Re: [PATCH v2 24/40] mm/mlock: eliminate weird VMA_IO_BIT abuse and simplify
From: sashiko-bot@kernel.org
Date: 2026-09-14 21:44:37
Also in:
bpf, dri-devel, kvm, kvmarm, linux-perf-users, linux-rdma, linux-s390, linux-scsi, selinux
Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider: - [Low] Casting `bool` to `void *` causes a compiler warning/error (-Wint-to-pointer-cast). -- commit 5ca9abe89befb83b06ebb688a6a1ab1fe25a4e0e Author: Lorenzo Stoakes (ARM) [off-list ref] mm/mlock: eliminate weird VMA_IO_BIT abuse and simplify This commit eliminates the use of VMA_IO_BIT during mlock/munlock operations which was previously used to prevent double-counting during migration. Instead, it threads a boolean through the page table walk and temporarily sets VMA_LOCKONFAULT_BIT without VMA_LOCKED_BIT to indicate a locking walk in progress. It also simplifies checking for VMA_LOCKED_BIT across various folio locking paths.
quoted hunk ↗ jump to hunk
diff --git a/mm/mlock.c b/mm/mlock.c index 39215a3eab1fb..4235a1518fc9e 100644 --- a/mm/mlock.c +++ b/mm/mlock.c
[ ... ]
quoted hunk ↗ jump to hunk
@@ -428,31 +417,29 @@ static void mlock_vma_pages_range(struct vm_area_struct *vma,
[ ... ]
lru_add_drain(); - walk_page_range_vma(vma, start, end, &mlock_walk_ops, NULL); + walk_page_range_vma(vma, start, end, &mlock_walk_ops, (void *)lock);
[Severity: Low] Does this direct cast from a boolean to a void pointer cause a compiler warning? Since a boolean is typically 1 byte and a void pointer is 8 bytes on 64-bit systems, casting directly might trigger a -Wint-to-pointer-cast warning. When building with -Werror enabled, this can result in a build failure. Would it be safer to cast it through unsigned long first, like (void *)(unsigned long)lock?
lru_add_drain();
-- Sashiko AI review · https://sashiko.dev/#/patchset/20260914-b4-mmap-prepare-vma-flag-sanify-v2-0-7d9781ed5361@kernel.org?part=24