On 2/13/21 5:05 PM, Michael Kerrisk (man-pages) wrote:
> Hi Alex,
>
> On 2/12/21 4:28 PM, Alejandro Colomar (man-pages) wrote:
>> Hi Michael,
>>
>> AFAICT, the syscalls [rt_]sigprocmask use '[old_]sigset_t' (there's the
>> '__user' modifier too), so 'kernel_' should be removed from the types,
>> shouldn't it?
>
> Looks like you're right. Will you send a patch?
>
> Thanks,
>
> Michael
Hi Michael,
I was wrong!
Thanks,
Alex
[[
C library/kernel differences
The kernel's definition of sigset_t differs in size from
that used by the C library. In this manual page, the for‐
mer is referred to as kernel_sigset_t (it is nevertheless
named sigset_t in the kernel sources).
The glibc wrapper function for sigprocmask() silently ig‐
nores attempts to block the two real-time signals that are
used internally by the NPTL threading implementation. See
nptl(7) for details.
The original Linux system call was named sigprocmask().
However, with the addition of real-time signals in Linux
2.2, the fixed-size, 32-bit sigset_t (referred to as
old_kernel_sigset_t in this manual page) type supported by
that system call was no longer fit for purpose. Conse‐
quently, a new system call, rt_sigprocmask(), was added to
support an enlarged sigset_t type (referred to as ker‐
nel_sigset_t in this manual page). The new system call
takes a fourth argument, size_t sigsetsize, which specifies
the size in bytes of the signal sets in set and oldset.
This argument is currently required to have a fixed archi‐
tecture specific value (equal to sizeof(kernel_sigset_t)).
The glibc sigprocmask() wrapper function hides these de‐
tails from us, transparently calling rt_sigprocmask() when
the kernel provides it.
]]
--
Alejandro Colomar
Linux man-pages comaintainer; https://www.kernel.org/doc/man-pages/
http://www.alejandro-colomar.es/
