-<h4>commit</h4><p>commit <a href="http://git.etalabs.net/cgi-bin/gitweb.cgi?p=musl;a=commit;h=4da289e1b3a0d09048d9be1dfc3a7280531a1e48">4da289e1b3a0d09048d9be1dfc3a7280531a1e48</a><br>Date: Tue Sep 13 21:21:53 2011 -0400<br><br> fix return types for aio_read and aio_write again<br> <br> previous fix was backwards and propagated the wrong type rather than<br> the right one...
+<h4>commit</h4><p>commit <a href="http://git.etalabs.net/cgi-bin/gitweb.cgi?p=musl;a=commit;h=3f72cdac73030761120cf32aeef44e7d03e2f1fa">3f72cdac73030761120cf32aeef44e7d03e2f1fa</a><br>Date: Sun Sep 18 10:14:37 2011 -0400<br><br> overhaul clone syscall wrapping<br> <br> several things are changed. first, i have removed the old __uniclone<br> function signature and replaced it with the "standard" linux<br> __clone/clone signature. this was necessary to expose clone to<br> applications anyway, and it makes it easier to port __clone to new<br> archs, since it's now testable independently of pthread_create.<br> <br> secondly, i have removed all references to the ugly ldt descriptor<br> structure (i386 only) from the c code and pthread structure. in places<br> where it is needed, it is now created on the stack just when it's<br> needed, in assembly code. thus, the i386 __clone function takes the<br> desired thread pointer as its argument, rather than an ldt descriptor<br> pointer, just like on all other sane archs. this should not affect<br> applications since there is really no way an application can use clone<br> with threads/tls in a way that doesn't horribly conflict with and<br> clobber the underlying implementation's use. applications are expected<br> to use clone only for creating actual processes, possibly with new<br> namespace features and whatnot.