linux-kernel - Re: [PATCH 1/1] iommu/arm-smmu-qcom: remove runtime pm enabling for TBU driver

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

Message-ID: <ZrsJLqTnq6tG2xp4@google.com>
Date: Tue, 13 Aug 2024 07:20:14 +0000
From: Pranjal Shrivastava <praan@...gle.com>
To: Zhenhua Huang <quic_zhenhuah@...cinc.com>
Cc: robdclark@...il.com, will@...nel.org, robin.murphy@....com,
	joro@...tes.org, iommu@...ts.linux.dev,
	linux-arm-msm@...r.kernel.org, linux-arm-kernel@...ts.infradead.org,
	linux-kernel@...r.kernel.org, quic_c_gdjako@...cinc.com,
	dmitry.baryshkov@...aro.org
Subject: Re: [PATCH 1/1] iommu/arm-smmu-qcom: remove runtime pm enabling for
 TBU driver

On Tue, Aug 13, 2024 at 10:37:33AM +0800, Zhenhua Huang wrote:
> 
> 
> On 2024/8/12 21:25, Pranjal Shrivastava wrote:
> > On Tue, Jul 30, 2024 at 06:30:43PM +0800, Zhenhua Huang wrote:
> > > TBU driver has no runtime pm support now, adding pm_runtime_enable()
> > > seems to be useless. Remove it.
> > > 
> > > Signed-off-by: Zhenhua Huang <quic_zhenhuah@...cinc.com>
> > > ---
> > >   drivers/iommu/arm/arm-smmu/arm-smmu-qcom.c | 6 ------
> > >   1 file changed, 6 deletions(-)
> > > 
> > > diff --git a/drivers/iommu/arm/arm-smmu/arm-smmu-qcom.c b/drivers/iommu/arm/arm-smmu/arm-smmu-qcom.c
> > > index 36c6b36ad4ff..aff2fe1fda13 100644
> > > --- a/drivers/iommu/arm/arm-smmu/arm-smmu-qcom.c
> > > +++ b/drivers/iommu/arm/arm-smmu/arm-smmu-qcom.c
> > > @@ -566,7 +566,6 @@ static struct acpi_platform_list qcom_acpi_platlist[] = {
> > >   static int qcom_smmu_tbu_probe(struct platform_device *pdev)
> > >   {
> > > -	struct device *dev = &pdev->dev;
> > >   	int ret;
> > >   	if (IS_ENABLED(CONFIG_ARM_SMMU_QCOM_DEBUG)) {
> > > @@ -575,11 +574,6 @@ static int qcom_smmu_tbu_probe(struct platform_device *pdev)
> > >   			return ret;
> > >   	}
> > > -	if (dev->pm_domain) {
> > > -		pm_runtime_set_active(dev);
> > > -		pm_runtime_enable(dev);
> > 
> > I assumed that this was required to avoid the TBU from being powered
> > down? If so, then I think we shall move it under the
> 
> Hi Pranjal,
> 
> In my sense, this was giving the TBU ability to power down when
> necessary(through pm callbacks)? While I haven't seen any RPM impl for TBU
> device.. hence having the doubt..
> 
> Thanks,
> Zhenhua

Apologies for being unclear. I just meant to ask if there was a reason
to add pm_runtime_set_active & enable in the tbu probe previously? And I
*assumed* that it was to set the device state as RPM_ACTIVE to avoid it
being RPM_SUSPENDED after enabling pm_runtime. 

I agree that there are no pm_runtime_suspend/resume calls within the TBU
driver. I'm just trying to understand why was pm_runtime enabled here
earlier (since it's not implemented) in order to ensure that removing it
doesn't cause further troubles?

I see Georgi added it as a part of
https://lore.kernel.org/all/20240704010759.507798-1-quic_c_gdjako@quicinc.com/

But I'm unsure why was it required to fix that bug?

> 
> > previous if condition, i.e. CONFIG_ARM_SMMU_QCOM_DEBUG?
> > 
> > If not, we can remove it give that the TBU would be powered ON as needed
> > 
> > > -	}
> > > -
> > >   	return 0;
> > >   }
> > > -- 
> > > 2.7.4
> > > 
> > > 
> > 
> > Thanks,
> > Pranjal

Thanks,
Pranjal