[PATCH 0/8] mm: globalize rest_of_page() macro

[PATCH 0/8] mm: globalize rest_of_page() macro

[Yury Norov]

The net/9p networking driver has a handy macro to calculate the
amount of bytes from a given pointer to the end of page. Move it
to core/mm, and apply tree-wide. No functional changes intended.

This series was originally introduced as a single patch #07/12 in:

https://lore.kernel.org/all/20260219181407.290201-1-ynorov@nvidia.com/

Split it for better granularity and submit separately.

Yury Norov (8):
  mm: add rest_of_page() macro
  fs: use rest_of_page() macro where appropriate
  net: use rest_of_page() macro where appropriate
  core: use rest_of_page() macro where appropriate
  spi: use rest_of_page() macro where appropriate
  KVM: use rest_of_page() macro where appropriate
  drivers: ALSA: use rest_of_page() macro where appropriate
  arch: use rest_of_page() macro where appropriate

 arch/arm64/kernel/patching.c                |  4 +---
 arch/powerpc/lib/code-patching.c            |  6 +++---
 arch/riscv/kernel/sbi.c                     |  4 ++--
 arch/s390/kvm/gaccess.c                     |  6 +++---
 arch/x86/kvm/emulate.c                      |  4 ++--
 drivers/block/null_blk/main.c               |  6 ++----
 drivers/gpu/drm/i915/gt/shmem_utils.c       |  5 ++---
 drivers/md/dm-pcache/backing_dev.h          |  2 +-
 drivers/net/ethernet/meta/fbnic/fbnic_tlv.c |  6 +++---
 drivers/s390/net/qeth_core_main.c           |  6 ++----
 drivers/spi/spi-pl022.c                     |  3 +--
 drivers/spi/spi.c                           |  4 +---
 fs/ext4/verity.c                            |  3 +--
 fs/f2fs/verity.c                            |  6 ++----
 fs/fuse/dev.c                               |  4 ++--
 fs/iomap/buffered-io.c                      |  2 +-
 fs/nfs/pagelist.c                           |  2 +-
 fs/remap_range.c                            |  3 +--
 fs/xfs/scrub/xfile.c                        |  3 +--
 include/crypto/scatterwalk.h                |  2 +-
 include/linux/highmem.h                     | 24 +++++++++------------
 include/linux/iomap.h                       |  2 +-
 include/linux/iov_iter.h                    |  3 +--
 include/linux/mm.h                          |  2 ++
 kernel/events/ring_buffer.c                 |  2 +-
 lib/bitmap-str.c                            |  2 +-
 lib/iov_iter.c                              |  5 ++---
 net/9p/trans_virtio.c                       |  6 ------
 sound/virtio/virtio_pcm_msg.c               |  4 ++--
 29 files changed, 53 insertions(+), 78 deletions(-)

-- 
2.43.0

[PATCH 1/8] mm: add rest_of_page() macro

[Yury Norov]

The net/9p networking driver has a handy macro to calculate the
amount of bytes from a given pointer to the end of page. Move it
to mm. The following patches apply it tree-wide.

Signed-off-by: Yury Norov <ynorov@nvidia.com>
---
 include/linux/mm.h    | 2 ++
 net/9p/trans_virtio.c | 6 ------
 2 files changed, 2 insertions(+), 6 deletions(-)

diff --git a/include/linux/mm.h b/include/linux/mm.h
index 5be3d8a8f806..6d1025c6249a 100644
--- a/include/linux/mm.h
+++ b/include/linux/mm.h
@@ -2793,6 +2793,8 @@ extern void pagefault_out_of_memory(void);
 #define offset_in_page(p)	((unsigned long)(p) & ~PAGE_MASK)
 #define offset_in_folio(folio, p) ((unsigned long)(p) & (folio_size(folio) - 1))
 
+#define rest_of_page(p)		(PAGE_SIZE - offset_in_page(p))
+
 /*
  * Parameter block passed down to zap_pte_range in exceptional cases.
  */
diff --git a/net/9p/trans_virtio.c b/net/9p/trans_virtio.c
index 4cdab7094b27..1ca53209d036 100644
--- a/net/9p/trans_virtio.c
+++ b/net/9p/trans_virtio.c
@@ -87,12 +87,6 @@ struct virtio_chan {
 
 static struct list_head virtio_chan_list;
 
-/* How many bytes left in this page. */
-static unsigned int rest_of_page(void *data)
-{
-	return PAGE_SIZE - offset_in_page(data);
-}
-
 /**
  * p9_virtio_close - reclaim resources of a channel
  * @client: client instance
-- 
2.43.0

[PATCH 2/8] fs: use rest_of_page() macro where appropriate

[Yury Norov]

Switch filesystem codebase to using the macro. No functional changes
intended.

Signed-off-by: Yury Norov <ynorov@nvidia.com>
---
 fs/ext4/verity.c       | 3 +--
 fs/f2fs/verity.c       | 6 ++----
 fs/fuse/dev.c          | 4 ++--
 fs/iomap/buffered-io.c | 2 +-
 fs/nfs/pagelist.c      | 2 +-
 fs/remap_range.c       | 3 +--
 fs/xfs/scrub/xfile.c   | 3 +--
 7 files changed, 9 insertions(+), 14 deletions(-)

diff --git a/fs/ext4/verity.c b/fs/ext4/verity.c
index ca61da53f313..3dc95581e4b1 100644
--- a/fs/ext4/verity.c
+++ b/fs/ext4/verity.c
@@ -74,8 +74,7 @@ static int pagecache_write(struct inode *inode, const void *buf, size_t count,
 		return -EFBIG;
 
