Re: [PATCH] Fix the man page

[Alejandro Colomar <alx.manpages@gmail.com>]

[-- Attachment #1.1: Type: text/plain, Size: 7521 bytes --]

On 5/20/23 00:21, enh wrote:
> On Fri, May 19, 2023 at 1:03 AM Alejandro Colomar
> <alx.manpages@gmail.com> wrote:
[...]

>> You'll also need to add _Nullable to the function prototypes in the
>> SYNOPSIS.
> 
> heh, funnily enough, zijunzhao's adding exactly those annotations to
> the bionic headers, which is how we stumbled across this.

:)

> 
>> BTW, I see that glibc still requires nonnull in gettimeofday(3).  It's
>> only settimeofday(3) that is nullable.
> 
> yeah, but the _kernel_ allows null in both (see kernel/time/time.c),
> and this is section 2 of the man page. it's unclear to me whether
> that's worth calling out up here, or should be down in the
> Linux-specific section? (this page seems to be written as if it cares
> about Hurd? "On a non-Linux kernel, with glibc ...".)

Yes, we care about hurd, and non-g libc, in this page and others.
Normally, I would be cautious, and specify the strictest prototype of
all implementations about which I care, but in this case, glibc seems
unnecessarily restrictive, and I'm ready to add _Nullable if the kernel
and some libc support that.  However, as a caution, please add a
(CAVEATS?) section documenting that glibc doesn't allow NULL there.
Or maybe keep the SYNOPSIS without _Nullable, and have some VERSIONS
section specifying that other libcs and the kernel support NULL.

> 
>> See:
>>
>>
>> $ grepc gettimeofday /usr/include/*/sys/time.h
>> /usr/include/x86_64-linux-gnu/sys/time.h:67:
>> extern int gettimeofday (struct timeval *__restrict __tv,
>>                          void *__restrict __tz) __THROW __nonnull ((1));
>>
>>
>> /usr/include/x86_64-linux-gnu/sys/time.h:75:
>> #  define gettimeofday __gettimeofday64
>>
>>
>> $ grepc settimeofday /usr/include/*/sys/time.h
>> /usr/include/x86_64-linux-gnu/sys/time.h:86:
>> extern int settimeofday (const struct timeval *__tv,
>>                          const struct timezone *__tz)
>>       __THROW;
>>
>>
>> /usr/include/x86_64-linux-gnu/sys/time.h:106:
>> #   define settimeofday __settimeofday64
>>
>>
>> And while NULL may be non-UB, the manual page is not very clear on why
>> someone would want to call these functions with NULL.  Could you
>> please also explain why would someone want to call these functions
>> with NULL?  (Let's discuss it in the list, and then we see what
>> wording we use for the page.)
> 
> so there are definitely several orders of magnitude more users of a
> null timezone --- most callers don't want that (not least because
> glibc seems to not fill it out anyway? musl certainly doesn't,
> implementing gettimeofday() in terms of clock_gettime() instead).
> 
> the users for a null time (and non-null timezone; i haven't seen
> anyone -- not even a test -- pass null for both!) are interesting.
> Android's init uses settimeofday() with only a timezone, but --
> although the timezone to use is settable in device-specific init
> scripts -- for devices in our tree, that's actually always 0. there's
> other proprietary (non-Google) code that uses gettimeofday() with only
> a timezone, and i'm honestly not sure whether it's deliberate that
> it's asking the _kernel_ for the timezone or not (i'm guessing they
> want the device timezone rather than the app timezone, but i do not
> believe there are devices where the _kernel_ timezone matches the
> device timezone?). certainly both of these use cases seem worth
> following up on, and i will do.

Please.  And thanks for the details!

> 
> but looking at musl (which ignores tz in both cases) it seems like the
> most impactful change would be to fix "The functions gettimeofday()
> and settimeofday() can get and set the time as well as a timezone" :-)

:-)

Makes sense; although it would be worth being clear about if that's
really true, and if libcs do support that as well as the kernel.

> 
> it also seems like an improvement to really call out the libc/kernel
> differences more clearly though.

