linux-kernel - Re: [PATCH v3 19/19] PCI: qcom: Expose link transition counts via debugfs for v2.4.0

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: Windows password security audit tool. GUI, reports in PDF.

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

Date:   Tue, 14 Mar 2023 16:34:31 +0530
From:   Manivannan Sadhasivam <manivannan.sadhasivam@...aro.org>
To:     Sricharan Ramabadhran <quic_srichara@...cinc.com>
Cc:     andersson@...nel.org, lpieralisi@...nel.org, kw@...ux.com,
        krzysztof.kozlowski+dt@...aro.org, robh@...nel.org,
        konrad.dybcio@...aro.org, linux-arm-msm@...r.kernel.org,
        devicetree@...r.kernel.org, linux-pci@...r.kernel.org,
        linux-kernel@...r.kernel.org
Subject: Re: [PATCH v3 19/19] PCI: qcom: Expose link transition counts via
 debugfs for v2.4.0

On Tue, Mar 14, 2023 at 04:25:29PM +0530, Sricharan Ramabadhran wrote:
> 
> 
> On 3/14/2023 4:06 PM, Sricharan Ramabadhran wrote:
> > 
> > 
> > On 3/10/2023 9:38 AM, Manivannan Sadhasivam wrote:
> > > Qualcomm PCIe controllers of version v2.4.0 have debug registers in the
> > > PARF region that count PCIe link transitions. Expose them over debugfs to
> > > userspace to help debug the low power issues.
> > > 
> > > Signed-off-by: Manivannan Sadhasivam <manivannan.sadhasivam@...aro.org>
> > > ---
> > >   drivers/pci/controller/dwc/pcie-qcom.c | 33 ++++++++++++++++++++++++++
> > >   1 file changed, 33 insertions(+)
> > > 
> > > diff --git a/drivers/pci/controller/dwc/pcie-qcom.c
> > > b/drivers/pci/controller/dwc/pcie-qcom.c
> > > index f99b7e7f3f73..0b41f007fa90 100644
> > > --- a/drivers/pci/controller/dwc/pcie-qcom.c
> > > +++ b/drivers/pci/controller/dwc/pcie-qcom.c
> > > @@ -37,6 +37,7 @@
> > >   /* PARF registers */
> > >   #define PARF_SYS_CTRL                0x00
> > >   #define PARF_PM_CTRL                0x20
> > > +#define PARF_PM_STTS                0x24
> > >   #define PARF_PCS_DEEMPH                0x34
> > >   #define PARF_PCS_SWING                0x38
> > >   #define PARF_PHY_CTRL                0x40
> > > @@ -84,6 +85,12 @@
> > >   /* PARF_PM_CTRL register fields */
> > >   #define REQ_NOT_ENTR_L1                BIT(5)
> > > +/* PARF_PM_STTS register fields */
> > > +#define PM_LINKST_IN_L1SUB            BIT(8)
> > > +#define PM_LINKST_IN_L0S            BIT(7)
> > > +#define PM_LINKST_IN_L2                BIT(5)
> > > +#define PM_LINKST_IN_L1                BIT(4)
> > > +
> > >   /* PARF_PCS_DEEMPH register fields */
> > >   #define PCS_DEEMPH_TX_DEEMPH_GEN1(x)        FIELD_PREP(GENMASK(21,
> > > 16), x)
> > >   #define PCS_DEEMPH_TX_DEEMPH_GEN2_3_5DB(x)   
> > > FIELD_PREP(GENMASK(13, 8), x)
> > > @@ -737,6 +744,31 @@ static int qcom_pcie_post_init_2_4_0(struct
> > > qcom_pcie *pcie)
> > >       return 0;
> > >   }
> > > +static int qcom_pcie_debugfs_func_2_4_0(struct seq_file *s, void *data)
> > > +{
> > > +    struct qcom_pcie *pcie = (struct qcom_pcie *)
> > > dev_get_drvdata(s->private);
> > > +
> > > +    seq_printf(s, "L0s transition count: %u\n",
> > > +           readl_relaxed(pcie->parf + PM_LINKST_IN_L0S));
> > > +
> > > +    seq_printf(s, "L1 transition count: %u\n",
> > > +           readl_relaxed(pcie->parf + PM_LINKST_IN_L1));
> > > +
> > > +    seq_printf(s, "L1.1 transition count: %u\n",
> > > +           readl_relaxed(pcie->parf + PM_LINKST_IN_L1SUB));
> > > +
> > > +    seq_printf(s, "L2 transition count: %u\n",
> > > +           readl_relaxed(pcie->parf + PM_LINKST_IN_L2));
> > > +
> > 
> >   Using bitmask as register offset ? instead use PM_STTS and bitmask it ?
> 
>  Also, since its 1 bit, all are status and not count.
>  Not sure, if you want it to limit this debug based on 'mhi' property
>  being populated ?
> 

Err... Look like I blindly copied the debugfs function from 2.7.0 :/ Sry, since
these are all just 1 bit, we cannot use it for getting the count. So I'll drop
this patch in next revision.

Thanks,
Mani

> Regards,
>  Sricharan

-- 
மணிவண்ணன் சதாசிவம்