 	while (count) {
-		size_t n = min_t(size_t, count,
-				 PAGE_SIZE - offset_in_page(pos));
+		size_t n = min_t(size_t, count, rest_of_page(pos));
 		struct folio *folio;
 		void *fsdata = NULL;
 		int res;
diff --git a/fs/f2fs/verity.c b/fs/f2fs/verity.c
index 92ebcc19cab0..1c3403fbf2a8 100644
--- a/fs/f2fs/verity.c
+++ b/fs/f2fs/verity.c
@@ -44,8 +44,7 @@ static int pagecache_read(struct inode *inode, void *buf, size_t count,
 			  loff_t pos)
 {
 	while (count) {
-		size_t n = min_t(size_t, count,
-				 PAGE_SIZE - offset_in_page(pos));
+		size_t n = min_t(size_t, count, rest_of_page(pos));
 		struct page *page;
 
 		page = read_mapping_page(inode->i_mapping, pos >> PAGE_SHIFT,
@@ -78,8 +77,7 @@ static int pagecache_write(struct inode *inode, const void *buf, size_t count,
 		return -EFBIG;
 
 	while (count) {
-		size_t n = min_t(size_t, count,
-				 PAGE_SIZE - offset_in_page(pos));
+		size_t n = min_t(size_t, count, rest_of_page(pos));
 		struct folio *folio;
 		void *fsdata = NULL;
 		int res;
diff --git a/fs/fuse/dev.c b/fs/fuse/dev.c
index 0b0241f47170..efd7e6ca929e 100644
--- a/fs/fuse/dev.c
+++ b/fs/fuse/dev.c
@@ -1157,8 +1157,8 @@ static int fuse_copy_folio(struct fuse_copy_state *cs, struct folio **foliop,
 			unsigned int copy = count;
 			unsigned int bytes_copied;
 
-			if (folio_test_highmem(folio) && count > PAGE_SIZE - offset_in_page(offset))
-				copy = PAGE_SIZE - offset_in_page(offset);
+			if (folio_test_highmem(folio) && count > rest_of_page(offset))
+				copy = rest_of_page(offset);
 
 			bytes_copied = fuse_copy_do(cs, &buf, &copy);
 			kunmap_local(mapaddr);
diff --git a/fs/iomap/buffered-io.c b/fs/iomap/buffered-io.c
index bc82083e420a..99e56ee6c3d6 100644
--- a/fs/iomap/buffered-io.c
+++ b/fs/iomap/buffered-io.c
@@ -852,7 +852,7 @@ static struct folio *__iomap_get_folio(struct iomap_iter *iter,
 	loff_t pos = iter->pos;
 
 	if (!mapping_large_folio_support(iter->inode->i_mapping))
-		len = min_t(size_t, len, PAGE_SIZE - offset_in_page(pos));
+		len = min_t(size_t, len, rest_of_page(pos));
 
 	if (iter->iomap.flags & IOMAP_F_FOLIO_BATCH) {
 		struct folio *folio = folio_batch_next(iter->fbatch);
diff --git a/fs/nfs/pagelist.c b/fs/nfs/pagelist.c
index a9373de891c9..221a90f57812 100644
--- a/fs/nfs/pagelist.c
+++ b/fs/nfs/pagelist.c
@@ -60,7 +60,7 @@ static struct page *nfs_page_iter_page_get(struct nfs_page_iter_page *i)
 
 	if (i->count != req->wb_bytes) {
 		size_t base = i->count + req->wb_pgbase;
-		size_t len = PAGE_SIZE - offset_in_page(base);
+		size_t len = rest_of_page(base);
 
 		page = nfs_page_to_page(req, base);
 		nfs_page_iter_page_advance(i, len);
diff --git a/fs/remap_range.c b/fs/remap_range.c
index 26afbbbfb10c..83f325e7f96b 100644
--- a/fs/remap_range.c
+++ b/fs/remap_range.c
@@ -199,8 +199,7 @@ static int vfs_dedupe_file_range_compare(struct file *src, loff_t srcoff,
 	while (len) {
 		struct folio *src_folio, *dst_folio;
 		void *src_addr, *dst_addr;
-		loff_t cmp_len = min(PAGE_SIZE - offset_in_page(srcoff),
-				     PAGE_SIZE - offset_in_page(dstoff));
+		loff_t cmp_len = min(rest_of_page(srcoff), rest_of_page(dstoff));
 
 		cmp_len = min(cmp_len, len);
 		if (cmp_len <= 0)
diff --git a/fs/xfs/scrub/xfile.c b/fs/xfs/scrub/xfile.c
index 05581571854d..95707407aa6b 100644
--- a/fs/xfs/scrub/xfile.c
+++ b/fs/xfs/scrub/xfile.c
@@ -135,8 +135,7 @@ xfile_load(
 			 * No data stored at this offset, just zero the output
 			 * buffer until the next page boundary.
 			 */
-			len = min_t(ssize_t, count,
-				PAGE_SIZE - offset_in_page(pos));
+			len = min_t(ssize_t, count, rest_of_page(pos));
 			memset(buf, 0, len);
 		} else {
 			if (filemap_check_wb_err(inode->i_mapping, 0)) {
-- 
2.43.0

[PATCH 3/8] net: use rest_of_page() macro where appropriate

[Yury Norov]

Switch networking codebase to using the macro. No functional changes
intended.

