All of lore.kernel.org
 help / color / mirror / Atom feed
From: Christoph Hellwig <hch@lst.de>
To: Chandan Babu R <chandan.babu@oracle.com>,
	"Darrick J. Wong" <djwong@kernel.org>,
	Hugh Dickins <hughd@google.com>,
	Andrew Morton <akpm@linux-foundation.org>
Cc: linux-xfs@vger.kernel.org, linux-mm@kvack.org
Subject: [PATCH 14/21] xfs: use shmem_get_folio in xfile_obj_store
Date: Fri, 26 Jan 2024 14:28:56 +0100	[thread overview]
Message-ID: <20240126132903.2700077-15-hch@lst.de> (raw)
In-Reply-To: <20240126132903.2700077-1-hch@lst.de>

Switch to using shmem_get_folio and manually dirtying the page instead
of abusing aops->write_begin and aops->write_end in xfile_get_page.

This simplifies the code by not doing indirect calls of not actually
exported interfaces that don't really fit the use case very well, and
happens to get us large folio support for free.

Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Darrick J. Wong <djwong@kernel.org>
---
 fs/xfs/scrub/xfile.c | 75 +++++++++++++++++---------------------------
 1 file changed, 29 insertions(+), 46 deletions(-)

diff --git a/fs/xfs/scrub/xfile.c b/fs/xfs/scrub/xfile.c
index 623bbde91ae3fe..c71c853c9ffdd7 100644
--- a/fs/xfs/scrub/xfile.c
+++ b/fs/xfs/scrub/xfile.c
@@ -183,11 +183,7 @@ xfile_store(
 	loff_t			pos)
 {
 	struct inode		*inode = file_inode(xf->file);
-	struct address_space	*mapping = inode->i_mapping;
-	const struct address_space_operations *aops = mapping->a_ops;
-	struct page		*page = NULL;
 	unsigned int		pflags;
-	int			error = 0;
 
 	if (count > MAX_RW_COUNT)
 		return -ENOMEM;
@@ -196,60 +192,47 @@ xfile_store(
 
 	trace_xfile_store(xf, pos, count);
 
+	/*
+	 * Increase the file size first so that shmem_get_folio(..., SGP_CACHE),
+	 * actually allocates a folio instead of erroring out.
+	 */
+	if (pos + count > i_size_read(inode))
+		i_size_write(inode, pos + count);
+
 	pflags = memalloc_nofs_save();
 	while (count > 0) {
-		void		*fsdata = NULL;
-		void		*p, *kaddr;
+		struct folio	*folio;
 		unsigned int	len;
-		int		ret;
+		unsigned int	offset;
 
-		len = min_t(ssize_t, count, PAGE_SIZE - offset_in_page(pos));
-
-		/*
-		 * We call write_begin directly here to avoid all the freezer
-		 * protection lock-taking that happens in the normal path.
-		 * shmem doesn't support fs freeze, but lockdep doesn't know
-		 * that and will trip over that.
-		 */
-		error = aops->write_begin(NULL, mapping, pos, len, &page,
-				&fsdata);
-		if (error) {
-			error = -ENOMEM;
+		if (shmem_get_folio(inode, pos >> PAGE_SHIFT, &folio,
+				SGP_CACHE) < 0)
 			break;
-		}
-
-		/*
-		 * xfile pages must never be mapped into userspace, so we skip
-		 * the dcache flush.  If the page is not uptodate, zero it
-		 * before writing data.
-		 */
-		kaddr = page_address(page);
-		if (!PageUptodate(page)) {
-			memset(kaddr, 0, PAGE_SIZE);
-			SetPageUptodate(page);
-		}
-		p = kaddr + offset_in_page(pos);
-		memcpy(p, buf, len);
-
-		ret = aops->write_end(NULL, mapping, pos, len, len, page,
-				fsdata);
-		if (ret < 0) {
-			error = -ENOMEM;
+		if (folio_test_hwpoison(folio) ||
+		    (folio_test_large(folio) &&
+		     folio_test_has_hwpoisoned(folio))) {
+			folio_unlock(folio);
+			folio_put(folio);
 			break;
 		}
 
-		if (ret != len) {
-			error = -ENOMEM;
-			break;
-		}
+		offset = offset_in_folio(folio, pos);
+		len = min_t(ssize_t, count, folio_size(folio) - offset);
+		memcpy(folio_address(folio) + offset, buf, len);
+
+		folio_mark_dirty(folio);
+		folio_unlock(folio);
+		folio_put(folio);
 
-		count -= ret;
-		pos += ret;
-		buf += ret;
+		count -= len;
+		pos += len;
+		buf += len;
 	}
 	memalloc_nofs_restore(pflags);
 
-	return error;
+	if (count)
+		return -ENOMEM;
+	return 0;
 }
 
 /* Find the next written area in the xfile data for a given offset. */
-- 
2.39.2


  parent reply	other threads:[~2024-01-26 13:30 UTC|newest]

