The quilt patch titled Subject: mm: khugepaged: fix the arguments order in khugepaged_collapse_file trace point has been removed from the -mm tree. Its filename was mm-khugepaged-fix-the-arguments-order-in-khugepaged_collapse_file-trace-point.patch
This patch was dropped because it was merged into the mm-hotfixes-stable branch of git://git.kernel.org/pub/scm/linux/kernel/git/akpm/mm
------------------------------------------------------ From: Yang Shi yang@os.amperecomputing.com Subject: mm: khugepaged: fix the arguments order in khugepaged_collapse_file trace point Date: Fri, 11 Oct 2024 18:17:02 -0700
The "addr" and "is_shmem" arguments have different order in TP_PROTO and TP_ARGS. This resulted in the incorrect trace result:
text-hugepage-644429 [276] 392092.878683: mm_khugepaged_collapse_file: mm=0xffff20025d52c440, hpage_pfn=0x200678c00, index=512, addr=1, is_shmem=0, filename=text-hugepage, nr=512, result=failed
The value of "addr" is wrong because it was treated as bool value, the type of is_shmem.
Fix the order in TP_PROTO to keep "addr" is before "is_shmem" since the original patch review suggested this order to achieve best packing.
And use "lx" for "addr" instead of "ld" in TP_printk because address is typically shown in hex.
After the fix, the trace result looks correct:
text-hugepage-7291 [004] 128.627251: mm_khugepaged_collapse_file: mm=0xffff0001328f9500, hpage_pfn=0x20016ea00, index=512, addr=0x400000, is_shmem=0, filename=text-hugepage, nr=512, result=failed
Link: https://lkml.kernel.org/r/20241012011702.1084846-1-yang@os.amperecomputing.c... Fixes: 4c9473e87e75 ("mm/khugepaged: add tracepoint to collapse_file()") Signed-off-by: Yang Shi yang@os.amperecomputing.com Cc: Gautam Menghani gautammenghani201@gmail.com Cc: Steven Rostedt (Google) rostedt@goodmis.org Cc: stable@vger.kernel.org [6.2+] Signed-off-by: Andrew Morton akpm@linux-foundation.org ---
include/trace/events/huge_memory.h | 4 ++-- mm/khugepaged.c | 2 +- 2 files changed, 3 insertions(+), 3 deletions(-)
--- a/include/trace/events/huge_memory.h~mm-khugepaged-fix-the-arguments-order-in-khugepaged_collapse_file-trace-point +++ a/include/trace/events/huge_memory.h @@ -208,7 +208,7 @@ TRACE_EVENT(mm_khugepaged_scan_file,
TRACE_EVENT(mm_khugepaged_collapse_file, TP_PROTO(struct mm_struct *mm, struct folio *new_folio, pgoff_t index, - bool is_shmem, unsigned long addr, struct file *file, + unsigned long addr, bool is_shmem, struct file *file, int nr, int result), TP_ARGS(mm, new_folio, index, addr, is_shmem, file, nr, result), TP_STRUCT__entry( @@ -233,7 +233,7 @@ TRACE_EVENT(mm_khugepaged_collapse_file, __entry->result = result; ),
- TP_printk("mm=%p, hpage_pfn=0x%lx, index=%ld, addr=%ld, is_shmem=%d, filename=%s, nr=%d, result=%s", + TP_printk("mm=%p, hpage_pfn=0x%lx, index=%ld, addr=%lx, is_shmem=%d, filename=%s, nr=%d, result=%s", __entry->mm, __entry->hpfn, __entry->index, --- a/mm/khugepaged.c~mm-khugepaged-fix-the-arguments-order-in-khugepaged_collapse_file-trace-point +++ a/mm/khugepaged.c @@ -2227,7 +2227,7 @@ rollback: folio_put(new_folio); out: VM_BUG_ON(!list_empty(&pagelist)); - trace_mm_khugepaged_collapse_file(mm, new_folio, index, is_shmem, addr, file, HPAGE_PMD_NR, result); + trace_mm_khugepaged_collapse_file(mm, new_folio, index, addr, is_shmem, file, HPAGE_PMD_NR, result); return result; }
_
Patches currently in -mm which might be from yang@os.amperecomputing.com are
linux-stable-mirror@lists.linaro.org