| Message ID | 20210715033704.692967-125-willy@infradead.org (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
| Series | Memory folios | expand |
On Thu, Jul 15, 2021 at 04:36:50AM +0100, Matthew Wilcox (Oracle) wrote: > We still only operate on a single page of data at a time due to using > kmap(). A more complex implementation would work on each page in a folio, > but it's not clear that such a complex implementation would be worthwhile. Does this break up a compound folio into smaller pages? > Signed-off-by: Matthew Wilcox (Oracle) <willy@infradead.org> > --- > fs/remap_range.c | 116 ++++++++++++++++++++++------------------------- > 1 file changed, 55 insertions(+), 61 deletions(-) > > diff --git a/fs/remap_range.c b/fs/remap_range.c > index e4a5fdd7ad7b..886e6ed2c6c2 100644 > --- a/fs/remap_range.c > +++ b/fs/remap_range.c > @@ -158,41 +158,41 @@ static int generic_remap_check_len(struct inode *inode_in, > } > > /* Read a page's worth of file data into the page cache. */ > -static struct page *vfs_dedupe_get_page(struct inode *inode, loff_t offset) > +static struct folio *vfs_dedupe_get_folio(struct inode *inode, loff_t pos) > { > - struct page *page; > + struct folio *folio; > > - page = read_mapping_page(inode->i_mapping, offset >> PAGE_SHIFT, NULL); > - if (IS_ERR(page)) > - return page; > - if (!PageUptodate(page)) { > - put_page(page); > + folio = read_mapping_folio(inode->i_mapping, pos >> PAGE_SHIFT, NULL); > + if (IS_ERR(folio)) > + return folio; > + if (!folio_test_uptodate(folio)) { > + folio_put(folio); > return ERR_PTR(-EIO); > } > - return page; > + return folio; > } > > /* > - * Lock two pages, ensuring that we lock in offset order if the pages are from > - * the same file. > + * Lock two folios, ensuring that we lock in offset order if the folios > + * are from the same file. > */ > -static void vfs_lock_two_pages(struct page *page1, struct page *page2) > +static void vfs_lock_two_folios(struct folio *folio1, struct folio *folio2) > { > /* Always lock in order of increasing index. */ > - if (page1->index > page2->index) > - swap(page1, page2); > + if (folio1->index > folio2->index) > + swap(folio1, folio2); > > - lock_page(page1); > - if (page1 != page2) > - lock_page(page2); > + folio_lock(folio1); > + if (folio1 != folio2) > + folio_lock(folio2); > } > > -/* Unlock two pages, being careful not to unlock the same page twice. */ > -static void vfs_unlock_two_pages(struct page *page1, struct page *page2) > +/* Unlock two folios, being careful not to unlock the same folio twice. */ > +static void vfs_unlock_two_folios(struct folio *folio1, struct folio *folio2) > { > - unlock_page(page1); > - if (page1 != page2) > - unlock_page(page2); > + folio_unlock(folio1); > + if (folio1 != folio2) > + folio_unlock(folio2); This could result in a lot of folio lock cycling. Do you think it's worth the effort to minimize this by keeping the folio locked if the next page is going to be from the same one? --D > } > > /* > @@ -200,77 +200,71 @@ static void vfs_unlock_two_pages(struct page *page1, struct page *page2) > * Caller must have locked both inodes to prevent write races. > */ > static int vfs_dedupe_file_range_compare(struct inode *src, loff_t srcoff, > - struct inode *dest, loff_t destoff, > + struct inode *dest, loff_t dstoff, > loff_t len, bool *is_same) > { > - loff_t src_poff; > - loff_t dest_poff; > - void *src_addr; > - void *dest_addr; > - struct page *src_page; > - struct page *dest_page; > - loff_t cmp_len; > - bool same; > - int error; > - > - error = -EINVAL; > - same = true; > + bool same = true; > + int error = -EINVAL; > + > while (len) { > - src_poff = srcoff & (PAGE_SIZE - 1); > - dest_poff = destoff & (PAGE_SIZE - 1); > - cmp_len = min(PAGE_SIZE - src_poff, > - PAGE_SIZE - dest_poff); > + struct folio *src_folio, *dst_folio; > + void *src_addr, *dst_addr; > + loff_t cmp_len = min(PAGE_SIZE - offset_in_page(srcoff), > + PAGE_SIZE - offset_in_page(dstoff)); > + > cmp_len = min(cmp_len, len); > if (cmp_len <= 0) > goto out_error; > > - src_page = vfs_dedupe_get_page(src, srcoff); > - if (IS_ERR(src_page)) { > - error = PTR_ERR(src_page); > + src_folio = vfs_dedupe_get_folio(src, srcoff); > + if (IS_ERR(src_folio)) { > + error = PTR_ERR(src_folio); > goto out_error; > } > - dest_page = vfs_dedupe_get_page(dest, destoff); > - if (IS_ERR(dest_page)) { > - error = PTR_ERR(dest_page); > - put_page(src_page); > + dst_folio = vfs_dedupe_get_folio(dest, dstoff); > + if (IS_ERR(dst_folio)) { > + error = PTR_ERR(dst_folio); > + folio_put(src_folio); > goto out_error; > } > > - vfs_lock_two_pages(src_page, dest_page); > + vfs_lock_two_folios(src_folio, dst_folio); > > /* > - * Now that we've locked both pages, make sure they're still > + * Now that we've locked both folios, make sure they're still > * mapped to the file data we're interested in. If not, > * someone is invalidating pages on us and we lose. > */ > - if (!PageUptodate(src_page) || !PageUptodate(dest_page) || > - src_page->mapping != src->i_mapping || > - dest_page->mapping != dest->i_mapping) { > + if (!folio_test_uptodate(src_folio) || !folio_test_uptodate(dst_folio) || > + src_folio->mapping != src->i_mapping || > + dst_folio->mapping != dest->i_mapping) { > same = false; > goto unlock; > } > > - src_addr = kmap_atomic(src_page); > - dest_addr = kmap_atomic(dest_page); > + src_addr = kmap_local_folio(src_folio, > + offset_in_folio(src_folio, srcoff)); > + dst_addr = kmap_local_folio(dst_folio, > + offset_in_folio(dst_folio, dstoff)); > > - flush_dcache_page(src_page); > - flush_dcache_page(dest_page); > + flush_dcache_folio(src_folio); > + flush_dcache_folio(dst_folio); > > - if (memcmp(src_addr + src_poff, dest_addr + dest_poff, cmp_len)) > + if (memcmp(src_addr, dst_addr, cmp_len)) > same = false; > > - kunmap_atomic(dest_addr); > - kunmap_atomic(src_addr); > + kunmap_local(dst_addr); > + kunmap_local(src_addr); > unlock: > - vfs_unlock_two_pages(src_page, dest_page); > - put_page(dest_page); > - put_page(src_page); > + vfs_unlock_two_folios(src_folio, dst_folio); > + folio_put(dst_folio); > + folio_put(src_folio); > > if (!same) > break; > > srcoff += cmp_len; > - destoff += cmp_len; > + dstoff += cmp_len; > len -= cmp_len; > } > > -- > 2.30.2 >
On Thu, Jul 15, 2021 at 03:08:40PM -0700, Darrick J. Wong wrote: > On Thu, Jul 15, 2021 at 04:36:50AM +0100, Matthew Wilcox (Oracle) wrote: > > We still only operate on a single page of data at a time due to using > > kmap(). A more complex implementation would work on each page in a folio, > > but it's not clear that such a complex implementation would be worthwhile. > > Does this break up a compound folio into smaller pages? No. We just operate on each page in turn. Splitting a folio is an expensive and unrealiable thing to do, so we avoid it unless necessary. > > +/* Unlock two folios, being careful not to unlock the same folio twice. */ > > +static void vfs_unlock_two_folios(struct folio *folio1, struct folio *folio2) > > { > > - unlock_page(page1); > > - if (page1 != page2) > > - unlock_page(page2); > > + folio_unlock(folio1); > > + if (folio1 != folio2) > > + folio_unlock(folio2); > > This could result in a lot of folio lock cycling. Do you think it's > worth the effort to minimize this by keeping the folio locked if the > next page is going to be from the same one? I think that might well be a worthwhile optimisation. I'd like to do that as a separate patch, though (and maybe somebody other than me could do it ;-)
diff --git a/fs/remap_range.c b/fs/remap_range.c index e4a5fdd7ad7b..886e6ed2c6c2 100644 --- a/fs/remap_range.c +++ b/fs/remap_range.c @@ -158,41 +158,41 @@ static int generic_remap_check_len(struct inode *inode_in, } /* Read a page's worth of file data into the page cache. */ -static struct page *vfs_dedupe_get_page(struct inode *inode, loff_t offset) +static struct folio *vfs_dedupe_get_folio(struct inode *inode, loff_t pos) { - struct page *page; + struct folio *folio; - page = read_mapping_page(inode->i_mapping, offset >> PAGE_SHIFT, NULL); - if (IS_ERR(page)) - return page; - if (!PageUptodate(page)) { - put_page(page); + folio = read_mapping_folio(inode->i_mapping, pos >> PAGE_SHIFT, NULL); + if (IS_ERR(folio)) + return folio; + if (!folio_test_uptodate(folio)) { + folio_put(folio); return ERR_PTR(-EIO); } - return page; + return folio; } /* - * Lock two pages, ensuring that we lock in offset order if the pages are from - * the same file. + * Lock two folios, ensuring that we lock in offset order if the folios + * are from the same file. */ -static void vfs_lock_two_pages(struct