From: Peter Zijlstra <peterz@infradead.org>
To: Joerg Roedel <joro@8bytes.org>
Cc: x86@kernel.org, Joerg Roedel <jroedel@suse.de>,
hpa@zytor.com, Andy Lutomirski <luto@kernel.org>,
Dave Hansen <dave.hansen@linux.intel.com>,
Jiri Slaby <jslaby@suse.cz>,
Dan Williams <dan.j.williams@intel.com>,
Tom Lendacky <thomas.lendacky@amd.com>,
Juergen Gross <jgross@suse.com>,
Kees Cook <keescook@chromium.org>,
David Rientjes <rientjes@google.com>,
Cfir Cohen <cfir@google.com>, Erdem Aktas <erdemaktas@google.com>,
Masami Hiramatsu <mhiramat@kernel.org>,
Mike Stunes <mstunes@vmware.com>,
Sean Christopherson <sean.j.christopherson@intel.com>,
Martin Radev <martin.b.radev@gmail.com>,
linux-kernel@vger.kernel.org, kvm@vger.kernel.org,
virtualization@lists.linux-foundation.org
Subject: Re: [PATCH v4 45/75] x86/sev-es: Adjust #VC IST Stack on entering NMI handler
Date: Wed, 15 Jul 2020 11:47:02 +0200 [thread overview]
Message-ID: <20200715094702.GF10769@hirez.programming.kicks-ass.net> (raw)
In-Reply-To: <20200714120917.11253-46-joro@8bytes.org>
On Tue, Jul 14, 2020 at 02:08:47PM +0200, Joerg Roedel wrote:
> @@ -489,6 +490,9 @@ DEFINE_IDTENTRY_RAW(exc_nmi)
> this_cpu_write(nmi_cr2, read_cr2());
> nmi_restart:
>
> + /* Needs to happen before DR7 is accessed */
> + sev_es_ist_enter(regs);
> +
> this_cpu_write(nmi_dr7, local_db_save());
>
> nmi_enter();
> @@ -502,6 +506,8 @@ DEFINE_IDTENTRY_RAW(exc_nmi)
>
> local_db_restore(this_cpu_read(nmi_dr7));
>
> + sev_es_ist_exit();
> +
> if (unlikely(this_cpu_read(nmi_cr2) != read_cr2()))
> write_cr2(this_cpu_read(nmi_cr2));
> if (this_cpu_dec_return(nmi_state))
I really hate all this #VC stuff :-(
So the above will make the NMI do 4 unconditional extra CALL+RET, a LOAD
(which will potentially miss) and a compare and branch.
How's that a win for normal people? Can we please turn all these
sev_es_*() hooks into something like:
DECLARE_STATIC_KEY_FALSE(sev_es_enabled_key);
static __always_inline void sev_es_foo()
{
if (static_branch_unlikely(&sev_es_enabled_key))
__sev_es_foo();
}
So that normal people will only see an extra NOP?
> diff --git a/arch/x86/kernel/sev-es.c b/arch/x86/kernel/sev-es.c
> index d415368f16ec..2a7cc72db1d5 100644
> --- a/arch/x86/kernel/sev-es.c
> +++ b/arch/x86/kernel/sev-es.c
> @@ -78,6 +78,67 @@ static void __init sev_es_setup_vc_stacks(int cpu)
> tss->x86_tss.ist[IST_INDEX_VC] = CEA_ESTACK_TOP(&cea->estacks, VC);
> }
>
> +static bool on_vc_stack(unsigned long sp)
noinstr or __always_inline
> +{
> + return ((sp >= __this_cpu_ist_bot_va(VC)) && (sp < __this_cpu_ist_top_va(VC)));
> +}
> +
> +/*
> + * This function handles the case when an NMI or an NMI-like exception
> + * like #DB is raised in the #VC exception handler entry code. In this
I've yet to find you handle the NMI-like cases..
> + * case the IST entry for VC must be adjusted, so that any subsequent VC
> + * exception will not overwrite the stack contents of the interrupted VC
> + * handler.
