Messages in this thread | | | From | "Michael Kerrisk (man-pages)" <> | Date | Thu, 23 Apr 2015 13:38:54 +0200 | Subject | Re: [PATCH v16] sys_membarrier(): system-wide memory barrier (generic, x86) |
| |
Hi Mathieu,
On 22 April 2015 at 23:01, Mathieu Desnoyers <mathieu.desnoyers@efficios.com> wrote: > ----- Original Message ----- >> On 22 April 2015 at 21:43, Mathieu Desnoyers >> <mathieu.desnoyers@efficios.com> wrote: >> > ----- Original Message ----- >> >> On Fri, 2015-04-17 at 11:06 -0400, Mathieu Desnoyers wrote: >> >> > [ Not in this patch: tests (Pranith Kumar has a patch for this), man >> >> > page. >> >> > ] >> >> >> >> Actually, we had discussed that for new syscalls, the changelog _would_ >> >> be the manpage -- at least included in it. >> > >> > Do you want the formatted content, or just the text ? Putting >> > manpage formatting in the changelog seems a bit odd. >> >> >> Plain text works for me. Just divide it up into the usuual man-pages >> sections (see man-pages(7), and I'll take care of the rest). > > OK. Should I do the writeup as a man(2) from the POV of a glibc > user ? It's a bit different when we get to return values and > error codes if we document kernel<->glibc or glibc<->user.
Ideally, from glibc POV, since that is the POV od user space consumers.
> Moreover, how can we get this into glibc ?
An eternal question...
https://lwn.net/Articles/534682/
I added Carlos, who can probably best summarize the current state of things. I think there's a wiki discussion on this topic, but I could not find a URL.
Cheers,
Michael
| |