system calls not intended for user space

system calls not intended for user space

[Alejandro Colomar (man-pages)]

Hello Michael,

There are some syscalls that are not intended to be used from user 
space.  See restart_syscall(2).  Should I document those as 
syscall(SYS_...) or not?  How are they called within the kernel?  As 
sys_restart_syscall()?

Thanks,

Alex


-- 
Alejandro Colomar
Linux man-pages comaintainer; https://www.kernel.org/doc/man-pages/
http://www.alejandro-colomar.es/

Re: system calls not intended for user space

[Michael Kerrisk (man-pages)]

Hi Alex,

On Fri, 14 May 2021 at 07:25, Alejandro Colomar (man-pages)
<alx.manpages@gmail.com> wrote:
>
> Hello Michael,
>
> There are some syscalls that are not intended to be used from user
> space.  See restart_syscall(2).  Should I document those as
> syscall(SYS_...) or not?

Yes, this function is an odd case. (Are there others? Perhaps
sigreturn(2) also.)

I think I would just ignore these two. (That is, no changes.)

> How are they called within the kernel?  As
> sys_restart_syscall()?

I'm not sure of the details, I'm sorry.

Thanks,

Michael

Re: system calls not intended for user space

[Alejandro Colomar (man-pages)]

Hi Michael,

On 5/17/21 3:04 AM, Michael Kerrisk (man-pages) wrote:
> Hi Alex,
> 
> On Fri, 14 May 2021 at 07:25, Alejandro Colomar (man-pages)
> <alx.manpages@gmail.com> wrote:
>>
>> Hello Michael,
>>
>> There are some syscalls that are not intended to be used from user
>> space.  See restart_syscall(2).  Should I document those as
>> syscall(SYS_...) or not?
> 
> Yes, this function is an odd case. (Are there others? Perhaps
> sigreturn(2) also.)

rt_sigqueueinfo.2 maybe, for which, by the way, I sent a patch recently, 
so you may prefer to ingore it.

> 
> I think I would just ignore these two. (That is, no changes.)

Agree.  I sent the patch for rt_sigqueueinfo.2 because in NOTES it 
already talked about syscall(2), and by the DESCRIPTION it looks like 
it's used by library writers (that's user space).

Thanks,

Alex

-- 
Alejandro Colomar
Linux man-pages comaintainer; https://www.kernel.org/doc/man-pages/
http://www.alejandro-colomar.es/

Re: system calls not intended for user space

[Alejandro Colomar (man-pages)]

Hi Michael,

On 5/17/21 9:47 AM, Alejandro Colomar (man-pages) wrote:
> Hi Michael,
> 
> On 5/17/21 3:04 AM, Michael Kerrisk (man-pages) wrote:
>> Hi Alex,
>>
>> On Fri, 14 May 2021 at 07:25, Alejandro Colomar (man-pages)
>> <alx.manpages@gmail.com> wrote:
>>>
>>> Hello Michael,
>>>
>>> There are some syscalls that are not intended to be used from user
>>> space.  See restart_syscall(2).  Should I document those as
>>> syscall(SYS_...) or not?
>>
>> Yes, this function is an odd case. (Are there others? Perhaps
>> sigreturn(2) also.)
> 
> rt_sigqueueinfo.2 maybe, for which, by the way, I sent a patch recently, 
> so you may prefer to ingore it.
> 
>>
>> I think I would just ignore these two. (That is, no changes.)
> 
> Agree.  I sent the patch for rt_sigqueueinfo.2 because in NOTES it 
> already talked about syscall(2), and by the DESCRIPTION it looks like 
> it's used by library writers (that's user space).


I see you've applied the patch (it was 2/8 that you applied today 
(<https://lore.kernel.org/linux-man/20210515182027.186403-2-alx.manpages@gmail.com/T/#u>)). 
  Check this before pushing and decide :)

Cheers,

Alex


-- 
Alejandro Colomar
Linux man-pages comaintainer; https://www.kernel.org/doc/man-pages/
http://www.alejandro-colomar.es/

Re: system calls not intended for user space

[Michael Kerrisk (man-pages)]

Hi Alex,

On Mon, 17 May 2021 at 19:51, Alejandro Colomar (man-pages)
<alx.manpages@gmail.com> wrote:
>
> Hi Michael,
>
> On 5/17/21 9:47 AM, Alejandro Colomar (man-pages) wrote:
> > Hi Michael,
> >
> > On 5/17/21 3:04 AM, Michael Kerrisk (man-pages) wrote:
> >> Hi Alex,
> >>
> >> On Fri, 14 May 2021 at 07:25, Alejandro Colomar (man-pages)
> >> <alx.manpages@gmail.com> wrote:
> >>>
> >>> Hello Michael,
> >>>
> >>> There are some syscalls that are not intended to be used from user
> >>> space.  See restart_syscall(2).  Should I document those as
> >>> syscall(SYS_...) or not?
> >>
> >> Yes, this function is an odd case. (Are there others? Perhaps
> >> sigreturn(2) also.)
> >
> > rt_sigqueueinfo.2 maybe, for which, by the way, I sent a patch recently,
> > so you may prefer to ingore it.
> >
> >>
> >> I think I would just ignore these two. (That is, no changes.)
> >
> > Agree.  I sent the patch for rt_sigqueueinfo.2 because in NOTES it
> > already talked about syscall(2), and by the DESCRIPTION it looks like
> > it's used by library writers (that's user space).
>
>
> I see you've applied the patch (it was 2/8 that you applied today
> (<https://lore.kernel.org/linux-man/20210515182027.186403-2-alx.manpages@gmail.com/T/#u>)).
>   Check this before pushing and decide :)

D'oh! Thanks for catching that. In the meantime, I will drop that patch.

Cheers,

Michael
-- 
Michael Kerrisk
Linux man-pages maintainer; http://www.kernel.org/doc/man-pages/
Linux/UNIX System Programming Training: http://man7.org/training/