From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from foss.arm.com (usa-sjc-mx-foss1.foss.arm.com [217.140.101.70]) by lists.ozlabs.org (Postfix) with ESMTP id 40N6Cr6Sr3zF1SH for ; Sat, 14 Apr 2018 04:45:28 +1000 (AEST) Date: Fri, 13 Apr 2018 19:45:23 +0100 From: Dave Martin To: Linus Torvalds Cc: Russell King - ARM Linux , Linux Kernel Mailing List , "Dmitry V. Levin" , "Eric W. Biederman" , sparclinux , ppc-dev , linux-arm-kernel Subject: Re: sparc/ppc/arm compat siginfo ABI regressions: sending SIGFPE via kill() returns wrong values in si_pid and si_uid Message-ID: <20180413184522.GD16308@e103592.cambridge.arm.com> References: <20180412124928.GA29458@altlinux.org> <20180412131404.GE16141@n2100.armlinux.org.uk> <20180412172051.GK16141@n2100.armlinux.org.uk> <20180413094211.GN16141@n2100.armlinux.org.uk> <20180413170827.GB16308@e103592.cambridge.arm.com> <20180413175407.GO16141@n2100.armlinux.org.uk> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii In-Reply-To: List-Id: Linux on PowerPC Developers Mail List List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , On Fri, Apr 13, 2018 at 11:23:36AM -0700, Linus Torvalds wrote: > On Fri, Apr 13, 2018 at 10:54 AM, Russell King - ARM Linux > wrote: > > > > FPE_FLTINV means "floating point invalid operation". Does it really > > cover the case where hardware has failed, or is it intended to cover > > the case where userspace did something wrong and asked for an invalid > > operation from the FP hardware? > > Note that the number of people who actually look at the si_code is > approximately zero. > > But the ones that _do_ check the si_code are certainly not going to > check it against a new code that they don't know about. > > I suspect that if you start searching for FLT_xyz occurrences in code, > approximately 100% of them are from the kernel code that generates > them, not from any actual users. > > So I'd be very surprised if you can find *anybody* who cares about > that exact value (with the possible exceptions of test-suites). > > Sadly, google code-search is no more. It was useful for things like that. I've found https://codesearch.debian.net/ useful for digging into this kind of question, though it tends to throw up a lot of false positives. Most uses I've seen do nothing more than use the FPE_xyz value to format diagnostic messages while dying. I struggled to find code that made a meaningful functional decision based on the value, though that's not proof... Cheers ---Dave From mboxrd@z Thu Jan 1 00:00:00 1970 From: Dave Martin Date: Fri, 13 Apr 2018 18:45:23 +0000 Subject: Re: sparc/ppc/arm compat siginfo ABI regressions: sending SIGFPE via kill() returns wrong values in Message-Id: <20180413184522.GD16308@e103592.cambridge.arm.com> List-Id: References: <20180412124928.GA29458@altlinux.org> <20180412131404.GE16141@n2100.armlinux.org.uk> <20180412172051.GK16141@n2100.armlinux.org.uk> <20180413094211.GN16141@n2100.armlinux.org.uk> <20180413170827.GB16308@e103592.cambridge.arm.com> <20180413175407.GO16141@n2100.armlinux.org.uk> In-Reply-To: MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit To: linux-arm-kernel@lists.infradead.org On Fri, Apr 13, 2018 at 11:23:36AM -0700, Linus Torvalds wrote: > On Fri, Apr 13, 2018 at 10:54 AM, Russell King - ARM Linux > wrote: > > > > FPE_FLTINV means "floating point invalid operation". Does it really > > cover the case where hardware has failed, or is it intended to cover > > the case where userspace did something wrong and asked for an invalid > > operation from the FP hardware? > > Note that the number of people who actually look at the si_code is > approximately zero. > > But the ones that _do_ check the si_code are certainly not going to > check it against a new code that they don't know about. > > I suspect that if you start searching for FLT_xyz occurrences in code, > approximately 100% of them are from the kernel code that generates > them, not from any actual users. > > So I'd be very surprised if you can find *anybody* who cares about > that exact value (with the possible exceptions of test-suites). > > Sadly, google code-search is no more. It was useful for things like that. I've found https://codesearch.debian.net/ useful for digging into this kind of question, though it tends to throw up a lot of false positives. Most uses I've seen do nothing more than use the FPE_xyz value to format diagnostic messages while dying. I struggled to find code that made a meaningful functional decision based on the value, though that's not proof... Cheers ---Dave From mboxrd@z Thu Jan 1 00:00:00 1970 From: Dave.Martin@arm.com (Dave Martin) Date: Fri, 13 Apr 2018 19:45:23 +0100 Subject: sparc/ppc/arm compat siginfo ABI regressions: sending SIGFPE via kill() returns wrong values in si_pid and si_uid In-Reply-To: References: <20180412124928.GA29458@altlinux.org> <20180412131404.GE16141@n2100.armlinux.org.uk> <20180412172051.GK16141@n2100.armlinux.org.uk> <20180413094211.GN16141@n2100.armlinux.org.uk> <20180413170827.GB16308@e103592.cambridge.arm.com> <20180413175407.GO16141@n2100.armlinux.org.uk> Message-ID: <20180413184522.GD16308@e103592.cambridge.arm.com> To: linux-arm-kernel@lists.infradead.org List-Id: linux-arm-kernel.lists.infradead.org On Fri, Apr 13, 2018 at 11:23:36AM -0700, Linus Torvalds wrote: > On Fri, Apr 13, 2018 at 10:54 AM, Russell King - ARM Linux > wrote: > > > > FPE_FLTINV means "floating point invalid operation". Does it really > > cover the case where hardware has failed, or is it intended to cover > > the case where userspace did something wrong and asked for an invalid > > operation from the FP hardware? > > Note that the number of people who actually look at the si_code is > approximately zero. > > But the ones that _do_ check the si_code are certainly not going to > check it against a new code that they don't know about. > > I suspect that if you start searching for FLT_xyz occurrences in code, > approximately 100% of them are from the kernel code that generates > them, not from any actual users. > > So I'd be very surprised if you can find *anybody* who cares about > that exact value (with the possible exceptions of test-suites). > > Sadly, google code-search is no more. It was useful for things like that. I've found https://codesearch.debian.net/ useful for digging into this kind of question, though it tends to throw up a lot of false positives. Most uses I've seen do nothing more than use the FPE_xyz value to format diagnostic messages while dying. I struggled to find code that made a meaningful functional decision based on the value, though that's not proof... Cheers ---Dave