linux-kernel - Re: [PATCH 2/3] ufs: core: Add a quirk for handling broken SDBS field in controller capabilities register

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]

Message-ID: <20240815034621.GB2562@thinkpad>
Date: Thu, 15 Aug 2024 09:16:21 +0530
From: Manivannan Sadhasivam <manivannan.sadhasivam@...aro.org>
To: Bart Van Assche <bvanassche@....org>
Cc: Alim Akhtar <alim.akhtar@...sung.com>,
	Avri Altman <avri.altman@....com>,
	"James E.J. Bottomley" <James.Bottomley@...senPartnership.com>,
	"Martin K. Petersen" <martin.petersen@...cle.com>,
	linux-scsi@...r.kernel.org, linux-kernel@...r.kernel.org,
	linux-arm-msm@...r.kernel.org, Kyoungrul Kim <k831.kim@...sung.com>
Subject: Re: [PATCH 2/3] ufs: core: Add a quirk for handling broken SDBS
 field in controller capabilities register

On Wed, Aug 14, 2024 at 10:29:11AM -0700, Bart Van Assche wrote:
> On 8/14/24 10:15 AM, Manivannan Sadhasivam via B4 Relay wrote:
> > From: Manivannan Sadhasivam <manivannan.sadhasivam@...aro.org>
> > 
> > 'Legacy Queue & Single Doorbell Support (SDBS)' field in the controller
> > capabilities register is supposed to be reserved for UFSHCI 3.0 based
> > controllers and should read as 0. But some controllers may report bogus
> > value of 1 due to the hardware bug. So let's add a quirk to handle those
> > controllers.
> > 
> > If the quirk is enabled by the controller driver and MCQ is not supported,
> > then 'hba->sdbs_sup' field will be ignored and the SCSI device will be
> > added in legacy/single doorbell mode.
> > 
> > Signed-off-by: Manivannan Sadhasivam <manivannan.sadhasivam@...aro.org>
> > ---
> >   drivers/ufs/core/ufshcd.c | 5 +++--
> >   include/ufs/ufshcd.h      | 7 +++++++
> >   2 files changed, 10 insertions(+), 2 deletions(-)
> > 
> > diff --git a/drivers/ufs/core/ufshcd.c b/drivers/ufs/core/ufshcd.c
> > index 168b9dbc3ada..acb6f261ecc2 100644
> > --- a/drivers/ufs/core/ufshcd.c
> > +++ b/drivers/ufs/core/ufshcd.c
> > @@ -10512,8 +10512,9 @@ int ufshcd_init(struct ufs_hba *hba, void __iomem *mmio_base, unsigned int irq)
> >   	}
> >   	if (!is_mcq_supported(hba)) {
> > -		if (!hba->sdbs_sup) {
> > -			dev_err(hba->dev, "%s: failed to initialize (legacy doorbell mode not supported)\n",
> > +		if (!(hba->quirks & UFSHCD_QUIRK_BROKEN_SDBS_CAP) && !hba->sdbs_sup) {
> > +			dev_err(hba->dev,
> > +				"%s: failed to initialize (legacy doorbell mode not supported)\n",
> >   				__func__);
> >   			err = -EINVAL;
> >   			goto out_disable;
> 
> Adding a check for the UFSHCD_QUIRK_BROKEN_SDBS_CAP quirk everywhere
> hba->sdbs_sup is tested is wrong. Please move this check to the code
> that assigns a value to hba->sdbs_sup.
> 

I thought it wouldn't matter since it is used in just one place, but I agree it
is best to keep it in assignment itself.

- Mani

-- 
மணிவண்ணன் சதாசிவம்