Everyone-
Does this look right? Non-IO error behavior during writeback (e.g.
ENOMEM) remains undefined.
thanks-
sage
---
Document correct use of page and address_space error bits for readpage,
writepage, and writepages on IO errors.
There are a few cases where ->writepage might get away with not marking
the mapping (the __writepage() wrapper in page-writeback.c will do it),
but only from certain call sites.
Signed-off-by: Sage Weil <sage@xxxxxxxxxxxx>
---
Documentation/filesystems/vfs.txt | 13 +++++++++++++
1 files changed, 13 insertions(+), 0 deletions(-)
diff --git a/Documentation/filesystems/vfs.txt b/Documentation/filesystems/vfs.txt
index f49eecf..63b8cd9 100644
--- a/Documentation/filesystems/vfs.txt
+++ b/Documentation/filesystems/vfs.txt
@@ -554,6 +554,11 @@ struct address_space_operations {
should return AOP_WRITEPAGE_ACTIVATE so that the VM will not keep
calling ->writepage on that page.
+ If an error occurs (EIO or ENOSPC), ->writepage should mark the
+ page via SetPageError, mark the address_space via
+ set_mapping_error, and (when possible) return an error to the
+ caller.
+
See the file "Locking" for more details.
readpage: called by the VM to read a page from backing store.
@@ -564,6 +569,10 @@ struct address_space_operations {
In this case, the page will be relocated, relocked and if
that all succeeds, ->readpage will be called again.
+ If an error occurs, ->readpage should return an error code if
+ possible. If it is an IO error, ->readpage should also mark
+ the page via SetPageError.
+
sync_page: called by the VM to notify the backing store to perform all
queued I/O operations for a page. I/O operations for other pages
associated with this address_space object may also be performed.
@@ -580,6 +589,10 @@ struct address_space_operations {
instead. This will choose pages from the address space that are
tagged as DIRTY and will pass them to ->writepage.
+ If an error occurs (EIO or ENOSPC), ->writepages should mark
+ the mapping via mapping_set_error(), and (when possible)
+ return an error to the caller.
+
set_page_dirty: called by the VM to set a page dirty.
This is particularly needed if an address space attaches
private data to a page, and that data needs to be updated when
--
1.5.6.5
--
To unsubscribe from this list: send the line "unsubscribe linux-fsdevel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