> + *
> + * The IST entry is adjusted unconditionally so that it can be also be
> + * unconditionally back-adjusted in sev_es_nmi_exit(). Otherwise a
> + * nested nmi_exit() call (#VC->NMI->#DB) may back-adjust the IST entry
> + * too early.
Is this comment accurate, I cannot find the patch touching
nmi_enter/exit()?
> + */
> +void noinstr sev_es_ist_enter(struct pt_regs *regs)
> +{
> + unsigned long old_ist, new_ist;
> + unsigned long *p;
> +
> + if (!sev_es_active())
> + return;
> +
> + /* Read old IST entry */
> + old_ist = __this_cpu_read(cpu_tss_rw.x86_tss.ist[IST_INDEX_VC]);
> +
> + /* Make room on the IST stack */
> + if (on_vc_stack(regs->sp))
> + new_ist = ALIGN_DOWN(regs->sp, 8) - sizeof(old_ist);
> + else
> + new_ist = old_ist - sizeof(old_ist);
> +
> + /* Store old IST entry */
> + p = (unsigned long *)new_ist;
> + *p = old_ist;
> +
> + /* Set new IST entry */
> + this_cpu_write(cpu_tss_rw.x86_tss.ist[IST_INDEX_VC], new_ist);
> +}
> +
> +void noinstr sev_es_ist_exit(void)
> +{
> + unsigned long ist;
> + unsigned long *p;
> +
> + if (!sev_es_active())
> + return;
> +
> + /* Read IST entry */
> + ist = __this_cpu_read(cpu_tss_rw.x86_tss.ist[IST_INDEX_VC]);
> +
> + if (WARN_ON(ist == __this_cpu_ist_top_va(VC)))
> + return;
> +
> + /* Read back old IST entry and write it to the TSS */
> + p = (unsigned long *)ist;
> + this_cpu_write(cpu_tss_rw.x86_tss.ist[IST_INDEX_VC], *p);
> +}
That's pretty disguisting :-(
next prev parent reply other threads:[~2020-07-15 9:47 UTC|newest]
Thread overview: 118+ messages / expand[flat|nested] mbox.gz Atom feed top
2020-07-14 12:08 [PATCH v4 00/75] x86: SEV-ES Guest Support Joerg Roedel
2020-07-14 12:08 ` [PATCH v4 01/75] KVM: SVM: Add GHCB definitions Joerg Roedel
2020-07-14 12:08 ` [PATCH v4 02/75] KVM: SVM: Add GHCB Accessor functions Joerg Roedel
2020-07-14 12:08 ` [PATCH v4 03/75] KVM: SVM: Use __packed shorthand Joerg Roedel
2020-07-14 12:08 ` [PATCH v4 04/75] x86/cpufeatures: Add SEV-ES CPU feature Joerg Roedel
2020-07-14 12:08 ` [PATCH v4 05/75] x86/traps: Move pf error codes to <asm/trap_pf.h> Joerg Roedel
2020-07-14 12:08 ` [PATCH v4 06/75] x86/insn: Make inat-tables.c suitable for pre-decompression code Joerg Roedel
2020-07-17 13:58 ` Masami Hiramatsu
2020-07-14 12:08 ` [PATCH v4 07/75] x86/umip: Factor out instruction fetch Joerg Roedel
2020-07-14 12:08 ` [PATCH v4 08/75] x86/umip: Factor out instruction decoding Joerg Roedel
2020-07-14 12:08 ` [PATCH v4 09/75] x86/insn: Add insn_get_modrm_reg_off() Joerg Roedel
2020-07-14 12:08 ` [PATCH v4 10/75] x86/insn: Add insn_has_rep_prefix() helper Joerg Roedel
2020-07-17 14:06 ` Masami Hiramatsu
2020-07-14 12:08 ` [PATCH v4 11/75] x86/boot/compressed/64: Disable red-zone usage Joerg Roedel
