On Thu, May 20, 2021 at 04:42:37PM -0700, Darrick J. Wong wrote: > > @@ -292,8 +292,8 @@ _xfs_buf_get_pages( > > if (bp->b_pages == NULL) > > return -ENOMEM; > > } > > - memset(bp->b_pages, 0, sizeof(struct page *) * page_count); > > } > > + memset(bp->b_pages, 0, sizeof(struct page *) * bp->b_page_count); > > Could this kmem_alloc be converted to kmem_zalloc? Yes. > And isn't the xfs_buf allocated with zalloc, which means we don't need > to zero b_page_array itself? Yes. > Confused about why this is needed. My series cleans up all this mess and the rebases the allocation change from Dave on top. Maybe that is a better start :)
