Current-Users archive

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index][Old Index]

Re: HEADS UP: I will be merging christos-time_t by the end of theweek

On Jan 14,  4:27pm, (Matthias Drochner) 
-- Subject: Re: HEADS UP: I will be merging christos-time_t by the end of the

| said:
| > What conclusions did we come to regarding PR 39215? I got behind on
| > the discussion there at the time and haven't caught up on it yet. 
| It has taken a while for me too to catch up -- I've just dug
| out my old patches introducing devmajor_t/devminor_t and
| built a -current kernel.
| I'll append just the basic kernel patches, didn't look at
| the rest yet. In any case there are no binary compatibility
| issues.
| > I think your(?) argument that minor numbers are used for array
| > indexing and whatnot and shouldn't be bigger than machine words was
| > fairly persuasive.
| I had to notice that something leaked into device drivers which
| shouldn't: Many drivers use macros like
| #define XXXUNIT(dev) (minor(dev) & 0x0f)
| This even lead to printf format changes to print out 64-bit
| device unit numbers. With minor() returning an "int" as before
| this wouldn't happen, and we'd avoid mixing types at other
| places.
| I'd say we should at least change major/minor to return ints
| again. Whether new names are introduced is a matter of
| taste, imo it helps clarity.

I think it is best to ask in tech-kern. Now is the good time to do
it, because we can change the format of the dev_t without killing
binary compatibility (because all the dev_t bearing syscalls have
been versioned and we can ride on the version bump). For example
I think that we can use:

/* Major, minor numbers, dev_t's. */
#define omajor(x)        ((uint32_t)((((x) & 0x000fff00) >>  8)))
#define ominor(x)        ((uint32_t)((((x) & 0xfff00000) >> 12) | \  
                                   (((x) & 0x000000ff) >>  0)))
#define omakedev(x,y)    ((dev_t)((((x) <<  8) & 0x000fff00) | \
                                 (((y) << 12) & 0xfff00000) | \
                                 (((y) <<  0) & 0x000000ff)))

#define major(x)        (uint32_t)(((x) >> 32) & 0xffffffffULL)
#define minor(x)        (uint32_t)(((x) >>  0) & 0xffffffffULL)
#define makedev(x,y)    (((dev_t)(x) << 32) | (y))

and then to convert from old style dev to new we can use:

        stat_to_stat50(st50, st) {
                st50->st_dev = omakedev(major(st->st_dev), minor(st->st_dev));
        stat50_to_stat(st, st50) {
                st->st_dev = makedev(omajor(st50->st_dev), 

Same goes for old on disk formats. We check the version of the fs,
and then we do the conversion depending on the version. I think it
makes sense to keep major and minor 32 bits, I don't see a reason
to split otherwise, like major be 24 and minor be 40.


Home | Main Index | Thread Index | Old Index