public inbox for [email protected]
 help / color / mirror / Atom feed
* [RFC liburing for-2.3] Kill src/syscall.c
@ 2022-07-11  9:23 Ammar Faizi
  0 siblings, 0 replies; only message in thread
From: Ammar Faizi @ 2022-07-11  9:23 UTC (permalink / raw)
  To: Jens Axboe
  Cc: Hao Xu, Pavel Begunkov, Alviro Iskandar Setiawan,
	Fernanda Ma'rouf, GNU/Weeb Mailing List,
	io-uring Mailing List


Hi Jens,

I haven't written any patch, but I want to get your comments first.
If you think it's a good idea, I will wire up a patch.

## Background story:

Since we support nolibc, the file src/syscall.c is no longer used by
any liburing internal functions. It's only used by `test/`.

Other architectures that still depend on libc are now using:

     src/generic/syscall.h

*not*:

     src/syscall.c

I should have killed src/syscall.c and fixed the tests earlier when
adding nolibc support, unfortunately, I didn't. I was mistaken that
if we killed src/syscall.c, it might break user. But no, no user uses
it.

We don't actually intend to export __sys_io_uring* functions to user.
This situation is reflected in liburing.map and liburing.h which don't
have those functions.

## Proposal:

Let's just kill src/syscall.c and use __sys_io_uring* or
____sys_io_uring* only.


## Two choices:

1. Keep using ____sys_io_uring* functions and delete __sys_io_uring*
    functions.

2. Back to use __sys_io_uring* functions and delete ____sys_io_uring*.

Both choices need to fix tests. I can do that. I personally like the
second choice since we don't have many underscores.

What do you think?

-- 
Ammar Faizi


^ permalink raw reply	[flat|nested] only message in thread

only message in thread, other threads:[~2022-07-11  9:24 UTC | newest]

Thread overview: (only message) (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2022-07-11  9:23 [RFC liburing for-2.3] Kill src/syscall.c Ammar Faizi

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox