> On Sun, Aug 11, 2024 at 06:19:49AM +0000, Avri Altman wrote: > > > On Fri, Aug 09, 2024 at 10:23:31AM +0300, Avri Altman wrote: > > > > The standard register map of UFSHCI is comprised of several groups. > > > > The first group (starting from offset 0x00), is the host capabilities group. > > > > It contains some interesting information, that otherwise is not > > > > available, e.g. the UFS version of the platform etc. > > > > > > > > Reviewed-by: Keoseong Park <keosung.park@xxxxxxxxxxx> > > > > Reviewed-by: Bart Van Assche <bvanassche@xxxxxxx> > > > > Signed-off-by: Avri Altman <avri.altman@xxxxxxx> > > > > --- > > > > Documentation/ABI/testing/sysfs-driver-ufs | 42 ++++++++++ > > > > drivers/ufs/core/ufs-sysfs.c | 95 ++++++++++++++++++++++ > > > > 2 files changed, 137 insertions(+) > > > > > > > > diff --git a/Documentation/ABI/testing/sysfs-driver-ufs > > > > b/Documentation/ABI/testing/sysfs-driver-ufs > > > > index fe943ce76c60..b6e0c3b806fd 100644 > > > > --- a/Documentation/ABI/testing/sysfs-driver-ufs > > > > +++ b/Documentation/ABI/testing/sysfs-driver-ufs > > > > @@ -1532,3 +1532,45 @@ Contact: Bean Huo <beanhuo@xxxxxxxxxx> > > > > Description: > > > > rtc_update_ms indicates how often the host should > > > > synchronize or > > > update the > > > > UFS RTC. If set to 0, this will disable UFS RTC periodic update. > > > > + > > > > +What: /sys/devices/platform/.../ufshci_capabilities/capabilities > > > > +Date: August 2024 > > > > +Contact: Avri Altman <avri.altman@xxxxxxx> > > > > +Description: > > > > + Host Capabilities register group: host controller capabilities register. > > > > + Symbol - CAP. Offset: 0x00 - 0x03. > > > > > > This doesn't look like an ABI description. You are merely specifying > > > the register name and offset that gets accessed while reading this attribute. > > > > > > Also, I'm not sure if we really want to expose HCI/MCQ capabilities as sysfs > ABI. > > Why not? > > sysfs is an userspace ABI. The information that is exposed is supposed to be > consumed by the sysadmins/userspace applications. > > Looking at the contents of the capabilities, I don't think anyone could interpret > what is being exposed unless they have access to the spec (which is not free). > > > Like the cover letter say, this info is otherwise unavailable. > > There are other ways to expose this information too, like debugfs. But what > exactly do you want users to do with it? OK. Will drop those. Thanks, Avri > > - Mani > > -- > மணிவண்ணன் சதாசிவம்
