| Message ID | 20231023092319.1507325-1-mpe@ellerman.id.au (mailing list archive) |
|---|---|
| State | Accepted |
| Commit | aad26d3b6af13c055b1d05dd253d2484551bde55 |
| Headers | show |
| Series | powerpc/32s: Implement local_flush_tlb_page_psize() | expand |
| Context | Check | Description |
|---|---|---|
| snowpatch_ozlabs/github-powerpc_ppctests | success | Successfully ran 8 jobs. |
| snowpatch_ozlabs/github-powerpc_selftests | success | Successfully ran 8 jobs. |
| snowpatch_ozlabs/github-powerpc_sparse | success | Successfully ran 4 jobs. |
| snowpatch_ozlabs/github-powerpc_clang | success | Successfully ran 6 jobs. |
| snowpatch_ozlabs/github-powerpc_kernel_qemu | success | Successfully ran 23 jobs. |
On Mon, 23 Oct 2023 20:23:19 +1100, Michael Ellerman wrote: > There's a single call to local_flush_tlb_page_psize() in the code > patching code. That call is never executed on 32-bit Book3S, > because it's guarded by mm_patch_enabled() which is essentially a > radix_enabled() check, which is always false on 32s. > > However depending on how the optimiser sees things it may still trip > over the BUILD_BUG() in the 32s stub of local_flush_tlb_page_psize(). > > [...] Applied to powerpc/next. [1/1] powerpc/32s: Implement local_flush_tlb_page_psize() https://git.kernel.org/powerpc/c/aad26d3b6af13c055b1d05dd253d2484551bde55 cheers
diff --git a/arch/powerpc/include/asm/book3s/32/tlbflush.h b/arch/powerpc/include/asm/book3s/32/tlbflush.h index 4be572908124..e43534da5207 100644 --- a/arch/powerpc/include/asm/book3s/32/tlbflush.h +++ b/arch/powerpc/include/asm/book3s/32/tlbflush.h @@ -80,7 +80,7 @@ static inline void local_flush_tlb_page(struct vm_area_struct *vma, static inline void local_flush_tlb_page_psize(struct mm_struct *mm, unsigned long vmaddr, int psize) { - BUILD_BUG(); + flush_range(mm, vmaddr, vmaddr); } static inline void local_flush_tlb_mm(struct mm_struct *mm)
There's a single call to local_flush_tlb_page_psize() in the code patching code. That call is never executed on 32-bit Book3S, because it's guarded by mm_patch_enabled() which is essentially a radix_enabled() check, which is always false on 32s. However depending on how the optimiser sees things it may still trip over the BUILD_BUG() in the 32s stub of local_flush_tlb_page_psize(). To avoid that, implement it in terms of flush_range() so that if it ever becomes called it should function, even if not optimally. Note that flush_range() deals with page aligning the address and so on, and that 32s doesn't support huge pages so there should be no issue with non-standard page sizes needing to be flushed. Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> --- arch/powerpc/include/asm/book3s/32/tlbflush.h | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)