Signed-off-by: Yury Norov <ynorov@nvidia.com>
---
 drivers/net/ethernet/meta/fbnic/fbnic_tlv.c | 6 +++---
 drivers/s390/net/qeth_core_main.c           | 6 ++----
 2 files changed, 5 insertions(+), 7 deletions(-)

diff --git a/drivers/net/ethernet/meta/fbnic/fbnic_tlv.c b/drivers/net/ethernet/meta/fbnic/fbnic_tlv.c
index 517ed8b2f1cb..2e80c25ba3c8 100644
--- a/drivers/net/ethernet/meta/fbnic/fbnic_tlv.c
+++ b/drivers/net/ethernet/meta/fbnic/fbnic_tlv.c
@@ -52,7 +52,7 @@ struct fbnic_tlv_msg *fbnic_tlv_msg_alloc(u16 msg_id)
  **/
 int fbnic_tlv_attr_put_flag(struct fbnic_tlv_msg *msg, const u16 attr_id)
 {
-	int attr_max_len = PAGE_SIZE - offset_in_page(msg) - sizeof(*msg);
+	int attr_max_len = rest_of_page(msg) - sizeof(*msg);
 	struct fbnic_tlv_hdr hdr = { 0 };
 	struct fbnic_tlv_msg *attr;
 
@@ -94,7 +94,7 @@ int fbnic_tlv_attr_put_flag(struct fbnic_tlv_msg *msg, const u16 attr_id)
 int fbnic_tlv_attr_put_value(struct fbnic_tlv_msg *msg, const u16 attr_id,
 			     const void *value, const int len)
 {
-	int attr_max_len = PAGE_SIZE - offset_in_page(msg) - sizeof(*msg);
+	int attr_max_len = rest_of_page(msg) - sizeof(*msg);
 	struct fbnic_tlv_hdr hdr = { 0 };
 	struct fbnic_tlv_msg *attr;
 
@@ -292,7 +292,7 @@ ssize_t fbnic_tlv_attr_get_string(struct fbnic_tlv_msg *attr, char *dst,
 struct fbnic_tlv_msg *fbnic_tlv_attr_nest_start(struct fbnic_tlv_msg *msg,
 						u16 attr_id)
 {
-	int attr_max_len = PAGE_SIZE - offset_in_page(msg) - sizeof(*msg);
+	int attr_max_len = rest_of_page(msg) - sizeof(*msg);
 	struct fbnic_tlv_msg *attr = &msg[le16_to_cpu(msg->hdr.len)];
 	struct fbnic_tlv_hdr hdr = { 0 };
 
diff --git a/drivers/s390/net/qeth_core_main.c b/drivers/s390/net/qeth_core_main.c
index cf5f760d0e02..5012c22d8f37 100644
--- a/drivers/s390/net/qeth_core_main.c
+++ b/drivers/s390/net/qeth_core_main.c
@@ -4087,8 +4087,7 @@ static unsigned int qeth_fill_buffer(struct qeth_qdio_out_buffer *buf,
 
 	/* map linear part into buffer element(s) */
 	while (length > 0) {
-		elem_length = min_t(unsigned int, length,
-				    PAGE_SIZE - offset_in_page(data));
+		elem_length = min_t(unsigned int, length, rest_of_page(data));
 
 		buffer->element[element].addr = virt_to_dma64(data);
 		buffer->element[element].length = elem_length;
@@ -4117,8 +4116,7 @@ static unsigned int qeth_fill_buffer(struct qeth_qdio_out_buffer *buf,
 		data = skb_frag_address(frag);
 		length = skb_frag_size(frag);
 		while (length > 0) {
-			elem_length = min_t(unsigned int, length,
-					    PAGE_SIZE - offset_in_page(data));
+			elem_length = min_t(unsigned int, length, rest_of_page(data));
 
 			buffer->element[element].addr = virt_to_dma64(data);
 			buffer->element[element].length = elem_length;
-- 
2.43.0

[PATCH 4/8] core: use rest_of_page() macro where appropriate

[Yury Norov]

Switch core and library code to using the macro. No functional
changes intended.

Signed-off-by: Yury Norov <ynorov@nvidia.com>
---
 include/crypto/scatterwalk.h |  2 +-
 include/linux/highmem.h      | 24 ++++++++++--------------
 include/linux/iomap.h        |  2 +-
 include/linux/iov_iter.h     |  3 +--
 kernel/events/ring_buffer.c  |  2 +-
 lib/bitmap-str.c             |  2 +-
 lib/iov_iter.c               |  5 ++---
 7 files changed, 17 insertions(+), 23 deletions(-)

diff --git a/include/crypto/scatterwalk.h b/include/crypto/scatterwalk.h
index 624fab589c2c..c671d5383c12 100644
--- a/include/crypto/scatterwalk.h
+++ b/include/crypto/scatterwalk.h
@@ -73,7 +73,7 @@ static inline unsigned int scatterwalk_clamp(struct scatter_walk *walk,
 	 * page due to the data not being aligned to the algorithm's alignmask.
 	 */
 	if (IS_ENABLED(CONFIG_HIGHMEM))
-		limit = PAGE_SIZE - offset_in_page(walk->offset);
+		limit = rest_of_page(walk->offset);
 	else
 		limit = PAGE_SIZE;
 
diff --git a/include/linux/highmem.h b/include/linux/highmem.h
index af03db851a1d..05528ba886fb 100644
--- a/include/linux/highmem.h
+++ b/include/linux/highmem.h
@@ -507,12 +507,10 @@ static inline void memcpy_folio(struct folio *dst_folio, size_t dst_off,
 		const char *src = kmap_local_folio(src_folio, src_off);
 		size_t chunk = len;
 
-		if (folio_test_highmem(dst_folio) &&
-		    chunk > PAGE_SIZE - offset_in_page(dst_off))
-			chunk = PAGE_SIZE - offset_in_page(dst_off);
-		if (folio_test_highmem(src_folio) &&
-		    chunk > PAGE_SIZE - offset_in_page(src_off))
-			chunk = PAGE_SIZE - offset_in_page(src_off);
+		if (folio_test_highmem(dst_folio) && chunk > rest_of_page(dst_off))
+			chunk = rest_of_page(dst_off);
+		if (folio_test_highmem(src_folio) && chunk > rest_of_page(src_off))
+			chunk = rest_of_page(src_off);
 		memcpy(dst, src, chunk);
 		kunmap_local(src);
 		kunmap_local(dst);
@@ -580,9 +578,8 @@ static inline void memcpy_from_folio(char *to, struct folio *folio,
 		const char *from = kmap_local_folio(folio, offset);
 		size_t chunk = len;
 
