[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <97c39c0cdb6b96d791fe05f8e5496a502a7e6ac6.camel@iokpp.de>
Date: Mon, 13 Oct 2025 17:42:56 +0200
From: Bean Huo <beanhuo@...pp.de>
To: Jens Wiklander <jens.wiklander@...aro.org>
Cc: avri.altman@....com, bvanassche@....org, alim.akhtar@...sung.com,
jejb@...ux.ibm.com, martin.petersen@...cle.com, can.guo@....qualcomm.com,
ulf.hansson@...aro.org, beanhuo@...ron.com, linux-scsi@...r.kernel.org,
linux-kernel@...r.kernel.org
Subject: Re: [PATCH v2 3/3] scsi: ufs: core: Add OP-TEE based RPMB driver
for UFS devices
On Mon, 2025-10-13 at 14:22 +0200, Jens Wiklander wrote:
> >
> > For certain memory vendors, the serial number is guaranteed to be unique
> > among
> > all devices.
>
> This is supposed to be generic and not rely on the behavior of only
> some vendors.
>
> >
> > For partitions or regions, we have appended the region number to the end of
> > the
> > CID — please check the patch for details.
>
> Yes, but how do you know that you don't overwrite a part of the serial number?
>
> >
> > Regarding improving CID uniqueness, we could include the OEM ID or product
> > number. However, this would make the CID longer than 16 bytes.
>
> UFS doesn't have a CID, but there's no need for one either. struct
> rpmb_descr has dev_id and dev_id_len. It can be any length, within
> reason.
>
> Cheers,
> Jens
Hi Jens,
how about combining wManufacturerID, wManufactureDate, wDeviceVersion, Serial
Number (in unicode), plus the region number?
Kind regards,
Bean
Powered by blists - more mailing lists