Re: [RFC 2/8] shmem: convert to use folio_test_hwpoison()

[Luis Chamberlain <mcgrof@xxxxxxxxxx>]

On Tue, Apr 25, 2023 at 01:00:25PM +0200, Pankaj Raghav wrote:
> On Fri, Apr 21, 2023 at 11:42:53PM +0100, Matthew Wilcox wrote:
> > On Fri, Apr 21, 2023 at 02:43:54PM -0700, Luis Chamberlain wrote:
> > > The PageHWPoison() call can be converted over to the respective folio call
> > > folio_test_hwpoison(). This introduces no functional changes.
> > 
> > Um, no.  Nobody should use folio_test_hwpoison(), it's a nonsense.
> > 
> > Individual pages are hwpoisoned.  You're only testing the head page
> > if you use folio_test_hwpoison().  There's folio_has_hwpoisoned() to
> > test if _any_ page in the folio is poisoned.  But blindly converting
> > PageHWPoison to folio_test_hwpoison() is wrong.
> 
> I see a pattern in shmem.c where first the head is tested and for large
> folios, any of pages in the folio is tested for poison flag. Should we
> factor it out as a helper in shmem.c and use it here?
> 
> static ssize_t shmem_file_splice_read(struct file *in, loff_t *ppos,
> ...
> 	if (folio_test_hwpoison(folio) ||
> 	    (folio_test_large(folio) &&
> 	     folio_test_has_hwpoisoned(folio))) {
> 	..

Hugh's commit 72887c976a7c9e ("shmem: minor fixes to splice-read
implementation") is on point about this :

  "Perhaps that ugliness can be improved at the mm end later"

So how about we put some lipstick on this guy now (notice right above it
a similar compound page check for is_page_hwpoison()):

diff --git a/include/linux/page-flags.h b/include/linux/page-flags.h
index 1c68d67b832f..6a4a571dbe50 100644
--- a/include/linux/page-flags.h
+++ b/include/linux/page-flags.h
@@ -883,6 +883,13 @@ static inline bool is_page_hwpoison(struct page *page)
 	return PageHuge(page) && PageHWPoison(compound_head(page));
 }
 
+static inline bool is_folio_hwpoison(struct folio *folio)
+{
+	if (folio_test_hwpoison(folio))
+		return true;
+	return folio_test_large(folio) && folio_test_has_hwpoisoned(folio);
+}
+
 /*
  * For pages that are never mapped to userspace (and aren't PageSlab),
  * page_type may be used.  Because it is initialised to -1, we invert the
diff --git a/mm/shmem.c b/mm/shmem.c
index ef7ad684f4fb..b7f47f6b75d5 100644
--- a/mm/shmem.c
+++ b/mm/shmem.c
@@ -3013,9 +3013,7 @@ static ssize_t shmem_file_splice_read(struct file *in, loff_t *ppos,
 		if (folio) {
 			folio_unlock(folio);
 
-			if (folio_test_hwpoison(folio) ||
-			    (folio_test_large(folio) &&
-			     folio_test_has_hwpoisoned(folio))) {
+			if (is_folio_hwpoison(folio)) {
 				error = -EIO;
 				break;
 			}