-		if (folio_test_partial_kmap(folio) &&
-		    chunk > PAGE_SIZE - offset_in_page(offset))
-			chunk = PAGE_SIZE - offset_in_page(offset);
+		if (folio_test_partial_kmap(folio) && chunk > rest_of_page(offset))
+			chunk = rest_of_page(offset);
 		memcpy(to, from, chunk);
 		kunmap_local(from);
 
@@ -608,9 +605,8 @@ static inline void memcpy_to_folio(struct folio *folio, size_t offset,
 		char *to = kmap_local_folio(folio, offset);
 		size_t chunk = len;
 
-		if (folio_test_partial_kmap(folio) &&
-		    chunk > PAGE_SIZE - offset_in_page(offset))
-			chunk = PAGE_SIZE - offset_in_page(offset);
+		if (folio_test_partial_kmap(folio) && chunk > rest_of_page(offset))
+			chunk = rest_of_page(offset);
 		memcpy(to, from, chunk);
 		kunmap_local(to);
 
@@ -642,7 +638,7 @@ static inline __must_check void *folio_zero_tail(struct folio *folio,
 	size_t len = folio_size(folio) - offset;
 
 	if (folio_test_partial_kmap(folio)) {
-		size_t max = PAGE_SIZE - offset_in_page(offset);
+		size_t max = rest_of_page(offset);
 
 		while (len > max) {
 			memset(kaddr, 0, max);
@@ -680,7 +676,7 @@ static inline void folio_fill_tail(struct folio *folio, size_t offset,
 	VM_BUG_ON(offset + len > folio_size(folio));
 
 	if (folio_test_partial_kmap(folio)) {
-		size_t max = PAGE_SIZE - offset_in_page(offset);
+		size_t max = rest_of_page(offset);
 
 		while (len > max) {
 			memcpy(to, from, max);
diff --git a/include/linux/iomap.h b/include/linux/iomap.h
index 99b7209dabd7..6ae549192adb 100644
--- a/include/linux/iomap.h
+++ b/include/linux/iomap.h
@@ -142,7 +142,7 @@ static inline void *iomap_inline_data(const struct iomap *iomap, loff_t pos)
  */
 static inline bool iomap_inline_data_valid(const struct iomap *iomap)
 {
-	return iomap->length <= PAGE_SIZE - offset_in_page(iomap->inline_data);
+	return iomap->length <= rest_of_page(iomap->inline_data);
 }
 
 /*
diff --git a/include/linux/iov_iter.h b/include/linux/iov_iter.h
index f9a17fbbd398..13a9ee653ef8 100644
--- a/include/linux/iov_iter.h
+++ b/include/linux/iov_iter.h
@@ -227,8 +227,7 @@ size_t iterate_xarray(struct iov_iter *iter, size_t len, void *priv, void *priv2
 		while (flen) {
 			void *base = kmap_local_folio(folio, offset);
 
-			part = min_t(size_t, flen,
-				     PAGE_SIZE - offset_in_page(offset));
+			part = min_t(size_t, flen, rest_of_page(offset));
 			remain = step(base, progress, part, priv, priv2);
 			kunmap_local(base);
 
diff --git a/kernel/events/ring_buffer.c b/kernel/events/ring_buffer.c
index 3e7de2661417..1db2868b90c9 100644
--- a/kernel/events/ring_buffer.c
+++ b/kernel/events/ring_buffer.c
@@ -590,7 +590,7 @@ long perf_output_copy_aux(struct perf_output_handle *aux_handle,
 	to &= (rb->aux_nr_pages << PAGE_SHIFT) - 1;
 
 	do {
-		tocopy = PAGE_SIZE - offset_in_page(from);
+		tocopy = rest_of_page(from);
 		if (to > from)
 			tocopy = min(tocopy, to - from);
 		if (!tocopy)
diff --git a/lib/bitmap-str.c b/lib/bitmap-str.c
index be745209507a..a357342d5d6c 100644
--- a/lib/bitmap-str.c
+++ b/lib/bitmap-str.c
@@ -58,7 +58,7 @@ EXPORT_SYMBOL(bitmap_parse_user);
 int bitmap_print_to_pagebuf(bool list, char *buf, const unsigned long *maskp,
 			    int nmaskbits)
 {
-	ptrdiff_t len = PAGE_SIZE - offset_in_page(buf);
+	ptrdiff_t len = rest_of_page(buf);
 
 	return list ? scnprintf(buf, len, "%*pbl\n", nmaskbits, maskp) :
 		      scnprintf(buf, len, "%*pb\n", nmaskbits, maskp);
diff --git a/lib/iov_iter.c b/lib/iov_iter.c
index 0a63c7fba313..c7e812349ca2 100644
--- a/lib/iov_iter.c
+++ b/lib/iov_iter.c
@@ -483,9 +483,8 @@ size_t copy_folio_from_iter_atomic(struct folio *folio, size_t offset,
 		char *to = kmap_local_folio(folio, offset);
 
 		n = bytes - copied;
-		if (folio_test_partial_kmap(folio) &&
-		    n > PAGE_SIZE - offset_in_page(offset))
-			n = PAGE_SIZE - offset_in_page(offset);
+		if (folio_test_partial_kmap(folio) && n > rest_of_page(offset))
+			n = rest_of_page(offset);
 
 		pagefault_disable();
 		n = __copy_from_iter(to, n, i);
-- 
2.43.0

[PATCH 5/8] spi: use rest_of_page() macro where appropriate

[Yury Norov]

Switch SPI code to using the macro. No functional changes intended.

