Messages in this thread Patch in this message | | | From | Andi Kleen <> | Subject | [PATCH] [2/6] kfifo: Make kfifo_in atomic | Date | Sun, 27 Dec 2009 22:03:12 +0100 (CET) |
| |
Right now kfifo_in allows copying in less than the input amount. This is unfortunately not a good idea on any record oriented users: if the size of the kfifo is not a multiple of the record (and that can easily happen due to the power-of-two requirement) then when the FIFO fills up partial records could be put in. Such a condition would be fatal for any record consumer who would get permanently desynchronized. In fact I doubt unless the input is a totally boundary less data stream I doubt anything could handle this.
Change kfifo_in() to always put in everything or nothing.
The return value is now always 0 or the full length.
Signed-off-by: Andi Kleen <ak@linux.intel.com>
--- kernel/kfifo.c | 9 ++++----- 1 file changed, 4 insertions(+), 5 deletions(-)
Index: linux/kernel/kfifo.c =================================================================== --- linux.orig/kernel/kfifo.c +++ linux/kernel/kfifo.c @@ -228,9 +228,8 @@ EXPORT_SYMBOL(__kfifo_in_n); * @from: the data to be added. * @len: the length of the data to be added. * - * This function copies at most @len bytes from the @from buffer into - * the FIFO depending on the free space, and returns the number of - * bytes copied. + * This function copies @len bytes from the @from buffer into + * the FIFO and returns 0 if there is not enough space. * * Note that with only one concurrent reader and one concurrent * writer, you don't need extra locking to use these functions. @@ -238,8 +237,8 @@ EXPORT_SYMBOL(__kfifo_in_n); unsigned int kfifo_in(struct kfifo *fifo, const void *from, unsigned int len) { - len = min(kfifo_avail(fifo), len); - + if (kfifo_avail(fifo) < len) + return 0; __kfifo_in_data(fifo, from, len, 0); __kfifo_add_in(fifo, len); return len;
| |