2020-07-14 12:08 ` [PATCH v4 12/75] x86/boot/compressed/64: Add IDT Infrastructure Joerg Roedel
2020-07-14 12:08 ` [PATCH v4 13/75] x86/boot/compressed/64: Rename kaslr_64.c to ident_map_64.c Joerg Roedel
2020-07-15 1:23 ` Kees Cook
2020-07-14 12:08 ` [PATCH v4 14/75] x86/boot/compressed/64: Add page-fault handler Joerg Roedel
2020-07-15 1:24 ` Kees Cook
2020-07-14 12:08 ` [PATCH v4 15/75] x86/boot/compressed/64: Always switch to own page-table Joerg Roedel
2020-07-15 1:23 ` Kees Cook
2020-07-14 12:08 ` [PATCH v4 16/75] x86/boot/compressed/64: Don't pre-map memory in KASLR code Joerg Roedel
2020-07-15 1:24 ` Kees Cook
2020-07-14 12:08 ` [PATCH v4 17/75] x86/boot/compressed/64: Change add_identity_map() to take start and end Joerg Roedel
2020-07-15 1:24 ` Kees Cook
2020-07-14 12:08 ` [PATCH v4 18/75] x86/boot/compressed/64: Add stage1 #VC handler Joerg Roedel
2020-07-14 12:08 ` [PATCH v4 19/75] x86/boot/compressed/64: Call set_sev_encryption_mask earlier Joerg Roedel
2020-07-14 12:08 ` [PATCH v4 20/75] x86/boot/compressed/64: Check return value of kernel_ident_mapping_init() Joerg Roedel
2020-07-14 12:08 ` [PATCH v4 21/75] x86/boot/compressed/64: Add set_page_en/decrypted() helpers Joerg Roedel
2020-07-14 12:08 ` [PATCH v4 22/75] x86/boot/compressed/64: Setup GHCB Based VC Exception handler Joerg Roedel
2020-07-14 12:08 ` [PATCH v4 23/75] x86/boot/compressed/64: Unmap GHCB page before booting the kernel Joerg Roedel
2020-07-14 12:08 ` [PATCH v4 24/75] x86/sev-es: Add support for handling IOIO exceptions Joerg Roedel
2020-07-14 12:08 ` [PATCH v4 25/75] x86/fpu: Move xgetbv()/xsetbv() into separate header Joerg Roedel
2020-07-14 12:08 ` [PATCH v4 26/75] x86/sev-es: Add CPUID handling to #VC handler Joerg Roedel
2020-07-14 12:08 ` [PATCH v4 27/75] x86/idt: Move IDT to data segment Joerg Roedel
2020-07-15 1:25 ` Kees Cook
2020-07-14 12:08 ` [PATCH v4 28/75] x86/idt: Split idt_data setup out of set_intr_gate() Joerg Roedel
2020-07-15 1:26 ` Kees Cook
2020-07-14 12:08 ` [PATCH v4 29/75] x86/idt: Move two function from k/idt.c to i/a/desc.h Joerg Roedel
2020-07-15 1:29 ` Kees Cook
2020-07-14 12:08 ` [PATCH v4 30/75] x86/head/64: Install boot GDT Joerg Roedel
2020-07-14 12:08 ` [PATCH v4 31/75] x86/head/64: Reload GDT after switch to virtual addresses Joerg Roedel
2020-07-14 12:08 ` [PATCH v4 32/75] x86/head/64: Load segment registers earlier Joerg Roedel
2020-07-14 12:08 ` [PATCH v4 33/75] x86/head/64: Switch to initial stack earlier Joerg Roedel
2020-07-14 12:08 ` [PATCH v4 34/75] x86/head/64: Build k/head64.c with -fno-stack-protector Joerg Roedel
2020-07-15 1:34 ` Kees Cook
2020-07-15 16:34 ` Joerg Roedel
2020-07-14 12:08 ` [PATCH v4 35/75] x86/head/64: Load IDT earlier Joerg Roedel
