed-off-by: Paul Burton <paul.burton@xxxxxxxxxx>
> > ---
> >
> > arch/mips/include/asm/cacheflush.h | 1 +
> > 1 file changed, 1 insertion(+)
> >
> > diff --git a/arch/mips/include/asm/cacheflush.h b/arch/mips/include/asm/cacheflush.h
> > index 723229f..7e9f468 100644
> > --- a/arch/mips/include/asm/cacheflush.h
> > +++ b/arch/mips/include/asm/cacheflush.h
> > @@ -132,6 +132,7 @@ static inline void kunmap_noncoherent(void)
> > static inline void flush_kernel_dcache_page(struct page *page)
> > {
> > BUG_ON(cpu_has_dc_aliases && PageHighMem(page));
> > + flush_dcache_page(page);
>
> Should we use instead __flush_dcache_page() that flushes immediately for mapped pages ? Steven J Hill's old patch set for highmem had done it like this.
Delayed flushing should be ok for lowmem where each page has a permanent
virtual address. With highmem the temporary address assigned by the kmap_*
function may change so the flush needs to be performed immediately.
Special case highmem without cache aliases - the exact virtual address
doesn't matter, so this should be fine.
Cache flushes are expensive so delaying if possible is always a good thing.
Steven's patches afair were trying to tackle the highmem with aliases case
so an immediately flush was required.
Ralf