Signed-off-by: Yury Norov <ynorov@nvidia.com>
---
 drivers/spi/spi-pl022.c | 3 +--
 drivers/spi/spi.c       | 4 +---
 2 files changed, 2 insertions(+), 5 deletions(-)

diff --git a/drivers/spi/spi-pl022.c b/drivers/spi/spi-pl022.c
index c82cc522776d..78fce33ff422 100644
--- a/drivers/spi/spi-pl022.c
+++ b/drivers/spi/spi-pl022.c
@@ -761,8 +761,7 @@ static void setup_dma_scatter(struct pl022 *pl022,
 			 * we just feed in this, else we stuff in as much
 			 * as we can.
 			 */
-			mapbytes = min_t(int, bytesleft,
-					 PAGE_SIZE - offset_in_page(bufp));
+			mapbytes = min_t(int, bytesleft, rest_of_page(bufp));
 
 			sg_set_page(sg, virt_to_page(bufp),
 				    mapbytes, offset_in_page(bufp));
diff --git a/drivers/spi/spi.c b/drivers/spi/spi.c
index 61f7bde8c7fb..cd4a18f3afaf 100644
--- a/drivers/spi/spi.c
+++ b/drivers/spi/spi.c
@@ -1162,9 +1162,7 @@ static int spi_map_buf_attrs(struct spi_controller *ctlr, struct device *dev,
 			 * the desc_len and the remaining buffer length that
 			 * fits in a page.
 			 */
-			min = min_t(size_t, desc_len,
-				    min_t(size_t, len,
-					  PAGE_SIZE - offset_in_page(buf)));
+			min = min_t(size_t, desc_len, min_t(size_t, len, rest_of_page(buf)));
 			if (vmalloced_buf)
 				vm_page = vmalloc_to_page(buf);
 			else
-- 
2.43.0

Re: [PATCH 5/8] spi: use rest_of_page() macro where appropriate

[Mark Brown]

[-- Attachment #1: Type: text/plain, Size: 181 bytes --]

On Tue, Mar 03, 2026 at 08:27:13PM -0500, Yury Norov wrote:
> Switch SPI code to using the macro. No functional changes intended.
> 

Acked-by: Mark Brown <broonie@kernel.org>

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 488 bytes --]

[PATCH 6/8] KVM: use rest_of_page() macro where appropriate

[Yury Norov]

Switch KVM code to using the macro. No functional changes intended.

Signed-off-by: Yury Norov <ynorov@nvidia.com>
---
 arch/s390/kvm/gaccess.c | 6 +++---
 arch/x86/kvm/emulate.c  | 4 ++--
 2 files changed, 5 insertions(+), 5 deletions(-)

diff --git a/arch/s390/kvm/gaccess.c b/arch/s390/kvm/gaccess.c
index 4630b2a067ea..40f85b7eca63 100644
--- a/arch/s390/kvm/gaccess.c
+++ b/arch/s390/kvm/gaccess.c
@@ -973,7 +973,7 @@ int access_guest_with_key(struct kvm_vcpu *vcpu, unsigned long ga, u8 ar,
 	if (rc)
 		goto out_unlock;
 	for (idx = 0; idx < nr_pages; idx++) {
-		fragment_len = min(PAGE_SIZE - offset_in_page(gpas[idx]), len);
+		fragment_len = min(rest_of_page(gpas[idx]), len);
 		if (try_fetch_prot_override && fetch_prot_override_applies(ga, fragment_len)) {
 			rc = access_guest_page_gpa(vcpu->kvm, mode, gpas[idx], data, fragment_len);
 		} else {
@@ -1015,7 +1015,7 @@ int access_guest_real(struct kvm_vcpu *vcpu, unsigned long gra,
 
 	while (len && !rc) {
 		gpa = kvm_s390_real_to_abs(vcpu, gra);
-		fragment_len = min(PAGE_SIZE - offset_in_page(gpa), len);
+		fragment_len = min(rest_of_page(gpa), len);
 		rc = access_guest_page_gpa(vcpu->kvm, mode, gpa, data, fragment_len);
 		len -= fragment_len;
 		gra += fragment_len;
@@ -1237,7 +1237,7 @@ int check_gpa_range(struct kvm *kvm, unsigned long gpa, unsigned long length,
 	int rc = 0;
 
 	while (length && !rc) {
-		fragment_len = min(PAGE_SIZE - offset_in_page(gpa), length);
+		fragment_len = min(rest_of_page(gpa), length);
 		rc = vm_check_access_key_gpa(kvm, access_key, mode, gpa);
 		length -= fragment_len;
 		gpa += fragment_len;
diff --git a/arch/x86/kvm/emulate.c b/arch/x86/kvm/emulate.c
index c8e292e9a24d..c060d1e2bb94 100644
--- a/arch/x86/kvm/emulate.c
+++ b/arch/x86/kvm/emulate.c
@@ -864,7 +864,7 @@ static int __do_insn_fetch_bytes(struct x86_emulate_ctxt *ctxt, int op_size)
 		return rc;
 
 	size = min_t(unsigned, 15UL ^ cur_size, max_size);
-	size = min_t(unsigned, size, PAGE_SIZE - offset_in_page(linear));
+	size = min_t(unsigned int, size, rest_of_page(linear));
 
 	/*
 	 * One instruction can only straddle two pages,
@@ -1372,7 +1372,7 @@ static int pio_in_emulated(struct x86_emulate_ctxt *ctxt,
 			address_mask(ctxt, reg_read(ctxt, VCPU_REGS_RCX)) : 1;
 		in_page = (ctxt->eflags & X86_EFLAGS_DF) ?
 			offset_in_page(reg_read(ctxt, VCPU_REGS_RDI)) :
-			PAGE_SIZE - offset_in_page(reg_read(ctxt, VCPU_REGS_RDI));
+			rest_of_page(reg_read(ctxt, VCPU_REGS_RDI));
 		n = min3(in_page, (unsigned int)sizeof(rc->data) / size, count);
 		if (n == 0)
 			n = 1;
-- 
2.43.0

[PATCH 7/8] drivers: ALSA: use rest_of_page() macro where appropriate

[Yury Norov]

Switch drivers to using the macro. No functional changes intended.

