linux-kernel - Re: [PATCH 1/1] scsi: ufs-mediatek: Add UFSHCD_QUIRK_BROKEN_LSDBS

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: <ZsOJKMg8xlpdgoi5@kuroko.kudu-justice.ts.net>
Date: Mon, 19 Aug 2024 20:17:10 +0200
From: Mary Guillemard <mary@...y.zone>
To: Manivannan Sadhasivam <manisadhasivam.linux@...il.com>
Cc: linux-kernel@...r.kernel.org, linux-scsi@...r.kernel.org,
	Peter Wang <peter.wang@...iatek.com>,
	Stanley Jhu <chu.stanley@...il.com>,
	"James E.J. Bottomley" <James.Bottomley@...senpartnership.com>,
	"Martin K. Petersen" <martin.petersen@...cle.com>,
	Matthias Brugger <matthias.bgg@...il.com>,
	AngeloGioacchino Del Regno <angelogioacchino.delregno@...labora.com>
Subject: Re: [PATCH 1/1] scsi: ufs-mediatek: Add UFSHCD_QUIRK_BROKEN_LSDBS_CAP

On Mon, Aug 19, 2024 at 05:38:52PM +0530, Manivannan Sadhasivam wrote:
> On Mon, Aug 19, 2024 at 12:24:42AM +0200, Mary Guillemard wrote:
> > MT8183 supports UFSHCI 2.1 spec, but report a bogus value of 1 in the
> > reserved part for the Legacy Single Doorbell Support (LSDBS) capability.
> > 
> 
> Wow... I never thought that this quirk will be used outside of Qcom SoCs...
>

Yeah I found that by trial and error some weeks ago and noticed your
serie while looking to upstream this change, quite funny to see other
vendors having the same quirk here.
 
> > This set UFSHCD_QUIRK_BROKEN_LSDBS_CAP when MCQ support is explicitly
> > disabled, allowing the device to be properly registered.
> > 
> > Signed-off-by: Mary Guillemard <mary@...y.zone>
> > ---
> >  drivers/ufs/host/ufs-mediatek.c | 3 +++
> >  1 file changed, 3 insertions(+)
> > 
> > diff --git a/drivers/ufs/host/ufs-mediatek.c b/drivers/ufs/host/ufs-mediatek.c
> > index 02c9064284e1..9a5919434c4e 100644
> > --- a/drivers/ufs/host/ufs-mediatek.c
> > +++ b/drivers/ufs/host/ufs-mediatek.c
> > @@ -1026,6 +1026,9 @@ static int ufs_mtk_init(struct ufs_hba *hba)
> >  	if (host->caps & UFS_MTK_CAP_DISABLE_AH8)
> >  		hba->caps |= UFSHCD_CAP_HIBERN8_WITH_CLK_GATING;
> >  
> > +	if (host->caps & UFS_MTK_CAP_DISABLE_MCQ)
> 
> How can this be the deciding factor? You said above that the issue is with
> MT8183 SoC. So why not just use the quirk only for that platform?
> 
> - Mani
>

So my current assumption is that it also affect other Mediatek SoCs
that are also based on UFS 2.1 spec but I cannot check this.

Instead, we know that if MCQ isn't supported, we must fallback to LSDB
as there is no other ways to drive the device.

UFS_MTK_CAP_DISABLE_MCQ (mediatek,ufs-disable-mcq) being unused upstream,
I think that's an acceptable fix.

Another way to handle this would be to add a new dt property and add it
to ufs_mtk_host_caps but I feel that my approach should be enough. 

> > +		hba->quirks |= UFSHCD_QUIRK_BROKEN_LSDBS_CAP;
> > +
> >  	ufs_mtk_init_clocks(hba);
> >  
> >  	/*
> > -- 
> > 2.46.0
> > 
> > 
> 
> -- 
> மணிவண்ணன் சதாசிவம்

- Mary