From mboxrd@z Thu Jan 1 00:00:00 1970 From: James Hogan Subject: Re: pt_regs leak into userspace (was Re: [PATCH v3 20/71] ARC: Signal handling) Date: Mon, 11 Feb 2013 11:01:47 +0000 Message-ID: <5118CF9B.60102@imgtec.com> References: <1359024639-21915-1-git-send-email-vgupta@synopsys.com> <1359024639-21915-9-git-send-email-vgupta@synopsys.com> <51189D2F.4030000@synopsys.com> <5118C458.2070205@synopsys.com> <5118C7B6.6040807@imgtec.com> Mime-Version: 1.0 Content-Type: text/plain; charset="UTF-8" Content-Transfer-Encoding: 7bit Return-path: Received: from multi.imgtec.com ([194.200.65.239]:29391 "EHLO multi.imgtec.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1754477Ab3BKLCC (ORCPT ); Mon, 11 Feb 2013 06:02:02 -0500 In-Reply-To: Sender: linux-arch-owner@vger.kernel.org List-ID: To: Jonas Bonn Cc: Vineet Gupta , Arnd Bergmann , "linux-arch@vger.kernel.org" , "linux-kernel@vger.kernel.org" , Al Viro Hi Jonas, On 11/02/13 10:53, Jonas Bonn wrote: > And now that I think about it some more, I think this is done > incorrectly in the openrisc arch, too, as the fast-path for > rt_sigreturn probably only restores the call-clobbered regs. > sigreturn probably needs to be special-cased to _always_ restore all > the regs on its way back to userspace. Not for the "signal" case, but > for the "setcontext" case; but these two are pretty-much > indistinguishable. Wouldn't setcontext restore a different stack, so the sigreturn wouldn't happen until the ucontext is switched back? Cheers James