Signed-off-by: Yury Norov <ynorov@nvidia.com>
---
 drivers/block/null_blk/main.c         | 6 ++----
 drivers/gpu/drm/i915/gt/shmem_utils.c | 5 ++---
 drivers/md/dm-pcache/backing_dev.h    | 2 +-
 sound/virtio/virtio_pcm_msg.c         | 4 ++--
 4 files changed, 7 insertions(+), 10 deletions(-)

diff --git a/drivers/block/null_blk/main.c b/drivers/block/null_blk/main.c
index f8c0fd57e041..89e9651b09a9 100644
--- a/drivers/block/null_blk/main.c
+++ b/drivers/block/null_blk/main.c
@@ -1147,8 +1147,7 @@ static blk_status_t copy_to_nullb(struct nullb *nullb, void *source,
 	sector_t sector;
 
 	while (count < n) {
-		temp = min3(nullb->dev->blocksize, n - count,
-			    PAGE_SIZE - offset_in_page(pos));
+		temp = min3(nullb->dev->blocksize, n - count, rest_of_page(pos));
 		sector = pos >> SECTOR_SHIFT;
 
 		if (null_cache_active(nullb) && !is_fua)
@@ -1181,8 +1180,7 @@ static void copy_from_nullb(struct nullb *nullb, void *dest, loff_t pos,
 	sector_t sector;
 
 	while (count < n) {
-		temp = min3(nullb->dev->blocksize, n - count,
-			    PAGE_SIZE - offset_in_page(pos));
+		temp = min3(nullb->dev->blocksize, n - count, rest_of_page(pos));
 		sector = pos >> SECTOR_SHIFT;
 
 		t_page = null_lookup_page(nullb, sector, false,
diff --git a/drivers/gpu/drm/i915/gt/shmem_utils.c b/drivers/gpu/drm/i915/gt/shmem_utils.c
index 5850adaebf82..9a0a6f67fef0 100644
--- a/drivers/gpu/drm/i915/gt/shmem_utils.c
+++ b/drivers/gpu/drm/i915/gt/shmem_utils.c
@@ -99,8 +99,7 @@ static int __shmem_rw(struct file *file, loff_t off,
 	unsigned long pfn;
 
 	for (pfn = off >> PAGE_SHIFT; len; pfn++) {
-		unsigned int this =
-			min_t(size_t, PAGE_SIZE - offset_in_page(off), len);
+		unsigned int this = min_t(size_t, rest_of_page(off), len);
 		struct page *page;
 		void *vaddr;
 
@@ -135,7 +134,7 @@ int shmem_read_to_iosys_map(struct file *file, loff_t off,
 
 	for (pfn = off >> PAGE_SHIFT; len; pfn++) {
 		unsigned int this =
-			min_t(size_t, PAGE_SIZE - offset_in_page(off), len);
+			min_t(size_t, rest_of_page(off), len);
 		struct page *page;
 		void *vaddr;
 
diff --git a/drivers/md/dm-pcache/backing_dev.h b/drivers/md/dm-pcache/backing_dev.h
index b371cba483b9..17e83b38b845 100644
--- a/drivers/md/dm-pcache/backing_dev.h
+++ b/drivers/md/dm-pcache/backing_dev.h
@@ -96,7 +96,7 @@ static inline u32 backing_dev_req_coalesced_max_len(const void *data, u32 len)
 
 	first_page = vmalloc_to_page(p);
 advance:
-	in_page = PAGE_SIZE - offset_in_page(p);
+	in_page = rest_of_page(p);
 	to_advance = min_t(u32, in_page, len - done);
 
 	done += to_advance;
diff --git a/sound/virtio/virtio_pcm_msg.c b/sound/virtio/virtio_pcm_msg.c
index a5c4e7027717..5d1b0dc08234 100644
--- a/sound/virtio/virtio_pcm_msg.c
+++ b/sound/virtio/virtio_pcm_msg.c
@@ -56,7 +56,7 @@ static int virtsnd_pcm_sg_num(u8 *data, unsigned int length)
 		phys_addr_t pg_address = page_to_phys(pg);
 		size_t pg_length;
 
-		pg_length = PAGE_SIZE - offset_in_page(data);
+		pg_length = rest_of_page(data);
 		if (pg_length > length)
 			pg_length = length;
 
@@ -96,7 +96,7 @@ static void virtsnd_pcm_sg_from(struct scatterlist *sgs, int nsgs, u8 *data,
 		struct page *pg = vmalloc_to_page(data);
 		size_t pg_length;
 
-		pg_length = PAGE_SIZE - offset_in_page(data);
+		pg_length = rest_of_page(data);
 		if (pg_length > length)
 			pg_length = length;
 
-- 
2.43.0

[PATCH 8/8] arch: use rest_of_page() macro where appropriate

[Yury Norov]

Switch arch code to using the macro. No functional changes intended.

