On Fri, Dec 16, 2022 at 08:53:44PM +0000, Matthew Wilcox (Oracle) wrote:
> Remove uses of kmap() and b_page. Also move the set_buffer_uptodate()
> call to after the memcpy() so that the memory barrier in
> set_buffer_uptodate() actually ensures that the old data can't be visible.
>
> Signed-off-by: Matthew Wilcox (Oracle) <willy@xxxxxxxxxxxxx>
> ---
> fs/reiserfs/journal.c | 12 +++++-------
> 1 file changed, 5 insertions(+), 7 deletions(-)
>
> diff --git a/fs/reiserfs/journal.c b/fs/reiserfs/journal.c
> index 9ce4ec296b74..faf2f09d82e1 100644
> --- a/fs/reiserfs/journal.c
> +++ b/fs/reiserfs/journal.c
> @@ -4200,21 +4200,19 @@ static int do_journal_end(struct reiserfs_transaction_handle *th, int flags)
> /* copy all the real blocks into log area. dirty log blocks */
> if (buffer_journaled(cn->bh)) {
> struct buffer_head *tmp_bh;
> + size_t offset = offset_in_folio(cn->bh->b_folio,
> + cn->bh->b_data);
> char *addr;
> - struct page *page;
> tmp_bh =
> journal_getblk(sb,
> SB_ONDISK_JOURNAL_1st_BLOCK(sb) +
> ((cur_write_start +
> jindex) %
> SB_ONDISK_JOURNAL_SIZE(sb)));
> + addr = kmap_local_folio(cn->bh->b_folio, offset);
> + memcpy(tmp_bh->b_data, addr, cn->bh->b_size);
> + kunmap_local(addr);
I think we should have a memcpy_{to|from}_folio() like we do for the pages.
Did I miss this in the earlier patch?
Ira
> set_buffer_uptodate(tmp_bh);
> - page = cn->bh->b_page;
> - addr = kmap(page);
> - memcpy(tmp_bh->b_data,
> - addr + offset_in_page(cn->bh->b_data),
> - cn->bh->b_size);
> - kunmap(page);
> mark_buffer_dirty(tmp_bh);
> jindex++;
> set_buffer_journal_dirty(cn->bh);
> --
> 2.35.1
>
