Re: [PATCH 2/4] mm: khugepaged: remove mm when all memory has been collapsed
From: David Hildenbrand (Red Hat)
Date: Fri Dec 19 2025 - 04:00:58 EST
On 12/19/25 06:24, Vernon Yang wrote:
On Thu, Dec 18, 2025 at 10:29:18AM +0100, David Hildenbrand (Red Hat) wrote:BTW, the 10s sleep is ridiculous :)
On 12/15/25 10:04, Vernon Yang wrote:
The following data is traced by bpftrace on a desktop system. After
the system has been left idle for 10 minutes upon booting, a lot of
SCAN_PMD_MAPPED or SCAN_PMD_NONE are observed during a full scan by
khugepaged.
@scan_pmd_status[1]: 1 ## SCAN_SUCCEED
@scan_pmd_status[4]: 158 ## SCAN_PMD_MAPPED
@scan_pmd_status[3]: 174 ## SCAN_PMD_NONE
total progress size: 701 MB
Total time : 440 seconds ## include khugepaged_scan_sleep_millisecs
The khugepaged_scan list save all task that support collapse into hugepage,
as long as the take is not destroyed, khugepaged will not remove it from
the khugepaged_scan list. This exist a phenomenon where task has already
collapsed all memory regions into hugepage, but khugepaged continues to
scan it, which wastes CPU time and invalid, and due to
khugepaged_scan_sleep_millisecs (default 10s) causes a long wait for
scanning a large number of invalid task, so scanning really valid task
is later.
After applying this patch, when all memory is either SCAN_PMD_MAPPED or
SCAN_PMD_NONE, the mm is automatically removed from khugepaged's scan
list. If the page fault or MADV_HUGEPAGE again, it is added back to
khugepaged.
I don't like that, as it assumes that memory within such a process would be
rather static, which is easily not the case (e.g., allocators just doing
MADV_DONTNEED to free memory).
If most stuff is collapsed to PMDs already, can't we just skip over these
regions a bit faster?
/* default scan 8*HPAGE_PMD_NR ptes (or vmas) every 10 second */
static unsigned int khugepaged_pages_to_scan __read_mostly;
The observed phenomenon is that when scanning these regions, the loop is
broken upon reaching the number of khugepaged_pages_to_scan, thereforce
the khugepaged enters 10s sleep.
I wonder whether we were more careful in the past regarding canning overhead due to the mmap read lock. Nowadays page faults typicaly use per-vma locks, so I wonder whether the scanning overhead is still a problem. (I assume there is more to optimize long-term)
--
Cheers
David