]> xenbits.xensource.com Git - qemu-xen.git/commit
translate-all: work page-by-page in tb_invalidate_phys_range_1
authorEmilio G. Cota <cota@braap.org>
Sat, 5 Aug 2017 06:04:40 +0000 (02:04 -0400)
committerRichard Henderson <richard.henderson@linaro.org>
Fri, 15 Jun 2018 17:42:55 +0000 (07:42 -1000)
commitae5486e273a4e368515a963a6d0076e20453eb72
tree764f8fa694532aa155c9b153ef106fc1441ae50d
parent94da9aec2a50f0c82e6c60939275c0337f03d5fe
translate-all: work page-by-page in tb_invalidate_phys_range_1

So that we pass a same-page range to tb_invalidate_phys_page_range,
instead of always passing an end address that could be on a different
page.

As discussed with Peter Maydell on the list [1], tb_invalidate_phys_page_range
doesn't actually do much with 'end', which explains why we have never
hit a bug despite going against what the comment on top of
tb_invalidate_phys_page_range requires:

> * Invalidate all TBs which intersect with the target physical address range
> * [start;end[. NOTE: start and end must refer to the *same* physical page.

The appended honours the comment, which avoids confusion.

While at it, rework the loop into a for loop, which is less error prone
(e.g. "continue" won't result in an infinite loop).

[1] https://lists.gnu.org/archive/html/qemu-devel/2017-07/msg09165.html

Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Signed-off-by: Emilio G. Cota <cota@braap.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
accel/tcg/translate-all.c