lkml.org 
[lkml]   [1999]   [Oct]   [21]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
SubjectRe: schedule_timeout() semantics/usage?
From
Date
Andrea Arcangeli <andrea@suse.de> writes:

> On Wed, 20 Oct 1999, David Hinds wrote:
>
> >Look at net/irda/*.c [..]
>
> /* Make sure nobody tries to transmit during the speed change */
> while (irda_lock((void *) &self->netdev.tbusy) == FALSE) {
> WARNING(__FUNCTION__ "(), device locked!\n");
> current->state = TASK_INTERRUPTIBLE;
> schedule_timeout(MSECS_TO_JIFFIES(10));
>
> if (n++ > 10) {
> WARNING(__FUNCTION__ "(), breaking loop!\n");
> break;
> }
> }

I've already sent a very large patch to Alan a couple of days ago, which
removes all the schedule_timeout's from the lower parts of the IrDA
stack. Instead I'm impl. things as state machines and do the timeouts using
timers. Some code can look a bit hairy because of this, but it's working
quite well.

struct irda_task;
typedef int (*TASK_CALLBACK) (struct irda_task *task);

struct irda_task {
queue_t q;
magic_t magic;

TASK_STATE state;
TASK_CALLBACK function;
TASK_CALLBACK finished;

struct irda_task *parent;
struct timer_list timer;

void *instance; /* Instance being called */
void *param; /* Parameter to be used by instance */
};

... then I can do things like this to make timeouts:

int tekram_reset(struct irda_task *task)
{
dongle_t *self = (dongle_t *) task->instance;
int ret = 0;

DEBUG(2, __FUNCTION__ "()\n");

ASSERT(task != NULL, return -1;);

/* Power off dongle */
self->set_dtr_rts(self->dev, FALSE, FALSE);

switch (task->state) {
case IRDA_TASK_INIT:
irda_task_next_state(task, IRDA_TASK_WAIT1);

/* Sleep 50 ms */
ret = MSECS_TO_JIFFIES(50);
break;
case IRDA_TASK_WAIT1:
/* Clear DTR, Set RTS */
self->set_dtr_rts(self->dev, FALSE, TRUE);

irda_task_next_state(task, IRDA_TASK_WAIT2);

/* Should sleep 1 ms, but 10 should not do any harm */
ret = MSECS_TO_JIFFIES(10);
break;

There are a couple of helping functions as well, but I guess you get my
point.

The reason is that the lower part is interrupt driven, and I previously had
to use a user-space daemon to help me get the process context for doing
schedule_timeout. Now there is no need for such a deamon anymore (at least
if my patch goes in, and when I have done a few more patches to IrLAN ;-)
Anyway, thanks for the info.

-- Dag

--
/ Dag Brattli | The Linux-IrDA Project /
// University of Tromsoe, Norway | Infrared communication for Linux //
/// http://www.cs.uit.no/~dagb | http://www.cs.uit.no/linux-irda/ ///

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/

\
 
 \ /
  Last update: 2005-03-22 13:54    [W:0.050 / U:0.236 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site