[PATCH 5.15 207/779] erofs: avoid consecutive detection for Highmem memory

From: Greg Kroah-Hartman
Date: Mon Aug 15 2022 - 14:39:58 EST


From: Gao Xiang <hsiangkao@xxxxxxxxxxxxxxxxx>

[ Upstream commit 448b5a1548d87c246c3d0c3df8480d3c6eb6c11a ]

Currently, vmap()s are avoided if physical addresses are
consecutive for decompressed buffers.

I observed that is very common for 4KiB pclusters since the
numbers of decompressed pages are almost 2 or 3.

However, such detection doesn't work for Highmem pages on
32-bit machines, let's fix it now.

Reported-by: Liu Jinbao <liujinbao1@xxxxxxxxxx>
Fixes: 7fc45dbc938a ("staging: erofs: introduce generic decompression backend")
Link: https://lore.kernel.org/r/20220708101001.21242-1-hsiangkao@xxxxxxxxxxxxxxxxx
Signed-off-by: Gao Xiang <hsiangkao@xxxxxxxxxxxxxxxxx>
Signed-off-by: Sasha Levin <sashal@xxxxxxxxxx>
---
fs/erofs/decompressor.c | 16 ++++++++++------
1 file changed, 10 insertions(+), 6 deletions(-)

diff --git a/fs/erofs/decompressor.c b/fs/erofs/decompressor.c
index ad3f31380e6b..8193c14bb111 100644
--- a/fs/erofs/decompressor.c
+++ b/fs/erofs/decompressor.c
@@ -93,14 +93,18 @@ static int z_erofs_lz4_prepare_destpages(struct z_erofs_decompress_req *rq,

if (page) {
__clear_bit(j, bounced);
- if (kaddr) {
- if (kaddr + PAGE_SIZE == page_address(page))
+ if (!PageHighMem(page)) {
+ if (!i) {
+ kaddr = page_address(page);
+ continue;
+ }
+ if (kaddr &&
+ kaddr + PAGE_SIZE == page_address(page)) {
kaddr += PAGE_SIZE;
- else
- kaddr = NULL;
- } else if (!i) {
- kaddr = page_address(page);
+ continue;
+ }
}
+ kaddr = NULL;
continue;
}
kaddr = NULL;
--
2.35.1