Sure.  Many pages have a dedicated subsection.  Feel free to add it
to this page.  I'd put it in .SH VERSIONS.

$ grep -rn 'C library/kernel differences' man*
man2/sigprocmask.2:120:.SS C library/kernel differences
man2/setreuid.2:168:.SS C library/kernel differences
man2/brk.2:126:.SS C library/kernel differences
man2/sigwaitinfo.2:114:.SS C library/kernel differences
man2/gethostname.2:122:.SS C library/kernel differences
man2/getcpu.2:73:.SS C library/kernel differences
man2/fork.2:263:.SS C library/kernel differences
man2/epoll_wait.2:258:.SS C library/kernel differences
man2/wait4.2:157:.SS C library/kernel differences
man2/stat.2:355:.SS C library/kernel differences
man2/sigsuspend.2:74:.SS C library/kernel differences
man2/access.2:40:                /* But see C library/kernel differences, below */
man2/access.2:285:.SS C library/kernel differences
man2/clock_getres.2:354:.SS C library/kernel differences
man2/getpid.2:57:.SS C library/kernel differences
man2/setfsgid.2:57:.SS C library/kernel differences
man2/getgroups.2:142:.SS C library/kernel differences
man2/sigaction.2:946:.SS C library/kernel differences
man2/setresuid.2:100:.SS C library/kernel differences
man2/readv.2:290:.SS C library/kernel differences
man2/readv.2:342:.SS Historical C library/kernel differences
man2/open.2:1393:.SS C library/kernel differences
man2/setgid.2:58:.SS C library/kernel differences
man2/setfsuid.2:97:.SS C library/kernel differences
man2/ptrace.2:2814:.SS C library/kernel differences
man2/poll.2:382:.SS C library/kernel differences
man2/sigreturn.2:130:.SS C library/kernel differences
man2/eventfd.2:267:.SS C library/kernel differences
man2/mmap.2:729:.SS C library/kernel differences
man2/_exit.2:104:.SS C library/kernel differences
man2/time.2:86:.SS C library/kernel differences
man2/pread.2:101:.SS C library/kernel differences
man2/seccomp.2:975:.IR "C library/kernel differences" .
man2/setuid.2:104:.SS C library/kernel differences
man2/sched_setaffinity.2:95:return 0 (but see "C library/kernel differences" below,
man2/sched_setaffinity.2:220:.SS C library/kernel differences
man2/select.2:559:.SS C library/kernel differences
man2/sigpending.2:52:.SS C library/kernel differences
man2/posix_fadvise.2:135:.SS C library/kernel differences
man2/gettimeofday.2:173:.SS C library/kernel differences
man2/uname.2:83:.SS C library/kernel differences
man2/clone.2:1576:.SS C library/kernel differences
man2/getpriority.2:185:.SS C library/kernel differences
man2/signalfd.2:331:.SS C library/kernel differences
man2/timer_create.2:225:.SS C library/kernel differences
man2/wait.2:464:.SS C library/kernel differences
man2/nice.2:80:.SS C library/kernel differences
man2/chmod.2:303:.SS C library/kernel differences
man2/seteuid.2:114:.SS C library/kernel differences
man3/getcwd.3:217:.SS C library/kernel differences
man3/killpg.3:98:.SS C library/kernel differences
man3/mq_notify.3:174:.SS C library/kernel differences
man3/mq_open.3:269:.SS C library/kernel differences
man3/sigqueue.3:118:.SS C library/kernel differences
man3type/epoll_event.3type:37:.SS C library/kernel differences
man7/man-pages.7:402:.I "C library/kernel differences"

Cheers,
Alex

> 
>> Thanks!
>> Alex
>>
>> --
>> <http://www.alejandro-colomar.es/>
>> GPG key fingerprint: A9348594CE31283A826FBDD8D57633D441E25BB5
>>

-- 
<http://www.alejandro-colomar.es/>
GPG key fingerprint: A9348594CE31283A826FBDD8D57633D441E25BB5

[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 833 bytes --]