From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from host.buserror.net (host.buserror.net [209.198.135.123]) (using TLSv1.2 with cipher ECDHE-RSA-AES128-GCM-SHA256 (128/128 bits)) (No client certificate requested) by lists.ozlabs.org (Postfix) with ESMTPS id 3wK2Cn2MnlzDq5W for ; Fri, 5 May 2017 16:33:40 +1000 (AEST) Message-ID: <1493966009.25397.39.camel@buserror.net> From: Scott Wood To: Karim Eshapa 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 Date: Fri, 05 May 2017 01:33:29 -0500 In-Reply-To: <1493964114-11296-1-git-send-email-karim.eshapa@gmail.com> References: <1493873917-7484-1-git-send-email-karim.eshapa@gmail.com> <1493964114-11296-1-git-send-email-karim.eshapa@gmail.com> Content-Type: text/plain; charset="UTF-8" Mime-Version: 1.0 Subject: Re: [PATCH v2] drivers:soc:fsl:qbman:qman.c: Sleep instead of stuck hacking jiffies. List-Id: Linux on PowerPC Developers Mail List List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , On Fri, 2017-05-05 at 08:01 +0200, Karim Eshapa wrote: > > On 5/4/2017 5:07 PM, Scott Wood wrote: > > > On Thu, 2017-05-04 at 06:58 +0200, Karim Eshapa wrote: > > > > +    stop = jiffies + 10000; > > > > +    /* > > > > +     * 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. > > > > +     */ > > > > > > OK, upon reading this more closely it seems the intent was to delay for > > > 10,000 > > > *processor cycles* and somehow that got turned into 10,000 jiffies > > > (which is > > > 40 seconds at the default Hz!).  We could just replace this whole thing > > > with > > > msleep(1) and still be far more paranoid than was originally intended. > > > > > > Claudiu and Roy, any comments? > > > > Yes the timing here is certainly off, the code changed a few times since > > the comment was originally written. > > An msleep(1) seems reasonable here to me. > > If the previous patch with msleep(1) is OK. > can I send a patch to slightly change the comments. Yes. -Scott From mboxrd@z Thu Jan 1 00:00:00 1970 From: oss@buserror.net (Scott Wood) Date: Fri, 05 May 2017 01:33:29 -0500 Subject: [PATCH v2] drivers:soc:fsl:qbman:qman.c: Sleep instead of stuck hacking jiffies. In-Reply-To: <1493964114-11296-1-git-send-email-karim.eshapa@gmail.com> References: <1493873917-7484-1-git-send-email-karim.eshapa@gmail.com> <1493964114-11296-1-git-send-email-karim.eshapa@gmail.com> Message-ID: <1493966009.25397.39.camel@buserror.net> To: linux-arm-kernel@lists.infradead.org List-Id: linux-arm-kernel.lists.infradead.org On Fri, 2017-05-05 at 08:01 +0200, Karim Eshapa wrote: > > On 5/4/2017 5:07 PM, Scott Wood wrote: > > > On Thu, 2017-05-04 at 06:58 +0200, Karim Eshapa wrote: > > > > +????stop = jiffies + 10000; > > > > +????/* > > > > +?????* 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. > > > > +?????*/ > > > > > > OK, upon reading this more closely it seems the intent was to delay for > > > 10,000 > > > *processor cycles* and somehow that got turned into 10,000 jiffies > > > (which is > > > 40 seconds at the default Hz!).??We could just replace this whole thing > > > with > > > msleep(1) and still be far more paranoid than was originally intended. > > > > > > Claudiu and Roy, any comments? > > > > Yes the timing here is certainly off, the code changed a few times since > > the comment was originally written. > > An msleep(1) seems reasonable here to me. > > If the previous patch with msleep(1) is OK. > can I send a patch to slightly change the comments. Yes. -Scott