| Message ID | 20181203142644.3038-1-wsa+renesas@sang-engineering.com (mailing list archive) |
|---|---|
| State | Superseded |
| Delegated to: | Geert Uytterhoeven |
| Headers | show |
| Series | mmc: document 'Reliable Write' bit in uapi header | expand |
Hi Wolfram, On Mon, Dec 3, 2018 at 3:27 PM Wolfram Sang <wsa+renesas@sang-engineering.com> wrote: > If we use it this way, people should know about it. > > Signed-off-by: Wolfram Sang <wsa+renesas@sang-engineering.com> Reviewed-by: Geert Uytterhoeven <geert+renesas@glider.be> > --- > > Still not super happy about shifting into the sign bit, but yeah... > don't break userspace, I guess. Indeed. > include/uapi/linux/mmc/ioctl.h | 5 ++++- > 1 file changed, 4 insertions(+), 1 deletion(-) > > diff --git a/include/uapi/linux/mmc/ioctl.h b/include/uapi/linux/mmc/ioctl.h > index 45f369dc0a42..ae659333e73f 100644 > --- a/include/uapi/linux/mmc/ioctl.h > +++ b/include/uapi/linux/mmc/ioctl.h > @@ -5,7 +5,10 @@ > #include <linux/types.h> > > struct mmc_ioc_cmd { > - /* Implies direction of data. true = write, false = read */ > + /* > + * Implies direction of data. true = write, false = read. nonzero = write, zero = read? > + * Bit 31 selects 'Reliable Write' for RPMB. Perhaps having a #define for this would be nice? Can also be used by the driver, instead of 1 << 31, which some static analyzers may complain about. > + */ > int write_flag; > > /* Application-specific command. true = precede with CMD55 */ Gr{oetje,eeting}s, Geert
> > Signed-off-by: Wolfram Sang <wsa+renesas@sang-engineering.com> > > Reviewed-by: Geert Uytterhoeven <geert+renesas@glider.be> Thanks, Geert! > > - /* Implies direction of data. true = write, false = read */ > > + /* > > + * Implies direction of data. true = write, false = read. > > nonzero = write, zero = read? Yes, even better. Will update. > > + * Bit 31 selects 'Reliable Write' for RPMB. > > Perhaps having a #define for this would be nice? > Can also be used by the driver, instead of 1 << 31, which some static > analyzers may complain about. Maybe. Should be an incremental patch then, I'd say.
diff --git a/include/uapi/linux/mmc/ioctl.h b/include/uapi/linux/mmc/ioctl.h index 45f369dc0a42..ae659333e73f 100644 --- a/include/uapi/linux/mmc/ioctl.h +++ b/include/uapi/linux/mmc/ioctl.h @@ -5,7 +5,10 @@ #include <linux/types.h> struct mmc_ioc_cmd { - /* Implies direction of data. true = write, false = read */ + /* + * Implies direction of data. true = write, false = read. + * Bit 31 selects 'Reliable Write' for RPMB. + */ int write_flag; /* Application-specific command. true = precede with CMD55 */
If we use it this way, people should know about it. Signed-off-by: Wolfram Sang <wsa+renesas@sang-engineering.com> --- Still not super happy about shifting into the sign bit, but yeah... don't break userspace, I guess. include/uapi/linux/mmc/ioctl.h | 5 ++++- 1 file changed, 4 insertions(+), 1 deletion(-)