Date: Fri, 19 Oct 2001 21:40:45 +0300 From: Ruslan Ermilov <ru@FreeBSD.org> To: Jason Evans <jasone@canonware.com> Cc: cvs-committers@FreeBSD.org, cvs-all@FreeBSD.org Subject: Re: cvs commit: src/lib/libc_r/uthread pthread_private.h Message-ID: <20011019214045.L24666@sunbay.com> In-Reply-To: <20011019105525.C53185@canonware.com>; from jasone@canonware.com on Fri, Oct 19, 2001 at 10:55:25AM -0700 References: <200110191722.f9JHM5O29047@freefall.freebsd.org> <20011019105525.C53185@canonware.com>
next in thread | previous in thread | raw e-mail | index | archive | help
On Fri, Oct 19, 2001 at 10:55:25AM -0700, Jason Evans wrote: > On Fri, Oct 19, 2001 at 10:22:04AM -0700, Ruslan Ermilov wrote: > > ru 2001/10/19 10:22:04 PDT > > > > Modified files: (Branch: RELENG_4) > > lib/libc_r/uthread pthread_private.h > > Log: > > Cleanup of _thread_sys* declarations: > > > > - Removed gratuitous declarations for non-syscalls. > > > > - Added declarations for aio_suspend(), kevent(), nanosleep(), rfork(), > > sched_yield(), sendfile(), sigblock(), sigsetmask(), and vfork(). > > > > - Fixed declarations for <sys/socket.h> syscalls (int -> socklen_t). > > > > - Style. > > Why weren't the changes that are equally applicable to -current made there > first, then MFCed? This looks to me like an obvious violation of stated > procedure. This surely would have been pointed out to you, had you > bothered to have the changes reviewed by me or Dan Eischen. > It doesn't equally applicable to -CURRENT actually, and I tried to get a review from Daniel for a related (unhide nfssvc(2)) change with no luck. In -CURRENT, we don't hide syscalls, but instead syscalls are available as __sys_*, and libc provides weak aliases for them which libc_r then overrides some of. We thus have a light-weighted version of libc_r in -CURRENT, as compared to -STABLE: -r--r--r-- 1 root wheel 1456412 Oct 19 20:06 /STABLE/usr/lib/libc_r.a -r--r--r-- 1 root wheel 252026 Oct 15 05:19 /CURRENT/usr/lib/libc_r.a In -STABLE, we rename syscalls that libc_r overrides to __sys_thread_*. So at least item 1 (the biggest) of this change isn't applicable to -CURRENT. Also, as we don't hide syscalls in -CURRENT, compiling a program statically with libc_r with either of nfssvc() or sigreturn() worked in -CURRENT, but was broken in -STABLE. But I agree that at least item 2 (<sys/socket.h> syscalls prototypes) applies to -CURRENT as well, and am sorry for not doing it there first. OTOH, in -CURRENT, real syscalls are avaibale as __sys_*, and the right thing to do would be to prototype only those syscalls that are used internally by libc_r. I'll prepare the patch on Monday for -CURRENT, and will send it to you for review. Cheers, -- Ruslan Ermilov Oracle Developer/DBA, ru@sunbay.com Sunbay Software AG, ru@FreeBSD.org FreeBSD committer, +380.652.512.251 Simferopol, Ukraine http://www.FreeBSD.org The Power To Serve http://www.oracle.com Enabling The Information Age To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe cvs-all" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20011019214045.L24666>