linux-kernel - Re: [PATCH V3 2/3] scsi: ufs-qcom: Add support for dumping MCQ registers

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite for Android: free password hash cracker in your pocket

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <awc2ql2x5amiahf7l47xqhgl7ugi4zpk5wz7qycgbqb52gh4yb@24za7q2rqqob>
Date: Mon, 24 Mar 2025 13:09:29 +0530
From: Manivannan Sadhasivam <manivannan.sadhasivam@...aro.org>
To: MANISH PANDEY <quic_mapa@...cinc.com>
Cc: "James E.J. Bottomley" <James.Bottomley@...senpartnership.com>, 
	"Martin K. Petersen" <martin.petersen@...cle.com>, linux-arm-msm@...r.kernel.org, linux-scsi@...r.kernel.org, 
	linux-kernel@...r.kernel.org, quic_nitirawa@...cinc.com, quic_cang@...cinc.com, 
	quic_nguyenb@...cinc.com
Subject: Re: [PATCH V3 2/3] scsi: ufs-qcom: Add support for dumping MCQ
 registers

On Wed, Mar 19, 2025 at 11:51:07AM +0530, MANISH PANDEY wrote:
> 
> 
> On 3/18/2025 12:14 PM, Manivannan Sadhasivam wrote:
> > On Thu, Mar 13, 2025 at 10:46:34AM +0530, Manish Pandey wrote:
> > > This patch adds functionality to dump MCQ registers.
> > > This will help in diagnosing issues related to MCQ
> > > operations by providing detailed register dumps.
> > > 
> > 
> > Same comment as previous patch. Also, make use of 75 column width.
> > 
> will Update in next patch set.>> Signed-off-by: Manish Pandey
> <quic_mapa@...cinc.com>
> > > ---
> > > 
> > > Changes in v3:
> > > - Addressed Bart's review comments by adding explanations for the
> > >    in_task() and usleep_range() calls.
> > > Changes in v2:
> > > - Rebased patchsets.
> > > - Link to v1: https://lore.kernel.org/linux-arm-msm/20241025055054.23170-1-quic_mapa@quicinc.com/
> > > ---
> > >   drivers/ufs/host/ufs-qcom.c | 60 +++++++++++++++++++++++++++++++++++++
> > >   drivers/ufs/host/ufs-qcom.h |  2 ++
> > >   2 files changed, 62 insertions(+)
> > > 
> > > diff --git a/drivers/ufs/host/ufs-qcom.c b/drivers/ufs/host/ufs-qcom.c
> > > index f5181773c0e5..fb9da04c0d35 100644
> > > --- a/drivers/ufs/host/ufs-qcom.c
> > > +++ b/drivers/ufs/host/ufs-qcom.c
> > > @@ -1566,6 +1566,54 @@ int ufs_qcom_testbus_config(struct ufs_qcom_host *host)
> > >   	return 0;
> > >   }
> > > +static void ufs_qcom_dump_mcq_hci_regs(struct ufs_hba *hba)
> > > +{
> > > +	/* sleep intermittently to prevent CPU hog during data dumps. */
> > > +	/* RES_MCQ_1 */
> > > +	ufshcd_dump_regs(hba, 0x0, 256 * 4, "MCQ HCI 1da0000-1da03f0 ");
> > > +	usleep_range(1000, 1100);
> > 
> > If your motivation is just to not hog the CPU, use cond_resched().
> > 
> > - Mani
> > 
> The intention here is to introduce a specific delay between each dump.

What is the reason for that?

> Therefore, i would like to use usleep_range() instead of cond_resched().
> Please let me know if i am getting it wrong..
> 

Without knowing the reason, I cannot judge. Your comment said that you do not
want to hog the CPU during dump. But now you are saying that you wanted to have
a delay. Both are contradictions.

- Mani

-- 
மணிவண்ணன் சதாசிவம்