Received-SPF: pass (google.com: best guess record for domain of linux-kernel-owner@vger.kernel.org designates 209.132.180.67 as permitted sender) client-ip=209.132.180.67;
Content-Type: text/plain;
        charset=utf-8
Mime-Version: 1.0 (1.0)
Subject: Re: [PATCH v2 4/4] x86/static_call: Add inline static call implementation for x86-64
From:   Andy Lutomirski <luto@amacapital.net>
In-Reply-To: <CAHk-=wgoeXuodaDkKNneiQmU3mxkt2C1MeyPF5_0NX1VSQV=XQ@mail.gmail.com>
Date:   Thu, 29 Nov 2018 09:35:42 -0800
Cc:     Josh Poimboeuf <jpoimboe@redhat.com>,
        Peter Zijlstra <peterz@infradead.org>,
        Andrew Lutomirski <luto@kernel.org>,
        the arch/x86 maintainers <x86@kernel.org>,
        Linux List Kernel Mailing <linux-kernel@vger.kernel.org>,
        Ard Biesheuvel <ard.biesheuvel@linaro.org>,
        Steven Rostedt <rostedt@goodmis.org>,
        Ingo Molnar <mingo@kernel.org>,
        Thomas Gleixner <tglx@linutronix.de>, mhiramat@kernel.org,
        jbaron@akamai.com, Jiri Kosina <jkosina@suse.cz>,
        David.Laight@aculab.com, bp@alien8.de, julia@ni.com,
        jeyu@kernel.org, Peter Anvin <hpa@zytor.com>
Content-Transfer-Encoding: quoted-printable
Message-Id: <B36138FF-6BA5-4868-A321-2E31CA4875F4@amacapital.net>
References: <20181126160217.GR2113@hirez.programming.kicks-ass.net> <20181126171036.chcbmb35ygpxziub@treble> <20181126175624.bruqfbkngbucpvxr@treble> <20181126200801.GW2113@hirez.programming.kicks-ass.net> <20181126212628.4apztfazichxnt7r@treble> <20181127084330.GX2113@hirez.programming.kicks-ass.net> <CALCETrXam8JBtXE3S97Vyb_=+pmjgr23SoZ-Dm5pmYtwZYA+7Q@mail.gmail.com> <20181129094210.GC2131@hirez.programming.kicks-ass.net> <ECF632D9-581B-40B7-9F0A-6B5607A4D0E3@amacapital.net> <20181129143853.GO2131@hirez.programming.kicks-ass.net> <20181129163342.tp5wlfcyiazwwyoh@treble> <CAHk-=wg5wUXJX5mF+22WXWM_vPESb+p6-z-ZrG7rC1k2O7u=NQ@mail.gmail.com> <0A629D30-ADCF-4159-9443-E5727146F65F@amacapital.net> <CAHk-=wgoeXuodaDkKNneiQmU3mxkt2C1MeyPF5_0NX1VSQV=XQ@mail.gmail.com>
To:     Linus Torvalds <torvalds@linux-foundation.org>
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk


> On Nov 29, 2018, at 9:29 AM, Linus Torvalds <torvalds@linux-foundation.org=
> wrote:
>=20
> On Thu, Nov 29, 2018 at 9:02 AM Andy Lutomirski <luto@amacapital.net> wrot=
e:
>>>=20
>>> - just restart the instruction (with the suggested "ptregs->rip --")
>>>=20
>>> - to avoid any "oh, we're not making progress" issues, just fix the
>>> instruction yourself to be the right call, by looking it up in the
>>> "what needs to be fixed" tables.
>>=20
>> I thought that too.  I think it deadlocks. CPU A does text_poke_bp().  CP=
U B is waiting for a spinlock with IRQs off.  CPU C holds the spinlock and h=
its the int3.  The int3 never goes away because CPU A is waiting for CPU B t=
o handle the sync_core IPI.
>>=20
>> Or do you think we can avoid the IPI while the int3 is there?
>=20
> I'm handwaving and thinking that CPU C that hits the int3 can just fix
> up the instruction directly in its own caches, and return.
>=20
> Yes, it does what he "text_poke" *will* do (so now the instruction
> gets rewritten _twice_), but who cares? It's idempotent.
>=20
>=20

But it=E2=80=99s out of order. I=E2=80=99m not concerned about the final IPI=
 =E2=80=94 I=E2=80=99m concerned about the IPI after the int3 write and befo=
re the int3 is removed again. If one CPU replaces 0xcc with 0xe8, another CP=
U could observe that before the last couple bytes of the call target are wri=
tten and observed by all CPUs.=