redonkable/remarkable-linux
redonkable/remarkable-linux
1
0
Fork 0
Code Releases Activity

arm64/lib: copy_page: use consistent prefetch stride

Browse source 
The optional prefetch instructions in the copy_page() routine are
inconsistent: at the start of the function, two cachelines are
prefetched beyond the one being loaded in the first iteration, but
in the loop, the prefetch is one more line ahead. This appears to
be unintentional, so let's fix it.

While at it, fix the comment style and white space.

Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Signed-off-by: Will Deacon <will.deacon@arm.com>
This commit is contained in:
Ard Biesheuvel 2017-07-12 15:44:14 +01:00 committed by Will Deacon
parent ece4b206be
commit 288be97cc7
1 changed files with 5 additions and 4 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

9
arch/arm64/lib/copy_page.S
View file

@ -30,9 +30,10 @@
*/
ENTRY(copy_page)
alternative_if ARM64_HAS_NO_HW_PREFETCH
# Prefetch two cache lines ahead.
prfm pldl1strm, [x1, #128]
prfm pldl1strm, [x1, #256]
// Prefetch three cache lines ahead.
prfm pldl1strm, [x1, #128]
prfm pldl1strm, [x1, #256]
prfm pldl1strm, [x1, #384]
alternative_else_nop_endif
ldp x2, x3, [x1]
@ -50,7 +51,7 @@ alternative_else_nop_endif
subs x18, x18, #128
alternative_if ARM64_HAS_NO_HW_PREFETCH
prfm pldl1strm, [x1, #384]
prfm pldl1strm, [x1, #384]
alternative_else_nop_endif
stnp x2, x3, [x0]