The patch titled
Subject: zram: don't use highmem for the bounce buffer in zram_bvec_{read,write}
has been added to the -mm mm-unstable branch. Its filename is
zram-dont-use-highmem-for-the-bounce-buffer-in-zram_bvec_readwrite.patch
This patch will shortly appear at
https://git.kernel.org/pub/scm/linux/kernel/git/akpm/25-new.git/tree/patches/zram-dont-use-highmem-for-the-bounce-buffer-in-zram_bvec_readwrite.patch
This patch will later appear in the mm-unstable branch at
git://git.kernel.org/pub/scm/linux/kernel/git/akpm/mm
Before you just go and hit "reply", please:
a) Consider who else should be cc'ed
b) Prefer to cc a suitable mailing list as well
c) Ideally: find the original patch on the mailing list and do a
reply-to-all to that, adding suitable additional cc's
*** Remember to use Documentation/process/submit-checklist.rst when testing your code ***
The -mm tree is included into linux-next via the mm-everything
branch at git://git.kernel.org/pub/scm/linux/kernel/git/akpm/mm
and is updated there every 2-3 working days
------------------------------------------------------
From: Christoph Hellwig <hch@xxxxxx>
Subject: zram: don't use highmem for the bounce buffer in zram_bvec_{read,write}
Date: Thu, 6 Apr 2023 16:40:53 +0200
There is no point in allocation a highmem page when we instantly need to
copy from it.
Link: https://lkml.kernel.org/r/20230406144102.149231-8-hch@xxxxxx
Signed-off-by: Christoph Hellwig <hch@xxxxxx>
Reviewed-by: Sergey Senozhatsky <senozhatsky@xxxxxxxxxxxx>
Acked-by: Minchan Kim <minchan@xxxxxxxxxx>
Cc: Jens Axboe <axboe@xxxxxxxxx>
Signed-off-by: Andrew Morton <akpm@xxxxxxxxxxxxxxxxxxxx>
---
drivers/block/zram/zram_drv.c | 17 +++++------------
1 file changed, 5 insertions(+), 12 deletions(-)
--- a/drivers/block/zram/zram_drv.c~zram-dont-use-highmem-for-the-bounce-buffer-in-zram_bvec_readwrite
+++ a/drivers/block/zram/zram_drv.c
@@ -1437,7 +1437,7 @@ static int zram_bvec_read(struct zram *z
page = bvec->bv_page;
if (is_partial_io(bvec)) {
/* Use a temporary buffer to decompress the page */
- page = alloc_page(GFP_NOIO|__GFP_HIGHMEM);
+ page = alloc_page(GFP_NOIO);
if (!page)
return -ENOMEM;
}
@@ -1446,12 +1446,8 @@ static int zram_bvec_read(struct zram *z
if (unlikely(ret))
goto out;
- if (is_partial_io(bvec)) {
- void *src = kmap_atomic(page);
-
- memcpy_to_bvec(bvec, src + offset);
- kunmap_atomic(src);
- }
+ if (is_partial_io(bvec))
+ memcpy_to_bvec(bvec, page_address(page) + offset);
out:
if (is_partial_io(bvec))
__free_page(page);
@@ -1595,12 +1591,11 @@ static int zram_bvec_write(struct zram *
vec = *bvec;
if (is_partial_io(bvec)) {
- void *dst;
/*
* This is a partial IO. We need to read the full page
* before to write the changes.
*/
- page = alloc_page(GFP_NOIO|__GFP_HIGHMEM);
+ page = alloc_page(GFP_NOIO);
if (!page)
return -ENOMEM;
@@ -1608,9 +1603,7 @@ static int zram_bvec_write(struct zram *
if (ret)
goto out;
- dst = kmap_atomic(page);
- memcpy_from_bvec(dst + offset, bvec);
- kunmap_atomic(dst);
+ memcpy_from_bvec(page_address(page) + offset, bvec);
bvec_set_page(&vec, page, PAGE_SIZE, 0);
}
_
Patches currently in -mm which might be from hch@xxxxxx are
zram-remove-valid_io_request.patch
zram-make-zram_bio_discard-more-self-contained.patch
zram-simplify-bvec-iteration-in-__zram_make_request.patch
zram-move-discard-handling-to-zram_submit_bio.patch
zram-return-early-on-error-in-zram_bvec_rw.patch
zram-refactor-highlevel-read-and-write-handling.patch
zram-dont-use-highmem-for-the-bounce-buffer-in-zram_bvec_readwrite.patch
zram-rename-__zram_bvec_read-to-zram_read_page.patch
zram-directly-call-zram_read_page-in-writeback_store.patch
zram-refactor-zram_bdev_read.patch
zram-dont-pass-a-bvec-to-__zram_bvec_write.patch
zram-refactor-zram_bdev_write.patch
zram-pass-a-page-to-read_from_bdev.patch
zram-dont-return-errors-from-read_from_bdev_async.patch
zram-fix-synchronous-reads.patch
zram-return-errors-from-read_from_bdev_sync.patch
