[PATCH] i2c_tools.c: add i2ctransfer utility

[Xabier Oneca -- xOneca]

Hi Nikolaus,

> >> +       while (*++argv) {
> >> +               char *arg_ptr = *argv;
> >> +               unsigned long len, raw_data;
> >> +               __u16 flags;
> >> +               __u8 data, *buf;
> >> +               char *end;
> >> +
> >> +               if (nmsgs > I2C_RDWR_IOCTL_MAX_MSGS)
> >> +                       bb_error_msg_and_die("Error: Too many messages (max: %d)\n",
> >> +                                            I2C_RDWR_IOCTL_MAX_MSGS);
> >
> > Off-by-one error? Seems you meant to use >=.
>
> The code is actually correct, as nmsgs is incremented below after
> writing to msgs[nmsgs] and nmsgs == I2C_RDWR_IOCTL_MAX_MSGS is valid.

Sorry, I don't see how can here nmsgs == I2C_RDWR_IOCTL_MAX_MSGS be
valid. nmsgs is the array index, so it can never reach
I2C_RDWR_IOCTL_MAX_MSGS.

> >> +                               } else {
> >> +                                       /* 'force' allows whole address range */
> >
> > Not sure about this statement. It seems to be a '-a' for this:
> > https://manpages.debian.org/unstable/i2c-tools/i2ctransfer.8.en.html#OPTIONS
> > It seems none of the I2C functions of Busybox but i2cdetect support -a.
>
> My version of i2ctransfer for i2c-tools does not (yet) support -a, see
> https://manpages.debian.org/testing/i2c-tools/i2ctransfer.8.en.html#OPTIONS
> but I think it is a good extension, so I added support for it.

Sorry, I didn't notice that I checked the man pages for unstable! :/

Only looking at the man, it is not clear what the behaviour should be
here. But if you checked that -f allows any address, then it's fine
for me.

Cheers,

Xabier Oneca_,,_