linux-kernel - Re: [PATCH v1 1/1] scsi: ufshcd: Properly set the device Icc Level

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: <6e36f0a315c13429bdad1ce704cbe878@codeaurora.org>
Date:   Wed, 16 Sep 2020 17:53:37 -0700
From:   nguyenb@...eaurora.org
To:     Bjorn Andersson <bjorn.andersson@...aro.org>
Cc:     cang@...eaurora.org, asutoshd@...eaurora.org,
        martin.petersen@...cle.com, linux-scsi@...r.kernel.org,
        linux-arm-msm@...r.kernel.org,
        Alim Akhtar <alim.akhtar@...sung.com>,
        Avri Altman <avri.altman@....com>,
        "James E.J. Bottomley" <jejb@...ux.ibm.com>,
        Stanley Chu <stanley.chu@...iatek.com>,
        Bean Huo <beanhuo@...ron.com>,
        Bart Van Assche <bvanassche@....org>,
        open list <linux-kernel@...r.kernel.org>
Subject: Re: [PATCH v1 1/1] scsi: ufshcd: Properly set the device Icc Level

On 2020-09-15 06:37, Bjorn Andersson wrote:
> On Tue 15 Sep 03:49 CDT 2020, nguyenb@...eaurora.org wrote:
> 
>> On 2020-09-14 19:54, Bjorn Andersson wrote:
>> > On Tue 01 Sep 01:19 UTC 2020, Bao D. Nguyen wrote:
>> >
>> > > UFS version 3.0 and later devices require Vcc and Vccq power supplies
>> > > with Vccq2 being optional. While earlier UFS version 2.0 and 2.1
>> > > devices, the Vcc and Vccq2 are required with Vccq being optional.
>> > > Check the required power supplies used by the device
>> > > and set the device's supported Icc level properly.
>> > >
>> > > Signed-off-by: Can Guo <cang@...eaurora.org>
>> > > Signed-off-by: Asutosh Das <asutoshd@...eaurora.org>
>> > > Signed-off-by: Bao D. Nguyen <nguyenb@...eaurora.org>
>> > > ---
>> > >  drivers/scsi/ufs/ufshcd.c | 5 +++--
>> > >  1 file changed, 3 insertions(+), 2 deletions(-)
>> > >
>> > > diff --git a/drivers/scsi/ufs/ufshcd.c b/drivers/scsi/ufs/ufshcd.c
>> > > index 06e2439..fdd1d3e 100644
>> > > --- a/drivers/scsi/ufs/ufshcd.c
>> > > +++ b/drivers/scsi/ufs/ufshcd.c
>> > > @@ -6845,8 +6845,9 @@ static u32
>> > > ufshcd_find_max_sup_active_icc_level(struct ufs_hba *hba,
>> > >  {
>> > >  	u32 icc_level = 0;
>> > >
>> > > -	if (!hba->vreg_info.vcc || !hba->vreg_info.vccq ||
>> > > -						!hba->vreg_info.vccq2) {
>> > > +	if (!hba->vreg_info.vcc ||
>> >
>> > How did you test this?
>> >
>> > devm_regulator_get() never returns NULL, so afaict this conditional will
>> > never be taken with either the old or new version of the code.
>> Thanks for your comment. The call flow is as follows:
>> ufshcd_pltfrm_init->ufshcd_parse_regulator_info->ufshcd_populate_vreg
>> In the ufshcd_populate_vreg() function, it looks for DT entries 
>> "%s-supply"
>> For UFS3.0+ devices, "vccq2-supply" is optional, so the vendor may 
>> choose
>> not to provide vccq2-supply in the DT.
>> As a result, a NULL is returned to hba->vreg_info.vccq2.
>> Same for UFS2.0 and UFS2.1 devices, a NULL may be returned to
>> hba->vreg_info.vccq if vccq-supply is not provided in the DT.
>> The current code only checks for !hba->vreg_info.vccq OR
>> !hba->vreg_info.vccq2. It will skip the setting for icc_level
>> if either vccq or vccq2 is not provided in the DT.
>> >
> 
> Thanks for the pointers, I now see that the there will only be struct
> ufs_vreg objects allocated for the items that has an associated
> %s-supply.
> 
> FYI, the idiomatic way to handle optional regulators is to use
> regulator_get_optional(), which will return -ENODEV for regulators not
> specified.
Thanks for the regulator_get_optional() suggestion. Do you have a strong 
opinion about
using regulator_get_optional() or would my proposal be ok? With 
regulator_get_optional(),
we need to make 3 calls and check each result while the current 
implementation is also reliable
simple quick check for NULL without any potential problem.

Thanks,
Bao
> 
> Regards,
> Bjorn
> 
>> > Regards,
>> > Bjorn
>> >
>> > > +		(!hba->vreg_info.vccq && hba->dev_info.wspecversion >= 0x300) ||
>> > > +		(!hba->vreg_info.vccq2 && hba->dev_info.wspecversion < 0x300)) {
>> > >  		dev_err(hba->dev,
>> > >  			"%s: Regulator capability was not set, actvIccLevel=%d",
>> > >  							__func__, icc_level);
>> > > --
>> > > The Qualcomm Innovation Center, Inc. is a member of the Code Aurora
>> > > Forum,
>> > > a Linux Foundation Collaborative Project
>> > >