[PATCH] axis-fifo: use size_t for write size calculations

[PATCH] axis-fifo: use size_t for write size calculations

[Lucas Jeffrey]

The write length is provided as size_t, but words_to_write and the
return value storage used narrower types. Use size_t for the word count
and ssize_t for the return value to keep the types consistent with the
write interface.

This avoids relying on implicit size limits when handling write sizes.

Signed-off-by: Lucas Jeffrey <luquijeffrey@gmail.com>
---
 drivers/staging/axis-fifo/axis-fifo.c | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)

diff --git a/drivers/staging/axis-fifo/axis-fifo.c b/drivers/staging/axis-fifo/axis-fifo.c
index 3aa2aa870ea9..d2b30b7b6845 100644
--- a/drivers/staging/axis-fifo/axis-fifo.c
+++ b/drivers/staging/axis-fifo/axis-fifo.c
@@ -214,9 +214,9 @@ static ssize_t axis_fifo_write(struct file *f, const char __user *buf,
 			       size_t len, loff_t *off)
 {
 	struct axis_fifo *fifo = f->private_data;
-	unsigned int words_to_write;
+	size_t words_to_write;
 	u32 *txbuf;
-	int ret;
+	ssize_t ret;
 
 	words_to_write = len / sizeof(u32);
 
@@ -257,7 +257,7 @@ static ssize_t axis_fifo_write(struct file *f, const char __user *buf,
 		goto end_unlock;
 	}
 
-	for (int i = 0; i < words_to_write; ++i)
+	for (size_t i = 0; i < words_to_write; ++i)
 		iowrite32(txbuf[i], fifo->base_addr + XLLF_TDFD_OFFSET);
 
 	iowrite32(len, fifo->base_addr + XLLF_TLR_OFFSET);
-- 
2.43.0

Re: [PATCH] axis-fifo: use size_t for write size calculations

[Dan Carpenter]

On Fri, Jul 31, 2026 at 12:43:07AM -0300, Lucas Jeffrey wrote:
> The write length is provided as size_t, but words_to_write and the
> return value storage used narrower types. Use size_t for the word count
> and ssize_t for the return value to keep the types consistent with the
> write interface.
> 
> This avoids relying on implicit size limits when handling write sizes.
> 
> Signed-off-by: Lucas Jeffrey <luquijeffrey@gmail.com>
> ---

This doesn't affect real life because len is capped in rw_verify_area().

The commit message needs to make it clear that this is just a clean up
and not a fix.  If we pretended that the len value in read/write
functions was an range from 0-ULONG_MAX then we would have a ton of
integer overflow bugs etc.

regards,
dan carpenter