<html><head></head><body>I can only think of three ways to solve this:<br>1) Custom headers pretending that return type is int (will potentially break non-buggy devices tough).<br>2) Interposed wrappers for all affected calls.<br>3) Replacement (asm) implementation for all affected calls.<br><br>It can't be done in VLC build system as the problem affects contribs too.<br><br><div class="gmail_quote">Le 4 avril 2019 14:45:53 GMT+03:00, "Hugo Beauzée-Luyssen" <hugo@beauzee.fr> a écrit :<blockquote class="gmail_quote" style="margin: 0pt 0pt 0pt 0.8ex; border-left: 1px solid rgb(204, 204, 204); padding-left: 1ex;">
<pre class="k9mail">On Thu, Apr 4, 2019, at 1:42 PM, Rémi Denis-Courmont wrote:<br><blockquote class="gmail_quote" style="margin: 0pt 0pt 1ex 0.8ex; border-left: 1px solid #729fcf; padding-left: 1ex;">Hi all,<br><br>5 minutes git archeology through bionic repo leads to this:<br><a href="https://android.googlesource.com/platform/bionic.git/+/3d19a8319b9c27af8aa5cfbf495da0fe7fa62d3e">https://android.googlesource.com/platform/bionic.git/+/3d19a8319b9c27af8aa5cfbf495da0fe7fa62d3e</a><br><br>The bug affect basically all system call wrappers with a signed 64-bits <br>return value type (including ssize_t).<br><br>Clearly this patch ain't going to work.<br><br>Le 4 avril 2019 10:54:44 GMT+03:00, "Hugo Beauzée-Luyssen" <br><hugo@beauzee.fr> a écrit : On Wed, Apr 3, 2019, at 9:24 PM, Rémi <br>Denis-Courmont wrote:<br><blockquote class="gmail_quote" style="margin: 0pt 0pt 1ex 0.8ex; border-left: 1px solid #ad7fa8; padding-left: 1ex;"><blockquote class="gmail_quote" style="margin: 0pt 0pt 1ex 0.8ex; border-left: 1px solid #8ae234; padding-left: 1ex;">I don't know what you did or did not imply. When somebody states that a patch <br>is ugly, I tend to assume that they dislike the form rather than the essence.<br><br>My chief concerns are that, on the one hand, other system calls may also <br>exhibit the bug, and on the other hand, contribs may be affected as well. In <br>other words, I am concerned not that the patch is ugly, but that does not <br>solve the problem(s).<br></blockquote>It solves the observed problem, but it doesn't solve problems I don't know about indeed.<br></blockquote></blockquote><br>Hi,<br><br>Let aside the passive aggressiveness of the reply, this is exactly what I was hoping for by submitting the patch, that someone who knows libc/kernel more than I do figures out what the problem is.<br>I looked at the syscall wrapper but did not check the __set_errno it called, that's my bad.<br><br>Thanks for finding the actual error.<br></pre></blockquote></div><br>-- <br>Envoyé de mon appareil Android avec Courriel K-9 Mail. Veuillez excuser ma brièveté.</body></html>