page *page1, struct page *page2) +static void vfs_lock_two_folios(struct folio *folio1, struct folio *folio2) { /* Always lock in order of increasing index. */ - if (page1->index > page2->index) - swap(page1, page2); + if (folio1->index > folio2->index) + swap(folio1, folio2); - lock_page(page1); - if (page1 != page2) - lock_page(page2); + folio_lock(folio1); + if (folio1 != folio2) + folio_lock(folio2); } -/* Unlock two pages, being careful not to unlock the same page twice. */ -static void vfs_unlock_two_pages(struct page *page1, struct page *page2) +/* Unlock two folios, being careful not to unlock the same folio twice. */ +static void vfs_unlock_two_folios(struct folio *folio1, struct folio *folio2) { - unlock_page(page1); - if (page1 != page2) - unlock_page(page2); + folio_unlock(folio1); + if (folio1 != folio2) + folio_unlock(folio2); } /* @@ -200,77 +200,71 @@ static void vfs_unlock_two_pages(struct page *page1, struct page *page2) * Caller must have locked both inodes to prevent write races. */ static int vfs_dedupe_file_range_compare(struct inode *src, loff_t srcoff, - struct inode *dest, loff_t destoff, + struct inode *dest, loff_t dstoff, loff_t len, bool *is_same) { - loff_t src_poff; - loff_t dest_poff; - void *src_addr; - void *dest_addr; - struct page *src_page; - struct page *dest_page; - loff_t cmp_len; - bool same; - int error; - - error = -EINVAL; - same = true; + bool same = true; + int error = -EINVAL; + while (len) { - src_poff = srcoff & (PAGE_SIZE - 1); - dest_poff = destoff & (PAGE_SIZE - 1); - cmp_len = min(PAGE_SIZE - src_poff, - PAGE_SIZE - dest_poff); + struct folio *src_folio, *dst_folio; + void *src_addr, *dst_addr; + loff_t cmp_len = min(PAGE_SIZE - offset_in_page(srcoff), + PAGE_SIZE - offset_in_page(dstoff)); + cmp_len = min(cmp_len, len); if (cmp_len <= 0) goto out_error; - src_page = vfs_dedupe_get_page(src, srcoff); - if (IS_ERR(src_page)) { - error = PTR_ERR(src_page); + src_folio = vfs_dedupe_get_folio(src, srcoff); + if (IS_ERR(src_folio)) { + error = PTR_ERR(src_folio); goto out_error; } - dest_page = vfs_dedupe_get_page(dest, destoff); - if (IS_ERR(dest_page)) { - error = PTR_ERR(dest_page); - put_page(src_page); + dst_folio = vfs_dedupe_get_folio(dest, dstoff); + if (IS_ERR(dst_folio)) { + error = PTR_ERR(dst_folio); + folio_put(src_folio); goto out_error; } - vfs_lock_two_pages(src_page, dest_page); + vfs_lock_two_folios(src_folio, dst_folio); /* - * Now that we've locked both pages, make sure they're still + * Now that we've locked both folios, make sure they're still * mapped to the file data we're interested in. If not, * someone is invalidating pages on us and we lose. */ - if (!PageUptodate(src_page) || !PageUptodate(dest_page) || - src_page->mapping != src->i_mapping || - dest_page->mapping != dest->i_mapping) { + if (!folio_test_uptodate(src_folio) || !folio_test_uptodate(dst_folio) || + src_folio->mapping != src->i_mapping || + dst_folio->mapping != dest->i_mapping) { same = false; goto unlock; } - src_addr = kmap_atomic(src_page); - dest_addr = kmap_atomic(dest_page); + src_addr = kmap_local_folio(src_folio, + offset_in_folio(src_folio, srcoff)); + dst_addr = kmap_local_folio(dst_folio, + offset_in_folio(dst_folio, dstoff)); - flush_dcache_page(src_page); - flush_dcache_page(dest_page); + flush_dcache_folio(src_folio); + flush_dcache_folio(dst_folio); - if (memcmp(src_addr + src_poff, dest_addr + dest_poff, cmp_len)) + if (memcmp(src_addr, dst_addr, cmp_len)) same = false; - kunmap_atomic(dest_addr); - kunmap_atomic(src_addr); + kunmap_local(dst_addr); + kunmap_local(src_addr); unlock: - vfs_unlock_two_pages(src_page, dest_page); - put_page(dest_page); - put_page(src_page); + vfs_unlock_two_folios(src_folio, dst_folio); + folio_put(dst_folio); + folio_put(src_folio); if (!same) break; srcoff += cmp_len; - destoff += cmp_len; + dstoff += cmp_len; len -= cmp_len; }
We still only operate on a single page of data at a time due to using kmap(). A more complex implementation would work on each page in a folio, but it's not clear that such a complex implementation would be worthwhile. Signed-off-by: Matthew Wilcox (Oracle) <willy@infradead.org> --- fs/remap_range.c | 116 ++++++++++++++++++++++------------------------- 1 file changed, 55 insertions(+), 61 deletions(-)