2020-07-14 12:08 ` [PATCH v4 36/75] x86/head/64: Move early exception dispatch to C code Joerg Roedel
2020-07-14 12:08 ` [PATCH v4 37/75] x86/sev-es: Add SEV-ES Feature Detection Joerg Roedel
2020-07-14 12:08 ` [PATCH v4 38/75] x86/sev-es: Print SEV-ES info into kernel log Joerg Roedel
2020-07-14 12:08 ` [PATCH v4 39/75] x86/sev-es: Compile early handler code into kernel image Joerg Roedel
2020-07-14 12:08 ` [PATCH v4 40/75] x86/sev-es: Setup early #VC handler Joerg Roedel
2020-07-14 12:08 ` [PATCH v4 41/75] x86/sev-es: Setup GHCB based boot " Joerg Roedel
2020-07-14 12:08 ` [PATCH v4 42/75] x86/sev-es: Setup per-cpu GHCBs for the runtime handler Joerg Roedel
2020-07-14 12:08 ` [PATCH v4 43/75] x86/sev-es: Allocate and Map stacks for #VC handler Joerg Roedel
2020-07-14 12:08 ` [PATCH v4 44/75] x86/sev-es: Allocate and setup IST entry for #VC Joerg Roedel
2020-07-14 12:08 ` [PATCH v4 45/75] x86/sev-es: Adjust #VC IST Stack on entering NMI handler Joerg Roedel
2020-07-15 9:47 ` Peter Zijlstra [this message]
2020-07-15 10:26 ` Joerg Roedel
2020-07-15 10:56 ` Peter Zijlstra
2020-07-14 12:08 ` [PATCH v4 46/75] x86/dumpstack/64: Add noinstr version of get_stack_info() Joerg Roedel
2020-07-14 12:08 ` [PATCH v4 47/75] x86/entry/64: Add entry code for #VC handler Joerg Roedel
2020-07-14 12:08 ` [PATCH v4 48/75] x86/sev-es: Add Runtime #VC Exception Handler Joerg Roedel
2020-07-14 12:08 ` [PATCH v4 49/75] x86/sev-es: Wire up existing #VC exit-code handlers Joerg Roedel
2020-07-14 12:08 ` [PATCH v4 50/75] x86/sev-es: Handle instruction fetches from user-space Joerg Roedel
2020-07-14 12:08 ` [PATCH v4 51/75] x86/sev-es: Handle MMIO events Joerg Roedel
2020-07-21 21:01 ` Mike Stunes
2020-07-22 7:55 ` Joerg Roedel
2020-07-22 8:05 ` Joerg Roedel
2020-07-22 22:53 ` Mike Stunes
2020-07-23 7:21 ` Joerg Roedel
2020-07-14 12:08 ` [PATCH v4 52/75] x86/sev-es: Handle MMIO String Instructions Joerg Roedel
2020-07-14 12:08 ` [PATCH v4 53/75] x86/sev-es: Handle MSR events Joerg Roedel
2020-07-14 12:08 ` [PATCH v4 54/75] x86/sev-es: Handle DR7 read/write events Joerg Roedel
2020-07-14 12:08 ` [PATCH v4 55/75] x86/sev-es: Handle WBINVD Events Joerg Roedel
2020-07-14 12:08 ` [PATCH v4 56/75] x86/sev-es: Handle RDTSC(P) Events Joerg Roedel
2020-07-14 12:08 ` [PATCH v4 57/75] x86/sev-es: Handle RDPMC Events Joerg Roedel
2020-07-14 12:09 ` [PATCH v4 58/75] x86/sev-es: Handle INVD Events Joerg Roedel
2020-07-14 12:09 ` [PATCH v4 59/75] x86/sev-es: Handle MONITOR/MONITORX Events Joerg Roedel
2020-07-14 12:09 ` [PATCH v4 60/75] x86/sev-es: Handle MWAIT/MWAITX Events Joerg Roedel
2020-07-14 12:09 ` [PATCH v4 61/75] x86/sev-es: Handle VMMCALL Events Joerg Roedel
2020-07-14 12:09 ` [PATCH v4 62/75] x86/sev-es: Handle #AC Events Joerg Roedel
