From mboxrd@z Thu Jan 1 00:00:00 1970 Date: Tue, 5 Feb 2019 14:29:53 +0100 From: Peter Zijlstra Subject: Re: [PATCH v2 06/20] x86/alternative: use temporary mm for text poking Message-ID: <20190205132953.GF17582@hirez.programming.kicks-ass.net> References: <20190129003422.9328-1-rick.p.edgecombe@intel.com> <20190129003422.9328-7-rick.p.edgecombe@intel.com> <20190205095853.GJ21801@zn.tnic> <20190205113146.GP17528@hirez.programming.kicks-ass.net> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <20190205113146.GP17528@hirez.programming.kicks-ass.net> To: Borislav Petkov Cc: Rick Edgecombe , Andy Lutomirski , Ingo Molnar , linux-kernel@vger.kernel.org, x86@kernel.org, hpa@zytor.com, Thomas Gleixner , Nadav Amit , Dave Hansen , linux_dti@icloud.com, linux-integrity@vger.kernel.org, linux-security-module@vger.kernel.org, akpm@linux-foundation.org, kernel-hardening@lists.openwall.com, linux-mm@kvack.org, will.deacon@arm.com, ard.biesheuvel@linaro.org, kristen@linux.intel.com, deneen.t.dock@intel.com, Nadav Amit , Kees Cook , Dave Hansen , Masami Hiramatsu List-ID: On Tue, Feb 05, 2019 at 12:31:46PM +0100, Peter Zijlstra wrote: > In general, text_poke() cannot fail: > > - suppose changing a single jump label requires poking multiple sites > (not uncommon), we fail halfway through and then have to undo the > first pokes, but those pokes fail again. > > - this then leaves us no way forward and no way back, we've got > inconsistent text state -> FAIL. Note that this exact fail scenario still exists in the CPU hotplug code. See kernel/cpu.c:cpuhp_thread_fun(): /* * If we fail on a rollback, we're up a creek without no * paddle, no way forward, no way back. We loose, thanks for * playing. */ WARN_ON_ONCE(st->rollback);