| Message ID | 20250112152658.20132-1-me@davidreaver.com |
|---|---|
| State | Superseded |
| Headers | show |
| Series | [v2] pm: Replace deprecated kmap_atomic() with kmap_local_page() | expand |
On Sun, Jan 12, 2025 at 4:27 PM David Reaver <me@davidreaver.com> wrote: > > kmap_atomic() is deprecated and should be replaced with kmap_local_page() > [1][2]. kmap_local_page() is faster in kernels with HIGHMEM enabled, can > take page faults, and allows preemption. > > According to [2], this replacement is safe as long as the code between > kmap_atomic() and kunmap_atomic() does not implicitly depend on disabling > page faults or preemption. In all of the call sites in this patch, the only > thing happening between mapping and unmapping pages is copy_page() calls, > and I don't suspect they depend on disabling page faults or preemption. > > [1] https://lwn.net/Articles/836144/ > [2] https://docs.kernel.org/mm/highmem.html#temporary-virtual-mappings > > Signed-off-by: David Reaver <me@davidreaver.com> > --- > > Apologies, I tried testing a CONFIG_HIGHMEM kernel in QEMU and clearly made > an error because there is an obvious typo. Any pointers to test setups for > testing CONFIG_HIGHMEM in a VM would be appreciated! Here is v2 of the > patch. > > V1 -> V2: Fix typo kmap_page_local() -> kmap_local_page() > > kernel/power/snapshot.c | 16 ++++++++-------- > 1 file changed, 8 insertions(+), 8 deletions(-) > > diff --git a/kernel/power/snapshot.c b/kernel/power/snapshot.c > index c9fb559a6399..87f4dde4a49d 100644 > --- a/kernel/power/snapshot.c > +++ b/kernel/power/snapshot.c > @@ -2270,9 +2270,9 @@ int snapshot_read_next(struct snapshot_handle *handle) > */ > void *kaddr; > > - kaddr = kmap_atomic(page); > + kaddr = kmap_local_page(page); > copy_page(buffer, kaddr); > - kunmap_atomic(kaddr); > + kunmap_local(kaddr); > handle->buffer = buffer; > } else { > handle->buffer = page_address(page); > @@ -2561,9 +2561,9 @@ static void copy_last_highmem_page(void) > if (last_highmem_page) { > void *dst; > > - dst = kmap_atomic(last_highmem_page); > + dst = kmap_local_page(last_highmem_page); > copy_page(dst, buffer); > - kunmap_atomic(dst); > + kunmap_local(dst); > last_highmem_page = NULL; > } > } > @@ -2881,13 +2881,13 @@ static inline void swap_two_pages_data(struct page *p1, struct page *p2, > { > void *kaddr1, *kaddr2; > > - kaddr1 = kmap_atomic(p1); > - kaddr2 = kmap_atomic(p2); > + kaddr1 = kmap_local_page(p1); > + kaddr2 = kmap_local_page(p2); > copy_page(buf, kaddr1); > copy_page(kaddr1, kaddr2); > copy_page(kaddr2, buf); > - kunmap_atomic(kaddr2); > - kunmap_atomic(kaddr1); > + kunmap_local(kaddr2); > + kunmap_local(kaddr1); > } > > /** Applied as 6.15 material, thanks!
diff --git a/kernel/power/snapshot.c b/kernel/power/snapshot.c index c9fb559a6399..87f4dde4a49d 100644 --- a/kernel/power/snapshot.c +++ b/kernel/power/snapshot.c @@ -2270,9 +2270,9 @@ int snapshot_read_next(struct snapshot_handle *handle) */ void *kaddr; - kaddr = kmap_atomic(page); + kaddr = kmap_local_page(page); copy_page(buffer, kaddr); - kunmap_atomic(kaddr); + kunmap_local(kaddr); handle->buffer = buffer; } else { handle->buffer = page_address(page); @@ -2561,9 +2561,9 @@ static void copy_last_highmem_page(void) if (last_highmem_page) { void *dst; - dst = kmap_atomic(last_highmem_page); + dst = kmap_local_page(last_highmem_page); copy_page(dst, buffer); - kunmap_atomic(dst); + kunmap_local(dst); last_highmem_page = NULL; } } @@ -2881,13 +2881,13 @@ static inline void swap_two_pages_data(struct page *p1, struct page *p2, { void *kaddr1, *kaddr2; - kaddr1 = kmap_atomic(p1); - kaddr2 = kmap_atomic(p2); + kaddr1 = kmap_local_page(p1); + kaddr2 = kmap_local_page(p2); copy_page(buf, kaddr1); copy_page(kaddr1, kaddr2); copy_page(kaddr2, buf); - kunmap_atomic(kaddr2); - kunmap_atomic(kaddr1); + kunmap_local(kaddr2); + kunmap_local(kaddr1); } /**
kmap_atomic() is deprecated and should be replaced with kmap_local_page() [1][2]. kmap_local_page() is faster in kernels with HIGHMEM enabled, can take page faults, and allows preemption. According to [2], this replacement is safe as long as the code between kmap_atomic() and kunmap_atomic() does not implicitly depend on disabling page faults or preemption. In all of the call sites in this patch, the only thing happening between mapping and unmapping pages is copy_page() calls, and I don't suspect they depend on disabling page faults or preemption. [1] https://lwn.net/Articles/836144/ [2] https://docs.kernel.org/mm/highmem.html#temporary-virtual-mappings Signed-off-by: David Reaver <me@davidreaver.com> --- Apologies, I tried testing a CONFIG_HIGHMEM kernel in QEMU and clearly made an error because there is an obvious typo. Any pointers to test setups for testing CONFIG_HIGHMEM in a VM would be appreciated! Here is v2 of the patch. V1 -> V2: Fix typo kmap_page_local() -> kmap_local_page() kernel/power/snapshot.c | 16 ++++++++-------- 1 file changed, 8 insertions(+), 8 deletions(-)