Signed-off-by: Yury Norov <ynorov@nvidia.com>
---
 arch/arm64/kernel/patching.c     | 4 +---
 arch/powerpc/lib/code-patching.c | 6 +++---
 arch/riscv/kernel/sbi.c          | 4 ++--
 3 files changed, 6 insertions(+), 8 deletions(-)

diff --git a/arch/arm64/kernel/patching.c b/arch/arm64/kernel/patching.c
index 1041bc67a3ee..4c3a4401719b 100644
--- a/arch/arm64/kernel/patching.c
+++ b/arch/arm64/kernel/patching.c
@@ -116,9 +116,7 @@ static void *__text_poke(text_poke_f func, void *addr, void *src, size_t len)
 
 	while (patched < len) {
 		ptr = addr + patched;
-		size = min_t(size_t, PAGE_SIZE - offset_in_page(ptr),
-			     len - patched);
-
+		size = min_t(size_t, rest_of_page(ptr), len - patched);
 		waddr = patch_map(ptr, FIX_TEXT_POKE0);
 		func(waddr, src, patched, size);
 		patch_unmap(FIX_TEXT_POKE0);
diff --git a/arch/powerpc/lib/code-patching.c b/arch/powerpc/lib/code-patching.c
index f84e0337cc02..186a9cb79ee3 100644
--- a/arch/powerpc/lib/code-patching.c
+++ b/arch/powerpc/lib/code-patching.c
@@ -463,7 +463,7 @@ static int __patch_instructions(u32 *patch_addr, u32 *code, size_t len, bool rep
 
 /*
  * A page is mapped and instructions that fit the page are patched.
- * Assumes 'len' to be (PAGE_SIZE - offset_in_page(addr)) or below.
+ * Assumes 'len' to be rest_of_page(addr) or below.
  */
 static int __do_patch_instructions_mm(u32 *addr, u32 *code, size_t len, bool repeat_instr)
 {
@@ -514,7 +514,7 @@ static int __do_patch_instructions_mm(u32 *addr, u32 *code, size_t len, bool rep
 
 /*
  * A page is mapped and instructions that fit the page are patched.
- * Assumes 'len' to be (PAGE_SIZE - offset_in_page(addr)) or below.
+ * Assumes 'len' to be rest_of_page(addr) or below.
  */
 static int __do_patch_instructions(u32 *addr, u32 *code, size_t len, bool repeat_instr)
 {
@@ -554,7 +554,7 @@ int patch_instructions(u32 *addr, u32 *code, size_t len, bool repeat_instr)
 		size_t plen;
 		int err;
 
-		plen = min_t(size_t, PAGE_SIZE - offset_in_page(addr), len);
+		plen = min_t(size_t, rest_of_page(addr), len);
 
 		local_irq_save(flags);
 		if (mm_patch_enabled())
diff --git a/arch/riscv/kernel/sbi.c b/arch/riscv/kernel/sbi.c
index c443337056ab..9a2f656f776f 100644
--- a/arch/riscv/kernel/sbi.c
+++ b/arch/riscv/kernel/sbi.c
@@ -602,7 +602,7 @@ int sbi_debug_console_write(const char *bytes, unsigned int num_bytes)
 	else
 		base_addr = __pa(bytes);
 	if (PAGE_SIZE < (offset_in_page(bytes) + num_bytes))
-		num_bytes = PAGE_SIZE - offset_in_page(bytes);
+		num_bytes = rest_of_page(bytes);
 
 	if (IS_ENABLED(CONFIG_32BIT))
 		ret = sbi_ecall(SBI_EXT_DBCN, SBI_EXT_DBCN_CONSOLE_WRITE,
@@ -631,7 +631,7 @@ int sbi_debug_console_read(char *bytes, unsigned int num_bytes)
 	else
 		base_addr = __pa(bytes);
 	if (PAGE_SIZE < (offset_in_page(bytes) + num_bytes))
-		num_bytes = PAGE_SIZE - offset_in_page(bytes);
+		num_bytes = rest_of_page(bytes);
 
 	if (IS_ENABLED(CONFIG_32BIT))
 		ret = sbi_ecall(SBI_EXT_DBCN, SBI_EXT_DBCN_CONSOLE_READ,
-- 
2.43.0

Re: [PATCH 0/8] mm: globalize rest_of_page() macro

[Jakub Kicinski]

On Tue,  3 Mar 2026 20:27:08 -0500 Yury Norov wrote:
> The net/9p networking driver has a handy macro to calculate the
> amount of bytes from a given pointer to the end of page. Move it
> to core/mm, and apply tree-wide. No functional changes intended.
> 
> This series was originally introduced as a single patch #07/12 in:
> 
> https://lore.kernel.org/all/20260219181407.290201-1-ynorov@nvidia.com/
> 
> Split it for better granularity and submit separately.

I don't get what the motivation is here. Another helper developers
and readers of the code will need to know about just to replace 
obvious and easy to comprehend math.

Re: [PATCH 0/8] mm: globalize rest_of_page() macro

[Jens Axboe]

On 3/3/26 7:28 PM, Jakub Kicinski wrote:
> On Tue,  3 Mar 2026 20:27:08 -0500 Yury Norov wrote:
>> The net/9p networking driver has a handy macro to calculate the
>> amount of bytes from a given pointer to the end of page. Move it
>> to core/mm, and apply tree-wide. No functional changes intended.
>>
>> This series was originally introduced as a single patch #07/12 in:
>>
>> https://lore.kernel.org/all/20260219181407.290201-1-ynorov@nvidia.com/
>>
>> Split it for better granularity and submit separately.
> 
> I don't get what the motivation is here. Another helper developers
> and readers of the code will need to know about just to replace 
> obvious and easy to comprehend math.

I fully agree, I had the same thought reading this.

-- 
Jens Axboe

Re: [PATCH 0/8] mm: globalize rest_of_page() macro

[Sean Christopherson]

On Tue, Mar 03, 2026, Jens Axboe wrote:
> On 3/3/26 7:28 PM, Jakub Kicinski wrote:
> > On Tue,  3 Mar 2026 20:27:08 -0500 Yury Norov wrote:
> >> The net/9p networking driver has a handy macro to calculate the
> >> amount of bytes from a given pointer to the end of page. Move it
> >> to core/mm, and apply tree-wide. No functional changes intended.
> >>
> >> This series was originally introduced as a single patch #07/12 in:
> >>
> >> https://lore.kernel.org/all/20260219181407.290201-1-ynorov@nvidia.com/
> >>
> >> Split it for better granularity and submit separately.
> > 
> > I don't get what the motivation is here. Another helper developers
> > and readers of the code will need to know about just to replace 
> > obvious and easy to comprehend math.
> 
> I fully agree, I had the same thought reading this.

