Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 23.128.96.18 as permitted sender) client-ip=23.128.96.18;
Date:   Tue, 02 Nov 2021 12:22:50 +0100
From:   "H. Peter Anvin" <hpa@zytor.com>
To:     Borislav Petkov <bp@alien8.de>,
        Lai Jiangshan <laijs@linux.alibaba.com>
CC:     Lai Jiangshan <jiangshanlai@gmail.com>,
        linux-kernel@vger.kernel.org, x86@kernel.org,
        Jan Beulich <jbeulich@suse.com>,
        Thomas Gleixner <tglx@linutronix.de>,
        Juergen Gross <jgross@suse.com>,
        xen-devel@lists.xenproject.org, Andy Lutomirski <luto@kernel.org>,
        Ingo Molnar <mingo@redhat.com>,
        Dave Hansen <dave.hansen@linux.intel.com>,
        Boris Ostrovsky <boris.ostrovsky@oracle.com>,
        Stefano Stabellini <sstabellini@kernel.org>
Subject: =?US-ASCII?Q?Re=3A_=5BPATCH_V4_04/50=5D_x86/xen=3A_Add_xenp?= =?US-ASCII?Q?v=5Frestore=5Fregs=5Fand=5Freturn=5Fto=5Fusermode=28=29?=
User-Agent: K-9 Mail for Android
In-Reply-To: <YYEJuIgQukcDzy1R@zn.tnic>
References: <20211026141420.17138-1-jiangshanlai@gmail.com> <20211026141420.17138-5-jiangshanlai@gmail.com> <YYD9ohN2Zcy4EdMb@zn.tnic> <d4ae23dd-377e-8316-909b-d5bdeacc0904@linux.alibaba.com> <YYEJuIgQukcDzy1R@zn.tnic>
Message-ID: <6F6D3FEC-9AF1-40E1-A7C2-394D21C40114@zytor.com>
MIME-Version: 1.0
Content-Type: text/plain;
 charset=utf-8
Content-Transfer-Encoding: quoted-printable
Precedence: bulk


On November 2, 2021 10:49:44 AM GMT+01:00, Borislav Petkov <bp@alien8=2Ede=
> wrote:
>On Tue, Nov 02, 2021 at 05:19:46PM +0800, Lai Jiangshan wrote:
>> It will add a 5-byte NOP at the beginning of the native
>> swapgs_restore_regs_and_return_to_usermode=2E
>
>So?
>

It would be interesting to have an "override function with jmp" alternativ=
es macro=2E It doesn't require any changes to the alternatives mechanism pr=
oper (but possibly to objtool): it would just insert an alternatives entry =
without adding any code including nops to the main path=2E It would of cour=
se only be applicable to a jmp, so a syntax like OVERRIDE_JMP feature, targ=
et rather than open-coding the instruction would probably be a good idea=2E

That would reduce the trade-off to zero=2E

>> I avoided adding unneeded code in the native code even if it is NOPs
>> and avoided melting xenpv-one into the native one which will reduce
>> the code readability=2E
>
>How does this reduce code readability?!
>
>diff --git a/arch/x86/entry/entry_64=2ES b/arch/x86/entry/entry_64=2ES
>index e38a4cf795d9=2E=2Ebf1de54a1fca 100644
>--- a/arch/x86/entry/entry_64=2ES
>+++ b/arch/x86/entry/entry_64=2ES
>@@ -567,6 +567,10 @@ __irqentry_text_end:
>=20
> SYM_CODE_START_LOCAL(common_interrupt_return)
> SYM_INNER_LABEL(swapgs_restore_regs_and_return_to_usermode, SYM_L_GLOBAL=
)
>+
>+	ALTERNATIVE "", "jmp xenpv_restore_regs_and_return_to_usermode", \
>+                X86_FEATURE_XENPV
>+
> #ifdef CONFIG_DEBUG_ENTRY
> 	/* Assert that pt_regs indicates user mode=2E */
> 	testb	$3, CS(%rsp)
>
>> I will follow your preference since a 5-byte NOP is so negligible in th=
e slow
>> path with an iret instruction=2E
>
>Yes, we do already gazillion things on those entry and exit paths=2E
>
>> Or other option that adds macros to wrap the ALTERNATIVE=2E
>> RESTORE_REGS_AND_RETURN_TO_USERMODE and
>> COND_RESTORE_REGS_AND_RETURN_TO_USERMODE (test %eax before jmp in nativ=
e case)
>
>No, the main goal is to keep the asm code as readable and as simple as
>possible=2E
>
>If macros or whatever need to be added, there better be a good reason
>for them=2E Saving a NOP is not one of them=2E
>
>Thx=2E
>

--=20
Sent from my Android device with K-9 Mail=2E Please excuse my brevity=2E