Any reason why we keep these two variables around for kiocb structure?
They are not used anywhere.
Patch to tidy up kiocb structure.
Signed-off-by: Ken Chen <[email protected]>
--- ./include/linux/aio.h.orig 2006-09-21 10:03:36.000000000 -0700
+++ ./include/linux/aio.h 2006-09-21 10:03:56.000000000 -0700
@@ -110,8 +110,6 @@
char __user *ki_buf; /* remaining iocb->aio_buf */
size_t ki_left; /* remaining bytes */
long ki_retried; /* just for testing */
- long ki_kicked; /* just for testing */
- long ki_queued; /* just for testing */
struct list_head ki_list; /* the aio core uses this
* for cancellation */
Chen, Kenneth W wrote:
> Any reason why we keep these two variables around for kiocb structure?
If there is a good one I've forgotten it.
> They are not used anywhere.
Indeed.
The ki_retried users all seem pretty questionable, too. How about
removing all that stuff?
> Signed-off-by: Ken Chen <[email protected]>
Signed-off-by: Zach Brown <[email protected]>
- z
Zach Brown wrote on Thursday, September 21, 2006 12:34 PM
> Chen, Kenneth W wrote:
> > Any reason why we keep these two variables around for kiocb structure?
>
> If there is a good one I've forgotten it.
>
> > They are not used anywhere.
>
> Indeed.
Let's remove them. We can always add them back if there is a need.
> The ki_retried users all seem pretty questionable, too. How about
> removing all that stuff?
Suparna wanted them around for debug purpose at one point. I don't know
whether that is still the case right now. At least I can wrap it around
with #if DEBUG.
- Ken
> Let's remove them. We can always add them back if there is a need.
Agreed, especially in this age of dynamic probing.
> Suparna wanted them around for debug purpose at one point. I don't know
> whether that is still the case right now. At least I can wrap it around
> with #if DEBUG.
I guess that's better than nothing. Given the near total lack of
-EIOCBRETRY users I'm just not convinced that they're worth it.
- z