2020-07-14 12:09 ` [PATCH v4 63/75] x86/sev-es: Handle #DB Events Joerg Roedel
2020-07-15 8:47 ` Peter Zijlstra
2020-07-15 9:13 ` Joerg Roedel
2020-07-15 9:51 ` Peter Zijlstra
2020-07-15 10:08 ` Joerg Roedel
2020-07-15 10:13 ` Peter Zijlstra
2020-07-15 10:31 ` Joerg Roedel
2020-07-14 12:09 ` [PATCH v4 64/75] x86/paravirt: Allow hypervisor specific VMMCALL handling under SEV-ES Joerg Roedel
2020-07-14 12:09 ` [PATCH v4 65/75] x86/kvm: Add KVM " Joerg Roedel
2020-07-14 12:09 ` [PATCH v4 66/75] x86/vmware: Add VMware specific handling for VMMCALL " Joerg Roedel
2020-07-14 12:09 ` [PATCH v4 67/75] x86/realmode: Add SEV-ES specific trampoline entry point Joerg Roedel
2020-07-14 12:09 ` [PATCH v4 68/75] x86/realmode: Setup AP jump table Joerg Roedel
2020-07-14 12:09 ` [PATCH v4 69/75] x86/head/64: Setup TSS early for secondary CPUs Joerg Roedel
2020-07-14 12:09 ` [PATCH v4 70/75] x86/head/64: Don't call verify_cpu() on starting APs Joerg Roedel
2020-07-15 1:40 ` Kees Cook
2020-07-15 9:26 ` Joerg Roedel
2020-07-15 15:26 ` Kees Cook
2020-07-15 15:48 ` Joerg Roedel
2020-07-15 19:49 ` Kees Cook
2020-07-20 15:29 ` Joerg Roedel
2020-07-14 12:09 ` [PATCH v4 71/75] x86/head/64: Rename start_cpu0 Joerg Roedel
2020-07-14 12:09 ` [PATCH v4 72/75] x86/sev-es: Support CPU offline/online Joerg Roedel
2020-07-14 12:09 ` [PATCH v4 73/75] x86/sev-es: Handle NMI State Joerg Roedel
2020-07-14 12:09 ` [PATCH v4 74/75] x86/efi: Add GHCB mappings when SEV-ES is active Joerg Roedel
2020-07-14 12:09 ` [PATCH v4 75/75] x86/sev-es: Check required CPU features for SEV-ES Joerg Roedel
2020-07-15 9:24 ` [PATCH v4 00/75] x86: SEV-ES Guest Support Peter Zijlstra
2020-07-15 9:34 ` Joerg Roedel
2020-07-15 9:55 ` Peter Zijlstra
2020-07-15 10:10 ` Joerg Roedel
2020-07-21 1:09 ` Erdem Aktas
2020-07-21 12:49 ` Joerg Roedel
2020-07-21 16:48 ` Erdem Aktas
2020-07-22 9:04 ` Joerg Roedel
2020-07-22 16:54 ` Erdem Aktas
2020-07-22 17:45 ` Joerg Roedel
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20200715094702.GF10769@hirez.programming.kicks-ass.net \
--to=peterz@infradead.org \
--cc=cfir@google.com \
--cc=dan.j.williams@intel.com \
--cc=dave.hansen@linux.intel.com \
--cc=erdemaktas@google.com \
--cc=hpa@zytor.com \
--cc=jgross@suse.com \
--cc=joro@8bytes.org \
--cc=jroedel@suse.de \
--cc=jslaby@suse.cz \
--cc=keescook@chromium.org \
--cc=kvm@vger.kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=luto@kernel.org \
--cc=martin.b.radev@gmail.com \
--cc=mhiramat@kernel.org \
--cc=mstunes@vmware.com \
--cc=rientjes@google.com \
--cc=sean.j.christopherson@intel.com \
--cc=thomas.lendacky@amd.com \
--cc=virtualization@lists.linux-foundation.org \
--cc=x86@kernel.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).