Thread overview: 48+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2024-01-26 13:28 put the xfs xfile abstraction on a diet v2 Christoph Hellwig
2024-01-26 13:28 ` [PATCH 01/21] mm: move mapping_set_update out of <linux/swap.h> Christoph Hellwig
2024-01-26 14:39   ` Matthew Wilcox
2024-01-26 13:28 ` [PATCH 02/21] shmem: move shmem_mapping out of line Christoph Hellwig
2024-01-26 14:40   ` Matthew Wilcox
2024-01-26 13:28 ` [PATCH 03/21] shmem: set a_ops earlier in shmem_symlink Christoph Hellwig
2024-01-26 14:41   ` Matthew Wilcox
2024-01-26 13:28 ` [PATCH 04/21] shmem: move the shmem_mapping assert into shmem_get_folio_gfp Christoph Hellwig
2024-01-26 15:09   ` Matthew Wilcox
2024-01-26 13:28 ` [PATCH 05/21] shmem: export shmem_get_folio Christoph Hellwig
2024-01-26 15:15   ` Matthew Wilcox
2024-01-26 13:28 ` [PATCH 06/21] shmem: export shmem_kernel_file_setup Christoph Hellwig
2024-01-26 15:45   ` Matthew Wilcox
2024-01-26 13:28 ` [PATCH 07/21] shmem: document how to "persist" data when using shmem_*file_setup Christoph Hellwig
2024-01-26 15:49   ` Matthew Wilcox
2024-01-28 16:54     ` Christoph Hellwig
2024-01-29 15:56       ` Matthew Wilcox
2024-01-29 16:04         ` Christoph Hellwig
2024-01-26 13:28 ` [PATCH 08/21] xfs: remove xfile_stat Christoph Hellwig
2024-01-26 13:28 ` [PATCH 09/21] xfs: remove the xfile_pread/pwrite APIs Christoph Hellwig
2024-01-26 16:21   ` Matthew Wilcox
2024-01-26 16:48     ` Darrick J. Wong
2024-01-26 13:28 ` [PATCH 10/21] xfs: don't try to handle non-update pages in xfile_obj_load Christoph Hellwig
2024-01-26 13:28 ` [PATCH 11/21] xfs: shmem_file_setup can't return NULL Christoph Hellwig
2024-01-26 13:28 ` [PATCH 12/21] xfs: don't modify file and inode flags for shmem files Christoph Hellwig
2024-01-26 13:28 ` [PATCH 13/21] xfs: don't allow highmem pages in xfile mappings Christoph Hellwig
2024-01-26 16:26   ` Matthew Wilcox
2024-01-26 13:28 ` Christoph Hellwig [this message]
2024-01-26 16:29   ` [PATCH 14/21] xfs: use shmem_get_folio in xfile_obj_store Matthew Wilcox
2024-01-26 13:28 ` [PATCH 15/21] xfs: use shmem_get_folio in in xfile_load Christoph Hellwig
2024-01-26 13:28 ` [PATCH 16/21] xfs: improve detection of lost xfile contents Christoph Hellwig
2024-01-26 16:33   ` Matthew Wilcox
2024-01-26 16:50     ` Darrick J. Wong
2024-01-28 16:55     ` Christoph Hellwig
2024-01-28 20:33       ` Matthew Wilcox
2024-01-26 13:28 ` [PATCH 17/21] xfs: add file_{get,put}_folio Christoph Hellwig
2024-01-26 16:40   ` Matthew Wilcox
2024-01-26 16:55     ` Darrick J. Wong
2024-01-27  1:26   ` Kent Overstreet
2024-01-27  1:32     ` Darrick J. Wong
2024-01-26 13:29 ` [PATCH 18/21] xfs: remove xfarray_sortinfo.page_kaddr Christoph Hellwig
2024-01-26 13:29 ` [PATCH 19/21] xfs: fix a comment in xfarray.c Christoph Hellwig
2024-01-26 13:29 ` [PATCH 20/21] xfs: convert xfarray_pagesort to deal with large folios Christoph Hellwig
2024-01-27  1:09   ` Kent Overstreet
2024-01-26 13:29 ` [PATCH 21/21] xfs: remove xfile_{get,put}_page Christoph Hellwig
2024-01-26 14:15 ` put the xfs xfile abstraction on a diet v2 Matthew Wilcox
2024-01-26 14:18   ` Christoph Hellwig
2024-01-26 16:47 ` Matthew Wilcox

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20240126132903.2700077-15-hch@lst.de \
    --to=hch@lst.de \
    --cc=akpm@linux-foundation.org \
    --cc=chandan.babu@oracle.com \
    --cc=djwong@kernel.org \
    --cc=hughd@google.com \
    --cc=linux-mm@kvack.org \
    --cc=linux-xfs@vger.kernel.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.