| Message ID | 20210203084137.25522-1-linmiaohe@huawei.com (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
| Series | mm/pgtable-generic.c: optimize the VM_BUG_ON condition in pmdp_huge_clear_flush() | expand |
On Wed, Feb 03, 2021 at 03:41:37AM -0500, Miaohe Lin wrote: > The developer will have trouble figuring out why the BUG actually triggered > when there is a complex expression in the VM_BUG_ON. Because we can only > identify the condition triggered BUG via line number provided by VM_BUG_ON. > Optimize this by spliting such a complex expression into two simple > conditions. > pmd_t pmd; > VM_BUG_ON(address & ~HPAGE_PMD_MASK); > - VM_BUG_ON(!pmd_present(*pmdp) || (!pmd_trans_huge(*pmdp) && > - !pmd_devmap(*pmdp))); > + VM_BUG_ON(!pmd_present(*pmdp)); > + /* Below assumes pmd_present() is true */ > + VM_BUG_ON(!pmd_trans_huge(*pmdp) && !pmd_devmap(*pmdp)); This is not a complex condition. We're in the huge PMD handling case and we're looking at a PMD which either isn't present or isn't huge. It might be useful to print out the PMD in such a case, but splitting this into the two cases of pmd-not-present and pmd-isn't-huge isn't particularly useful. I think you know that, or you wouldn't feel the need to put in a comment explaining it!
Hi: On 2021/2/4 20:36, Matthew Wilcox wrote: > On Wed, Feb 03, 2021 at 03:41:37AM -0500, Miaohe Lin wrote: >> The developer will have trouble figuring out why the BUG actually triggered >> when there is a complex expression in the VM_BUG_ON. Because we can only >> identify the condition triggered BUG via line number provided by VM_BUG_ON. >> Optimize this by spliting such a complex expression into two simple >> conditions. > >> pmd_t pmd; >> VM_BUG_ON(address & ~HPAGE_PMD_MASK); >> - VM_BUG_ON(!pmd_present(*pmdp) || (!pmd_trans_huge(*pmdp) && >> - !pmd_devmap(*pmdp))); >> + VM_BUG_ON(!pmd_present(*pmdp)); >> + /* Below assumes pmd_present() is true */ >> + VM_BUG_ON(!pmd_trans_huge(*pmdp) && !pmd_devmap(*pmdp)); > > This is not a complex condition. We're in the huge PMD handling case > and we're looking at a PMD which either isn't present or isn't huge. > It might be useful to print out the PMD in such a case, but splitting > this into the two cases of pmd-not-present and pmd-isn't-huge isn't > particularly useful. > Many thanks for your time. You are right that it would be more helpful if we had a VM_BUG_ON_PMD() that we could print the pmd's value and permit diagnosis from that. I think splitting this into the two cases is the best we can do now while lacking of such helper. > I think you know that, or you wouldn't feel the need to put in a > comment explaining it! > . > Thanks again.
diff --git a/mm/pgtable-generic.c b/mm/pgtable-generic.c index fa1375f3e3b2..c2210e1cdb51 100644 --- a/mm/pgtable-generic.c +++ b/mm/pgtable-generic.c @@ -135,8 +135,9 @@ pmd_t pmdp_huge_clear_flush(struct vm_area_struct *vma, unsigned long address, { pmd_t pmd; VM_BUG_ON(address & ~HPAGE_PMD_MASK); - VM_BUG_ON(!pmd_present(*pmdp) || (!pmd_trans_huge(*pmdp) && - !pmd_devmap(*pmdp))); + VM_BUG_ON(!pmd_present(*pmdp)); + /* Below assumes pmd_present() is true */ + VM_BUG_ON(!pmd_trans_huge(*pmdp) && !pmd_devmap(*pmdp)); pmd = pmdp_huge_get_and_clear(vma->vm_mm, address, pmdp); flush_pmd_tlb_range(vma, address, address + HPAGE_PMD_SIZE); return pmd;
The developer will have trouble figuring out why the BUG actually triggered when there is a complex expression in the VM_BUG_ON. Because we can only identify the condition triggered BUG via line number provided by VM_BUG_ON. Optimize this by spliting such a complex expression into two simple conditions. Suggested-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Miaohe Lin <linmiaohe@huawei.com> --- mm/pgtable-generic.c | 5 +++-- 1 file changed, 3 insertions(+), 2 deletions(-)