diff options
| author | Vlastimil Babka (SUSE) <vbabka@kernel.org> | 2026-11-03 17:16:51 +0100 |
|---|---|---|
| committer | Vlastimil Babka (SUSE) <vbabka@kernel.org> | 2026-11-10 09:19:09 +0100 |
| commit | 2122887b04f3f2d76daa80055c2c273926032fa0 (patch) | |
| tree | 193530cb438716b685c4d41a195d67316cc8150e /include/linux/platform_data | |
| parent | 58e81bb7de5aac8ec5e4aa21d401e7dafbd4058b (diff) | |
mm/page_alloc: remove the zone_spans_pfn() check from get_pfnblock_bitmap_bitidx()
Whenever the flags of a pageblock are read or written, a kernel with
CONFIG_DEBUG_VM makes sure that the pfn lies within the zone of the page.
Unlike bad_range(), that check reads zone->zone_start_pfn and
zone->spanned_pages without zone->span_seqlock.
Memory hotplug changes the two fields one after the other. If the lower
end of a zone moves, the zone seems to end before its last pages in
between. A page of that part which is freed at that moment, on another
CPU, ends in
page dumped because: VM_BUG_ON_PAGE(!zone_spans_pfn(page_zone(page), pfn))
kernel BUG at mm/page_alloc.c:372!
free_unref_folios+0x55e/0xed0
folios_put_refs+0x2b5/0x530
An attempt has been made [1] to fix this with the span seqlock check.
But that can trigger all sorts of problems in restricted contexts, as
noted by sashiko [2].
Since this is just a sanity debug check and anything with BUG_ON is
frowned upon anyway, just remove it altogether.
Reported-by: Josef Bacik <josef@toxicpanda.com>
Closes: https://lore.kernel.org/all/20261027-pgalloc-ai-hotfixes-v1-9-9c7d1aca801e@kernel.org/ [1]
Link: https://lore.kernel.org/all/sashiko-outbox-188081@kernel.org/ [2]
Fixes: 86051ca5eaf5 ("mm: fix usemap initialization")
Cc: stable@vger.kernel.org
Link: https://patch.msgid.link/20261103-pgalloc-ai-hotfixes-v2-5-e0d2890ab74f@kernel.org
Signed-off-by: Vlastimil Babka (SUSE) <vbabka@kernel.org>
Diffstat (limited to 'include/linux/platform_data')
0 files changed, 0 insertions, 0 deletions
