Date:   Fri, 17 Mar 2023 14:00:16 -0700
From:   "H. Peter Anvin" <hpa@zytor.com>
To:     andrew.cooper3@citrix.com, Lai Jiangshan <jiangshanlai@gmail.com>,
        Xin Li <xin3.li@intel.com>
CC:     linux-kernel@vger.kernel.org, x86@kernel.org, kvm@vger.kernel.org,
        tglx@linutronix.de, mingo@redhat.com, bp@alien8.de,
        dave.hansen@linux.intel.com, peterz@infradead.org,
        seanjc@google.com, pbonzini@redhat.com, ravi.v.shankar@intel.com
Subject: =?US-ASCII?Q?Re=3A_=5BPATCH_v5_28/34=5D_x86/fred=3A_fixup_fault_?= =?US-ASCII?Q?on_ERETU_by_jumping_to_fred=5Fentrypoint=5Fuser?=
User-Agent: K-9 Mail for Android
In-Reply-To: <ed318bd6-25b2-efcf-0cc4-c57699f6654a@citrix.com>
References: <20230307023946.14516-1-xin3.li@intel.com> <20230307023946.14516-29-xin3.li@intel.com> <CAJhGHyC6LgCwdDTkiy2TaQ8wzBQQfrx8ni7fY8vH-bUT2kR8pg@mail.gmail.com> <ed318bd6-25b2-efcf-0cc4-c57699f6654a@citrix.com>
Message-ID: <D3DB8A8F-7293-4715-AAB6-213F00CB521C@zytor.com>
MIME-Version: 1.0
Content-Type: text/plain;
 charset=utf-8
Content-Transfer-Encoding: quoted-printable
Precedence: bulk

On March 17, 2023 2:55:44 AM PDT, andrew=2Ecooper3@citrix=2Ecom wrote:
>On 17/03/2023 9:39 am, Lai Jiangshan wrote:
>>> +#ifdef CONFIG_X86_FRED
>>> +static bool ex_handler_eretu(const struct exception_table_entry *fixu=
p,
>>> +                            struct pt_regs *regs, unsigned long error=
_code)
>>> +{
>>> +       struct pt_regs *uregs =3D (struct pt_regs *)(regs->sp - offset=
of(struct pt_regs, ip));
>>> +       unsigned short ss =3D uregs->ss;
>>> +       unsigned short cs =3D uregs->cs;
>>> +
>>> +       fred_info(uregs)->edata =3D fred_event_data(regs);
>>> +       uregs->ssx =3D regs->ssx;
>>> +       uregs->ss =3D ss;
>>> +       uregs->csx =3D regs->csx;
>>> +       uregs->current_stack_level =3D 0;
>>> +       uregs->cs =3D cs;
>> Hello
>>
>> If the ERETU instruction had tried to return from NMI to ring3 and just=
 faulted,
>> is NMI still blocked?
>>
>> We know that IRET unconditionally enables NMI, but I can't find any clu=
e in the
>> FRED's manual=2E
>>
>> In the pseudocode of ERETU in the manual, it seems that NMI is only ena=
bled when
>> ERETU succeeds with bit28 in csx set=2E  If so, this code will fail to =
reenable
>> NMI if bit28 is not explicitly re-set in csx=2E
>
>IRET clearing NMI blocking is the source of an immense amount of grief,
>and ultimately the reason why Linux and others can't use supervisor
>shadow stacks at the moment=2E
>
>Changing this property, so NMIs only get unblocked on successful
>execution of an ERET{S,U}, was a key demand of the FRED spec=2E
>
>i=2Ee=2E until you have successfully ERET*'d, you're still logically in t=
he
>NMI handler and NMIs need to remain blocked even when handling the #GP
>from a bad ERET=2E
>
>~Andrew

This is correct=2E