On Fri, Nov 05, 2004 at 02:47:02AM -0500, Jeff Garzik wrote:
> Nishanth Aravamudan wrote:
> >On Sun, Oct 31, 2004 at 06:06:23AM -0500, Jeff Garzik wrote:
> >
> >>janitor@xxxxxxxxxxxxxx wrote:
> >>
> >>>@@ -528,7 +517,7 @@ abort(struct net_device *dev, u32 abort_
> >>> // wait for abort to complete
> >>> while (GT96100ETH_READ(gp, GT96100_ETH_SDMA_COMM) & abort_bits) {
> >>> // snooze for 20 msec and check again
> >>>- gt96100_delay(1);
> >>>+ msleep_interruptible(20);
> >>>
> >>> if (--timedout == 0) {
> >>
> >>
> >>don't change the behavior of the driver, even if it disagrees with the
> >>comment.
> >
> >
> >Ok, please find the corrected patch below.
> >
> >Description: Uses msleep_interruptible() instead of gt96100_delay()
> >to guarantee the task delays as expected. Removes prototype and
> >definition of now unused gt96100_delay() function. Corrects one comment
> >to correspond to the code.
> >
> >Signed-off-by: Nishanth Aravamudan <nacc@xxxxxxxxxx>
> >
> >
> >--- 2.6.10-rc1-vanilla/drivers/net/gt96100eth.c 2004-10-30
> >15:33:30.000000000 -0700
> >+++ 2.6.10-rc1/drivers/net/gt96100eth.c 2004-11-01
> >11:05:37.000000000 -0800
> >@@ -59,7 +59,6 @@
> > // prototypes
> > static void* dmaalloc(size_t size, dma_addr_t *dma_handle);
> > static void dmafree(size_t size, void *vaddr);
> >-static void gt96100_delay(int msec);
> > static int gt96100_add_hash_entry(struct net_device *dev,
> > unsigned char* addr);
> > static void read_mib_counters(struct gt96100_private *gp);
> >@@ -183,16 +182,6 @@ static void dmafree(size_t size, void *v
> > free_pages((unsigned long)vaddr, get_order(size));
> > }
> >
> >-static void gt96100_delay(int ms)
> >-{
> >- if (in_interrupt())
> >- return;
> >- else {
> >- current->state = TASK_INTERRUPTIBLE;
> >- schedule_timeout(ms*HZ/1000);
>
> I'm still worried about this patch, because
>
> 1) in_interrupt() check disappears, and
This is true... I have changed the patch significantly, leaving the
gt96100_delay() function intact and simply replacing the else result
with msleep_interruptible(). Does this newer patch look better to you?
Thanks for all the input!
> 2) does driver already include linux/delay.h ?
Yes, it does.
Description: Uses msleep_interruptible() instead of schedule_timeout()
in the gt96100_delay() function. Corrects one comment to correspond to
the code.
Signed-off-by: Nishanth Aravamudan <nacc@xxxxxxxxxx>
--- 2.6.10-rc1-vanilla/drivers/net/gt96100eth.c 2004-11-05 09:42:37.000000000
-0800
+++ 2.6.10-rc1/drivers/net/gt96100eth.c 2004-11-05 09:45:48.000000000 -0800
@@ -187,10 +187,8 @@ static void gt96100_delay(int ms)
{
if (in_interrupt())
return;
- else {
- current->state = TASK_INTERRUPTIBLE;
- schedule_timeout(ms*HZ/1000);
- }
+ else
+ msleep_interruptible(ms);
}
static int
@@ -527,7 +525,7 @@ abort(struct net_device *dev, u32 abort_
// wait for abort to complete
while (GT96100ETH_READ(gp, GT96100_ETH_SDMA_COMM) & abort_bits) {
- // snooze for 20 msec and check again
+ // snooze for 1 msec and check again
gt96100_delay(1);
if (--timedout == 0) {
|