Hi Kevin,

On 2025/1/3 21:27, Kevin Brodsky wrote:

On 03/01/2025 10:35, Qi Zheng wrote:

quoted

On 2025/1/3 17:13, Qi Zheng wrote:

quoted

On 2025/1/3 16:02, Kevin Brodsky wrote:

quoted

On 03/01/2025 04:48, Qi Zheng wrote:

quoted

[...]

In __tlb_batch_free_encoded_pages(), we can indeed detect PageTable()
and call pagetable_dtor() to dtor the page table pages.
But __tlb_batch_free_encoded_pages() is also used to free normal pages
(not page table pages), so I don't want to add overhead there.

Interesting, can a tlb batch refer to pages than are not PTPs then?

Yes, you can see the caller of __tlb_remove_folio_pages() or
tlb_remove_page_size().

I had a brief look but clearly not a good enough one! I hadn't realised
that "table" in tlb_remove_table() means PTP, while "page" in
tlb_remove_page() can mean any page, and it's making more sense now.

[...]

quoted

quoted

For arm, the call to pagetable_dtor() is indeed missed in the
non-MMU_GATHER_RCU_TABLE_FREE case. This needs to be fixed. But we
can't fix this by adding pagetable_dtor() to tlb_remove_table(),
because some architectures call tlb_remove_table() but don't support
page table statistics, like sparc.

When I investigated this for my own series, I found that the only case
where ctor/dtor are not called for page-sized page tables is 32-bit
sparc (see table at the end of [1]). However only 64-bit sparc makes use
of tlb_remove_table() (at PTE level, where ctor/dtor are already called).

Thanks for providing this information.

So really calling pagetable_dtor() from tlb_remove_table() in the
non-MMU_GATHER_TABLE_FREE case seems to be the obvious thing to do.

Right. Currently, only powerpc, sparc and x86 will directly call
tlb_remove_table(), and all of them are in the MMU_GATHER_TABLE_FREE
case. Therefore, I think the modification you mentioned below is
feasible.

In summary, currently only arm calls tlb_remove_table() in the
non-MMU_GATHER_RCU_TABLE_FREE case. So I think we can add this fix
directly to patch #8. If I haven't missed anything, I'll send an
updated patch #8.

Once this is done, we should be able to replace all those confusing
calls to tlb_remove_page() on PTPs with tlb_remove_table() and remove
the explicit call to pagetable_dtor(). AIUI this is essentially what
Peter suggested on v3 [2].

Since this patch series is mainly for bug fix, I think that these things
can be done in separate patch series later.

[1]
https://lore.kernel.org/linux-mm/20241219164425.2277022-1-kevin.brodsky@arm.com/ (local)
[2]
https://lore.kernel.org/linux-mm/20250103111457.GC22934@noisy.programming.kicks-ass.net/ (local)

[...]

quoted

Or can we just not let tlb_remove_table() fall back to
tlb_remove_page()? Like the following:

diff --git a/include/asm-generic/tlb.h b/include/asm-generic/tlb.h
index a59205863f431..354ffaa4bd120 100644
--- a/include/asm-generic/tlb.h
+++ b/include/asm-generic/tlb.h

@@ -195,8 +195,6 @@
   *  various ptep_get_and_clear() functions.
   */

-#ifdef CONFIG_MMU_GATHER_TABLE_FREE
-
  struct mmu_table_batch {
  #ifdef CONFIG_MMU_GATHER_RCU_TABLE_FREE
         struct rcu_head         rcu;

@@ -219,16 +217,6 @@ static inline void __tlb_remove_table(void *table)

  extern void tlb_remove_table(struct mmu_gather *tlb, void *table);

-#else /* !CONFIG_MMU_GATHER_HAVE_TABLE_FREE */
-
-/*
- * Without MMU_GATHER_TABLE_FREE the architecture is assumed to have

page based
- * page directories and we can use the normal page batching to free
them.
- */
-#define tlb_remove_table(tlb, page) tlb_remove_page((tlb), (page))

We still need a different implementation of tlb_remove_table() in this
case. We could define it inline here:

static inline void tlb_remove_table(struct mmu_gather *tlb, void *table)
{
     struct page *page = table;

     pagetable_dtor(page_ptdesc(page));
     tlb_remove_page(page);
}

Right. As I said above, will add this to the updated patch #8.

Thanks!

- Kevin