+1 from KVM-land.

Re: [PATCH 0/8] mm: globalize rest_of_page() macro

[Yury Norov]

On Tue, Mar 03, 2026 at 06:49:36PM -0800, Sean Christopherson wrote:
> On Tue, Mar 03, 2026, Jens Axboe wrote:
> > On 3/3/26 7:28 PM, Jakub Kicinski wrote:
> > > On Tue,  3 Mar 2026 20:27:08 -0500 Yury Norov wrote:
> > >> The net/9p networking driver has a handy macro to calculate the
> > >> amount of bytes from a given pointer to the end of page. Move it
> > >> to core/mm, and apply tree-wide. No functional changes intended.
> > >>
> > >> This series was originally introduced as a single patch #07/12 in:
> > >>
> > >> https://lore.kernel.org/all/20260219181407.290201-1-ynorov@nvidia.com/
> > >>
> > >> Split it for better granularity and submit separately.
> > > 
> > > I don't get what the motivation is here. Another helper developers
> > > and readers of the code will need to know about just to replace 
> > > obvious and easy to comprehend math.
> > 
> > I fully agree, I had the same thought reading this.
> 
> +1 from KVM-land.

My motivation is that it helps to simplify constructions like this:

-               loff_t cmp_len = min(PAGE_SIZE - offset_in_page(srcoff),
-                                    PAGE_SIZE - offset_in_page(dstoff));
+               loff_t cmp_len = min(rest_of_page(srcoff), rest_of_page(dstoff));

Or this:

-               if (folio_test_highmem(dst_folio) &&
-                   chunk > PAGE_SIZE - offset_in_page(dst_off))
-                       chunk = PAGE_SIZE - offset_in_page(dst_off);
-               if (folio_test_highmem(src_folio) &&
-                   chunk > PAGE_SIZE - offset_in_page(src_off))
-                       chunk = PAGE_SIZE - offset_in_page(src_off);
+               if (folio_test_highmem(dst_folio) && chunk > rest_of_page(dst_off))
+                       chunk = rest_of_page(dst_off);
+               if (folio_test_highmem(src_folio) && chunk > rest_of_page(src_off))
+                       chunk = rest_of_page(src_off);

To a point where I don't have to use my brains to decode them. I agree
it's an easy math. It's just too bulky to my (and 9p guys too) taste.

Thanks,
Yury

Re: [PATCH 0/8] mm: globalize rest_of_page() macro

[Jens Axboe]

On 3/3/26 8:32 PM, Yury Norov wrote:
> My motivation is that it helps to simplify constructions like this:
> 
> -               loff_t cmp_len = min(PAGE_SIZE - offset_in_page(srcoff),
> -                                    PAGE_SIZE - offset_in_page(dstoff));
> +               loff_t cmp_len = min(rest_of_page(srcoff), rest_of_page(dstoff));
> 
> Or this:
> 
> -               if (folio_test_highmem(dst_folio) &&
> -                   chunk > PAGE_SIZE - offset_in_page(dst_off))
> -                       chunk = PAGE_SIZE - offset_in_page(dst_off);
> -               if (folio_test_highmem(src_folio) &&
> -                   chunk > PAGE_SIZE - offset_in_page(src_off))
> -                       chunk = PAGE_SIZE - offset_in_page(src_off);
> +               if (folio_test_highmem(dst_folio) && chunk > rest_of_page(dst_off))
> +                       chunk = rest_of_page(dst_off);
> +               if (folio_test_highmem(src_folio) && chunk > rest_of_page(src_off))
> +                       chunk = rest_of_page(src_off);
> 
> To a point where I don't have to use my brains to decode them. I agree
> it's an easy math. It's just too bulky to my (and 9p guys too) taste.

The thing is, now I have to go lookup what on earth rest_of_page() does,
whereas PAGE_SIZE - offset_in_page(page) is immediately obvious. It's a
classic case of "oh let's add this helper to simplify things" which
really just makes it worse, because now you have to jump to the
definition of rest_of_page().

IOW, just no.

-- 
Jens Axboe

Re: [PATCH 0/8] mm: globalize rest_of_page() macro

[Peter Zijlstra]

On Tue, Mar 03, 2026 at 06:49:36PM -0800, Sean Christopherson wrote:
> On Tue, Mar 03, 2026, Jens Axboe wrote:
> > On 3/3/26 7:28 PM, Jakub Kicinski wrote:
> > > On Tue,  3 Mar 2026 20:27:08 -0500 Yury Norov wrote:
> > >> The net/9p networking driver has a handy macro to calculate the
> > >> amount of bytes from a given pointer to the end of page. Move it
> > >> to core/mm, and apply tree-wide. No functional changes intended.
> > >>
> > >> This series was originally introduced as a single patch #07/12 in:
> > >>
> > >> https://lore.kernel.org/all/20260219181407.290201-1-ynorov@nvidia.com/
> > >>
> > >> Split it for better granularity and submit separately.
> > > 
> > > I don't get what the motivation is here. Another helper developers
> > > and readers of the code will need to know about just to replace 
> > > obvious and easy to comprehend math.
> > 
> > I fully agree, I had the same thought reading this.
> 
> +1 from KVM-land.

Right, this. I hate these pointless helpers that obscure perfectly
sensible and obvious code -- and for me that includes things like
rounddown() and DIV_ROUND_UP().

It just makes the code harder to read.