linux-kernel.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Karim Eshapa <karim.eshapa@gmail.com>
To: oss@buserror.net
Cc: claudiu.manoil@nxp.com, roy.pledge@nxp.com,
	colin.king@canonical.com, linuxppc-dev@lists.ozlabs.org,
	linux-arm-kernel@lists.infradead.org,
	linux-kernel@vger.kernel.org,
	Karim Eshapa <karim.eshapa@gmail.com>
Subject: RE:drivers:soc:fsl:qbman:qman.c: unsigned long jiffies value
Date: Sun, 30 Apr 2017 03:09:26 +0200	[thread overview]
Message-ID: <1493514566-12920-1-git-send-email-karim.eshapa@gmail.com> (raw)
In-Reply-To: <1493508775.25397.26.camel@buserror.net>

On Sat, 29 Apr 2017 18:32:55 -0500, Scott Wood wrote:
>On Sat, 2017-04-29 at 22:43 +0200, Karim Eshapa wrote:
>
>> unsigned long jiffies value sorry for that.
>>
> You mean unsigned long msecs?
>

Yes, I mean usecs. 

>>
>> Signed-off-by: Karim Eshapa <karim.eshapa@gmail.com>
>> ---
>>  drivers/soc/fsl/qbman/qman.c | 2 +-
>>  1 file changed, 1 insertion(+), 1 deletion(-)
>>
>> diff --git a/drivers/soc/fsl/qbman/qman.c b/drivers/soc/fsl/qbman/qman.c
>> index e0df4d1..6e1a44a 100644
>> --- a/drivers/soc/fsl/qbman/qman.c
>> +++ b/drivers/soc/fsl/qbman/qman.c
>> @@ -1084,7 +1084,7 @@ static int drain_mr_fqrni(struct qm_portal *p)
>>                * entries well before the ring has been fully consumed, so
>>                * we're being *really* paranoid here.
>>                */
>> -             unsigned int udel_time = jiffies_to_usecs(10000);
>> +             unsigned long udel_time = jiffies_to_usecs(10000);
>>  
>>               usleep_range(udel_time/2, udel_time);
>>               msg = qm_mr_current(p);

>If unsigned int isn't big enough, then unsigned long won't be either on 32-
>bit.  With such a long delay why not use msleep()?
>

I agree with you in long int.
After looking at Documentation/timers/timers-howto.txt I think
the msleep() is better as we actually have long delay.
may be we can use jiffies_to_msecs() with msleep() in case of we still have 
this large jiffies difference.

>As for the previous patch[1], you're halving the minimum timeout which may not
>be correct.
>
>

For the minimum timeout, I've read the following comments.

		 /*
                  * if MR was full and h/w had other FQRNI entries to produce, we
                  * need to allow it time to produce those entries once the
                  * existing entries are consumed. A worst-case situation
                  * (fully-loaded system) means h/w sequencers may have to do 3-4
                  * other things before servicing the portal's MR pump, each of
                  * which (if slow) may take ~50 qman cycles (which is ~200
                  * processor cycles). So rounding up and then multiplying this
                  * worst-case estimate by a factor of 10, just to be
                  * ultra-paranoid, goes as high as 10,000 cycles. NB, we consume
                  * one entry at a time, so h/w has an opportunity to produce new
                  * entries well before the ring has been fully consumed, so
                  * we're being *really* paranoid here.
		  */
		  u64 now, then = jiffies;
		
		do {
	                now = jiffies;
                 } while ((then + 10000) > now);  

He needs to guarantee certain action so, he made a very large factor of saftey
therefore I've used sleep in range with approximate delay. But we still need 
the driver owner to define appropriate value to put.

>[1] When fixing a patch you've already posted that hasn't yet been applied,
>send a replacement (v2) patch rather than a separate fix.
>

Thanks so much, I'm just newbies :)


Karim   

  reply	other threads:[~2017-04-30  1:10 UTC|newest]

Thread overview: 12+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2017-04-29 19:46 [PATCH] drivers:soc:fsl:qbman:qman.c: Sleep instead of stuck hacking jiffies Karim Eshapa
2017-04-29 20:43 ` [PATCH] drivers:soc:fsl:qbman:qman.c: unsigned long jiffies value Karim Eshapa
2017-04-29 23:32   ` Scott Wood
2017-04-30  1:09     ` Karim Eshapa [this message]
2017-05-04  4:58     ` [PATCH v2] drivers:soc:fsl:qbman:qman.c: Sleep instead of stuck hacking jiffies Karim Eshapa
2017-05-04 21:07       ` Scott Wood
2017-05-04 23:30         ` Roy Pledge
2017-05-05  5:45       ` [PATCH v3] " Karim Eshapa
2017-06-25  2:46         ` [v3] " Scott Wood
2017-06-27 16:38           ` Leo Li
2017-05-05  6:01       ` [PATCH v2] " Karim Eshapa
2017-05-05  6:33         ` Scott Wood

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=1493514566-12920-1-git-send-email-karim.eshapa@gmail.com \
    --to=karim.eshapa@gmail.com \
    --cc=claudiu.manoil@nxp.com \
    --cc=colin.king@canonical.com \
    --cc=linux-arm-kernel@lists.infradead.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linuxppc-dev@lists.ozlabs.org \
    --cc=oss@buserror.net \
    --cc=roy.pledge@nxp.com \
    --subject='RE:drivers:soc:fsl:qbman:qman.c: unsigned long jiffies value' \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).