fcntl adding F_LOG2PHYS* flags.

[devnexen]

to get the current device address, in practice its current offset only.

[SteveLauC]

The F_LOG2PHYS looks good to me, I am confused about the meaning and usage of F_LOG2PHYS_EXT:

From man page

The F_LOG2PHYS_EXT command operates on the same structure as F_LOG2PHYS but treats it as an in/out:

        struct log2phys {
            u_int32_t l2p_flags;        /* unused so far */
            off_t     l2p_contigbytes;  /* IN: number of bytes to be queried;
                                           OUT: number of contiguous bytes allocated at this position */
            off_t     l2p_devoffset;    /* IN: bytes into file;
                                           OUT: bytes into device */
        };

1. The interface proposed in this PR only exposes that l2p_devoffset argument, but unlike F_LOG2PHYS, l2p_contigbytes is also used, should we expose it?

2. What does "difference with F_LOG2PHYS is the struct passed is used as both IN/OUT" mean? From the unit test, looks like it is similar to lseek(2), if so, why are we using a mutable reference to off_t here? Or it is possible that the returned offset could be different from the value passed in?

[devnexen]

to be honest F_LOGPHYS_EXT was kind of "late addition" I only noticed it just before I commited :)

The interface proposed in this PR only exposes that l2p_devoffset argument, but unlike F_LOG2PHYS, l2p_contigbytes is also used, should we expose it?

Yes. Better just exposing the struct altogether in this case.

What does "difference with F_LOG2PHYS is the struct passed is used as both IN/OUT" mean? From the unit test, looks like it is similar to lseek(2), if so, why are we using a mutable reference to off_t here? Or it is possible that the returned offset could be different from the value passed in?

I may need to rephrase it, was just explaining the comment above about those fields being both IN/OUT for F_LOG2PHYS_EXT. Value of l2p_devoffset can change thus why I add mut qualifier.

[SteveLauC]

to be honest F_LOGPHYS_EXT was kind of "late addition" I only noticed it just before I commited :)

Got it:)

Yes. Better just exposing the struct altogether in this case.

Yeah, that makes sense to me.

I may need to rephrase it,

Thanks!

Value of l2p_devoffset can change thus why I add mut qualifier.

Then what's the IN semantic here? i.e., the meaning of the value passed in🤔

[SteveLauC]

Do you mind telling me the functionality of that _EXT flag, I think I am still confused about it:<

[devnexen]

It is the same as LOG2PHYS only l2p_devoffset/l2p_contigbytes are set after call but also can be set before the call whereas LOG2PHYS set l2p_devoffset after the call.

[SteveLauC]

Though I am still unclear about this flag, but it should not matter as we expose everything needed in this interface, so merging:)

[SteveLauC]

Thanks!