From mboxrd@z Thu Jan 1 00:00:00 1970 From: ebiederm@xmission.com (Eric W. Biederman) Subject: Re: [REVIEW][PATCH 03/26] signal/arm64: Use force_sig not force_sig_fault for SIGKILL Date: Thu, 23 May 2019 09:59:20 -0500 Message-ID: <87d0k9gqt3.fsf@xmission.com> References: <20190523003916.20726-1-ebiederm@xmission.com> <20190523003916.20726-4-ebiederm@xmission.com> <20190523101702.GG26646@fuggles.cambridge.arm.com> Mime-Version: 1.0 Content-Type: text/plain Return-path: In-Reply-To: <20190523101702.GG26646@fuggles.cambridge.arm.com> (Will Deacon's message of "Thu, 23 May 2019 11:17:02 +0100") Sender: linux-kernel-owner@vger.kernel.org To: Will Deacon Cc: linux-kernel@vger.kernel.org, Linux Containers , Oleg Nesterov , linux-arch@vger.kernel.org, Dave Martin , James Morse List-Id: linux-arch.vger.kernel.org Will Deacon writes: > On Wed, May 22, 2019 at 07:38:53PM -0500, Eric W. Biederman wrote: >> It really only matters to debuggers but the SIGKILL does not have any >> si_codes that use the fault member of the siginfo union. Correct this >> the simple way and call force_sig instead of force_sig_fault when the >> signal is SIGKILL. >> >> Cc: stable@vger.kernel.org >> Cc: Dave Martin >> Cc: James Morse >> Cc: Will Deacon >> Fixes: af40ff687bc9 ("arm64: signal: Ensure si_code is valid for all fault signals") >> Signed-off-by: "Eric W. Biederman" >> --- >> arch/arm64/kernel/traps.c | 5 +++++ >> 1 file changed, 5 insertions(+) >> >> diff --git a/arch/arm64/kernel/traps.c b/arch/arm64/kernel/traps.c >> index ade32046f3fe..0feb17bdcaa0 100644 >> --- a/arch/arm64/kernel/traps.c >> +++ b/arch/arm64/kernel/traps.c >> @@ -282,6 +282,11 @@ void arm64_notify_die(const char *str, struct pt_regs *regs, >> current->thread.fault_address = 0; >> current->thread.fault_code = err; >> >> + if (signo == SIGKILL) { >> + arm64_show_signal(signo, str); >> + force_sig(signo, current); >> + return; >> + } > > I know it's a bit of a misnomer, but I'd rather do this check inside > arm64_force_sig_fault, since I think we have other callers (e.g. > do_bad_area()) which also blindly pass in SIGKILL here. Sigh. You are right. I thought I had checked for that when I made my change there. But do_bad_area will definitely do that, and that was one of the cases that jumped out at me as needing to be fixed, when I skimmed the arm code. I will respin this patch to move that lower. > We could rename the thing if necessary. I would not mind but as long as we aren't misusing the generic bits I won't have alarm bells going of in my head when I look at their users. Eric From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from out02.mta.xmission.com ([166.70.13.232]:52242 "EHLO out02.mta.xmission.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1730719AbfEWO7a (ORCPT ); Thu, 23 May 2019 10:59:30 -0400 From: ebiederm@xmission.com (Eric W. Biederman) References: <20190523003916.20726-1-ebiederm@xmission.com> <20190523003916.20726-4-ebiederm@xmission.com> <20190523101702.GG26646@fuggles.cambridge.arm.com> Date: Thu, 23 May 2019 09:59:20 -0500 In-Reply-To: <20190523101702.GG26646@fuggles.cambridge.arm.com> (Will Deacon's message of "Thu, 23 May 2019 11:17:02 +0100") Message-ID: <87d0k9gqt3.fsf@xmission.com> MIME-Version: 1.0 Content-Type: text/plain Subject: Re: [REVIEW][PATCH 03/26] signal/arm64: Use force_sig not force_sig_fault for SIGKILL Sender: linux-arch-owner@vger.kernel.org List-ID: To: Will Deacon Cc: linux-kernel@vger.kernel.org, Linux Containers , Oleg Nesterov , linux-arch@vger.kernel.org, Dave Martin , James Morse Message-ID: <20190523145920.q3rGPJEu6xMpPJNw46o2rezU9lEjs8jHJ-E6LiyRePc@z> Will Deacon writes: > On Wed, May 22, 2019 at 07:38:53PM -0500, Eric W. Biederman wrote: >> It really only matters to debuggers but the SIGKILL does not have any >> si_codes that use the fault member of the siginfo union. Correct this >> the simple way and call force_sig instead of force_sig_fault when the >> signal is SIGKILL. >> >> Cc: stable@vger.kernel.org >> Cc: Dave Martin >> Cc: James Morse >> Cc: Will Deacon >> Fixes: af40ff687bc9 ("arm64: signal: Ensure si_code is valid for all fault signals") >> Signed-off-by: "Eric W. Biederman" >> --- >> arch/arm64/kernel/traps.c | 5 +++++ >> 1 file changed, 5 insertions(+) >> >> diff --git a/arch/arm64/kernel/traps.c b/arch/arm64/kernel/traps.c >> index ade32046f3fe..0feb17bdcaa0 100644 >> --- a/arch/arm64/kernel/traps.c >> +++ b/arch/arm64/kernel/traps.c >> @@ -282,6 +282,11 @@ void arm64_notify_die(const char *str, struct pt_regs *regs, >> current->thread.fault_address = 0; >> current->thread.fault_code = err; >> >> + if (signo == SIGKILL) { >> + arm64_show_signal(signo, str); >> + force_sig(signo, current); >> + return; >> + } > > I know it's a bit of a misnomer, but I'd rather do this check inside > arm64_force_sig_fault, since I think we have other callers (e.g. > do_bad_area()) which also blindly pass in SIGKILL here. Sigh. You are right. I thought I had checked for that when I made my change there. But do_bad_area will definitely do that, and that was one of the cases that jumped out at me as needing to be fixed, when I skimmed the arm code. I will respin this patch to move that lower. > We could rename the thing if necessary. I would not mind but as long as we aren't misusing the generic bits I won't have alarm bells going of in my head when I look at their users. Eric