Re: Endless calls to xas_split_alloc() due to corrupted xarray entry

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



On 6/20/24 5:58 AM, David Hildenbrand wrote:
On 19.06.24 17:48, Linus Torvalds wrote:
On Wed, 19 Jun 2024 at 07:31, Matthew Wilcox <willy@xxxxxxxxxxxxx> wrote:

Actually, it's 11.  We can't split an order-12 folio because we'd have
to allocate two levels of radix tree, and I decided that was too much
work.  Also, I didn't know that ARM used order-13 PMD size at the time.

I think this is the best fix (modulo s/12/11/).

Can we use some more descriptive thing than the magic constant 11 that
is clearly very subtle.

Is it "XA_CHUNK_SHIFT * 2 - 1"

That's my best guess as well :)


IOW, something like

    #define MAX_XAS_ORDER (XA_CHUNK_SHIFT * 2 - 1)
    #define MAX_PAGECACHE_ORDER min(HPAGE_PMD_ORDER,12)

except for the non-TRANSPARENT_HUGEPAGE case where it currently does

   #define MAX_PAGECACHE_ORDER    8

and I assume that "8" is just "random round value, smaller than 11"?

Yes, that matches my understanding.

Maybe to be safe for !THP as well, something ike:

+++ b/include/linux/pagemap.h
@@ -354,11 +354,18 @@ static inline void mapping_set_gfp_mask(struct address_space *m, gfp_t mask)
   * a good order (that's 1MB if you're using 4kB pages)
   */
  #ifdef CONFIG_TRANSPARENT_HUGEPAGE
-#define MAX_PAGECACHE_ORDER    HPAGE_PMD_ORDER
+#define WANTED_MAX_PAGECACHE_ORDER    HPAGE_PMD_ORDER
  #else
-#define MAX_PAGECACHE_ORDER    8
+#define WANTED_MAX_PAGECACHE_ORDER    8
  #endif

+/*
+ * xas_split_alloc() does not support arbitrary orders yet. This implies no
+ * 512MB THP on arm64 with 64k.
+ */
+#define MAX_XAS_ORDER        (XA_CHUNK_SHIFT * 2 - 1)
+#define MAX_PAGECACHE_ORDER    min(MAX_XAS_ORDER, WANTED_MAX_PAGECACHE_ORDER)
+
  /**
   * mapping_set_large_folios() - Indicate the file supports large folios.
   * @mapping: The file.

Thanks David. I'm checking if shmem needs the similar limitation and test patches.
I will post them for review once they're ready.

Thanks,
Gavin





[Index of Archives]     [XFS Filesystem Development (older mail)]     [Linux Filesystem Development]     [Linux Audio Users]     [Yosemite Trails]     [Linux Kernel]     [Linux RAID]     [Linux SCSI]


  Powered by Linux