Yeoh Chun Yeow wrote: > Dear Jan, > > --> Would be nice to see it inside drivers/, and to have the required > changes to the build system done as well - when time permits. Then generate > a true patch from your changes (just like here [1]) and repost.
Please fix your mail client to use standard style for citing the
original message!
> I have managed to create Makefile for my driver in the RTnet tree and build
> successfully in rtnet/driver/at91/ directory but not in the rtnet/
> directory. Could you help on this?
I can, but I can't promise when I will be able to do this. Why not look
out for patterns: Pick some existing driver, say rt_pcnet32, search for
string occurrences of "pcnet32" (case-independent, of course), and
duplicate&adopt them to your driver. That's what I would do as well :),
as I don't add driver to the tree every day.
>
> The printk statements in the do_request_cal_job are as follow:
>
> static struct tdma_job *do_request_cal_job(struct tdma_priv *tdma, struct
> tdma_request_cal *job, rtdm_lockctx_t lockctx)
> {
> struct rt_proc_call *call;
> struct tdma_job *prev_job;
> int err;
>
> printk("[Debug] callibration start 1\n");
>
> if ((job->period != 1) &&
> (tdma->current_cycle % job->period != job->phasing))
> return &job->head;
>
> printk("[Debug] callibration start 2\n");
>
> /* remove job until we get a reply */
> __list_del(job->head.entry.prev, job->head.entry.next);
> job->head.ref_count--;
> prev_job = tdma->current_job =
> list_entry(job->head.entry.prev, struct tdma_job, entry);
> prev_job->ref_count++;
> tdma->job_list_revision++;
>
> printk("[Debug] callibration start 3\n");
>
> rtdm_lock_put_irqrestore(&tdma->lock, lockctx);
>
> printk("[Debug] callibration start 4 [delay] %lld\n",
> tdma->current_cycle_start + job->offset);
>
> rtdm_task_sleep_until(tdma->current_cycle_start + job->offset);
>
> printk("[Debug] callibration start 5\n");
>
> err = tdma_xmit_request_cal_frame(tdma,
> tdma->current_cycle + job->period, job->offset);
>
> rtdm_lock_get_irqsave(&tdma->lock, lockctx);
>
> printk("delay: %lld\n", tdma->current_cycle_start + job->offset -
> rtdm_clock_read());
>
> /* terminate call on error */
> if (err < 0) {
> call = tdma->calibration_call;
> tdma->calibration_call = NULL;
>
> if (call) {
> rtdm_lock_put_irqrestore(&tdma->lock, lockctx);
> rtpc_complete_call(call, err);
> rtdm_lock_get_irqsave(&tdma->lock, lockctx);
> }
> }
>
> return prev_job;
> }
> Result of Kernel Message:
> [Debug] Req Cal Stage
> [Debug] callibration start 1
> [Debug] callibration start 2
> [Debug] callibration start 3
> [Debug] callibration start 4 [delay] 1189571585923819617
Ah, that's more logical than the earlier report!
> RTcfg: rtcfg_do_main_event() rtdev=1, event=RTCFG_FRM_STAGE_1_CFG,
> state=RTCFG_MAIN_CLIENT_1
>
> Is this because of the delay is too long? Any ideas?
That delay is definitely too long (1 billion seconds...). Please look
closer which component of the calculation is borken. I guess it is
current_cycle_start, but it may also be the job offset. Then look at
their origin: job offset should come unmodified from the ioctl, ie.
tdmacfg, current_cycle_start is calculated here [1]. Regarding the
latter: it will be interesting to check what timestamp your driver took
for the related Sync frame. Please also check this.
Thanks,
Jan
[1]http://www.rts.uni-hannover.de/rtnet/lxr/source/stack/rtmac/tdma/tdma_proto.c#246
PS: Already thought about subscribing to rtnet-users...?
signature.asc
Description: OpenPGP digital signature
------------------------------------------------------------------------- This SF.net email is sponsored by: Microsoft Defy all challenges. Microsoft(R) Visual Studio 2005. http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/
_______________________________________________ RTnet-users mailing list RTnet-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/rtnet-users
