Re: [PATCH v1 1/1] scripts: dtc: mask flags bit when check i2c addr

[Rayagonda Kokatanur]

On Thu, Mar 26, 2020 at 10:04 AM Rayagonda Kokatanur
<rayagonda.kokatanur@xxxxxxxxxxxx> wrote:
>
> On Thu, Mar 26, 2020 at 12:37 AM Rob Herring <robh+dt@xxxxxxxxxx> wrote:
> >
> > On Mon, Mar 2, 2020 at 9:56 PM Rayagonda Kokatanur
> > <rayagonda.kokatanur@xxxxxxxxxxxx> wrote:
> > >
> > > On Fri, Feb 28, 2020 at 7:20 PM Rob Herring <robh+dt@xxxxxxxxxx> wrote:
> > > >
> > > > On Fri, Feb 28, 2020 at 2:48 AM Rayagonda Kokatanur
> > > > <rayagonda.kokatanur@xxxxxxxxxxxx> wrote:
> > > > >
> > > > > Generally i2c addr should not be greater than 10-bit. The highest 2 bits
> > > > > are used for I2C_TEN_BIT_ADDRESS and I2C_OWN_SLAVE_ADDRESS. Need to mask
> > > > > these flags if check slave addr valid.
> > > > >
> > > > > Signed-off-by: Rayagonda Kokatanur <rayagonda.kokatanur@xxxxxxxxxxxx>
> > > > > ---
> > > > >  scripts/dtc/Makefile | 2 +-
> > > > >  scripts/dtc/checks.c | 5 +++++
> > > > >  2 files changed, 6 insertions(+), 1 deletion(-)
> > > >
> > > > dtc changes must be submitted against upstream dtc.
> > >
> > > Please let me know link to clone the upstream dtc branch.
> > > >
> > > >
> > > > > diff --git a/scripts/dtc/Makefile b/scripts/dtc/Makefile
> > > > > index 3acbb410904c..c5e8d6a9e73c 100644
> > > > > --- a/scripts/dtc/Makefile
> > > > > +++ b/scripts/dtc/Makefile
> > > > > @@ -9,7 +9,7 @@ dtc-objs        := dtc.o flattree.o fstree.o data.o livetree.o treesource.o \
> > > > >  dtc-objs       += dtc-lexer.lex.o dtc-parser.tab.o
> > > > >
> > > > >  # Source files need to get at the userspace version of libfdt_env.h to compile
> > > > > -HOST_EXTRACFLAGS := -I $(srctree)/$(src)/libfdt
> > > > > +HOST_EXTRACFLAGS := -I $(srctree)/$(src)/libfdt -I$(srctree)/tools/include
> > > > >
> > > > >  ifeq ($(shell pkg-config --exists yaml-0.1 2>/dev/null && echo yes),)
> > > > >  ifneq ($(CHECK_DTBS),)
> > > > > diff --git a/scripts/dtc/checks.c b/scripts/dtc/checks.c
> > > > > index 756f0fa9203f..17c9ed4137b5 100644
> > > > > --- a/scripts/dtc/checks.c
> > > > > +++ b/scripts/dtc/checks.c
> > > > > @@ -3,6 +3,7 @@
> > > > >   * (C) Copyright David Gibson <dwg@xxxxxxxxxxx>, IBM Corporation.  2007.
> > > > >   */
> > > > >
> > > > > +#include <linux/bits.h>
> > > >
> > > > Not a UAPI header not that that would be much better as dtc also builds on Mac.
> > > >
> > > > >  #include "dtc.h"
> > > > >  #include "srcpos.h"
> > > > >
> > > > > @@ -17,6 +18,9 @@
> > > > >  #define TRACE(c, fmt, ...)     do { } while (0)
> > > > >  #endif
> > > > >
> > > > > +#define I2C_TEN_BIT_ADDRESS    BIT(31)
> > > > > +#define I2C_OWN_SLAVE_ADDRESS  BIT(30)
> > > > > +
> > > > >  enum checkstatus {
> > > > >         UNCHECKED = 0,
> > > > >         PREREQ,
> > > > > @@ -1048,6 +1052,7 @@ static void check_i2c_bus_reg(struct check *c, struct dt_info *dti, struct node
> > > > >
> > > > >         for (len = prop->val.len; len > 0; len -= 4) {
> > > > >                 reg = fdt32_to_cpu(*(cells++));
> > > > > +               reg &= ~(I2C_OWN_SLAVE_ADDRESS | I2C_TEN_BIT_ADDRESS);
> > > >
> > > > I'd just mask the top byte so we don't have to update on the next flag we add.
> > > Do you mean something like this, shown below ?
> > > reg &= 0xFFFF_FC000;
> >
> > Yes, but as I said, the 'top byte', so 0xFF000000.
> Thank you, will do as per your suggestion and send v2.
Hi Rob,
I have pushed separate patch which is prepared against dtc master branch.
Please review.

> >
> > Rob