All of lore.kernel.org
 help / color / mirror / Atom feed
From: Andrew Cooper <andrew.cooper3@citrix.com>
To: Xen-devel <xen-devel@lists.xenproject.org>
Cc: "Andrew Cooper" <andrew.cooper3@citrix.com>,
	"Jan Beulich" <JBeulich@suse.com>,
	"Roger Pau Monné" <roger.pau@citrix.com>, "Wei Liu" <wl@xen.org>
Subject: [PATCH 4/8] x86/spec-ctrl: Extend all SPEC_CTRL_{ENTER,EXIT}_* comments
Date: Wed, 13 Sep 2023 21:27:54 +0100	[thread overview]
Message-ID: <20230913202758.508225-5-andrew.cooper3@citrix.com> (raw)
In-Reply-To: <20230913202758.508225-1-andrew.cooper3@citrix.com>

... to better explain how they're used.

Doing so highlights that SPEC_CTRL_EXIT_TO_XEN is missing a VERW flush for the
corner case when e.g. an NMI hits late in an exit-to-guest path.

Leave a TODO, which will be addressed in subsequent patches which arrange for
DO_COND_VERW to be safe within SPEC_CTRL_EXIT_TO_XEN.

Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com>
---
CC: Jan Beulich <JBeulich@suse.com>
CC: Roger Pau Monné <roger.pau@citrix.com>
CC: Wei Liu <wl@xen.org>

This was decided not to be XSA-worthy, as guests can't usefully control when
IST events occur.
---
 xen/arch/x86/include/asm/spec_ctrl_asm.h | 35 ++++++++++++++++++++----
 1 file changed, 30 insertions(+), 5 deletions(-)

diff --git a/xen/arch/x86/include/asm/spec_ctrl_asm.h b/xen/arch/x86/include/asm/spec_ctrl_asm.h
index ee75f2bced42..77f6c35bb9c5 100644
--- a/xen/arch/x86/include/asm/spec_ctrl_asm.h
+++ b/xen/arch/x86/include/asm/spec_ctrl_asm.h
@@ -218,7 +218,10 @@
     wrmsr
 .endm
 
-/* Use after an entry from PV context (syscall/sysenter/int80/int82/etc). */
+/*
+ * Used after a synchronous entry from PV context.  SYSCALL, SYSENTER, INT,
+ * etc.  Will always interrupt a guest speculation context.
+ */
 .macro SPEC_CTRL_ENTRY_FROM_PV
 /*
  * Requires %rsp=regs/cpuinfo, %rdx=0
@@ -233,7 +236,11 @@
         X86_FEATURE_SC_MSR_PV
 .endm
 
-/* Use in interrupt/exception context.  May interrupt Xen or PV context. */
+/*
+ * Used after a synchronous interrupt or exception.  May interrupt Xen or PV
+ * context, but will not interrupt Xen with a guest speculation context,
+ * outside of fatal error cases.
+ */
 .macro SPEC_CTRL_ENTRY_FROM_INTR
 /*
  * Requires %rsp=regs, %r14=stack_end, %rdx=0
@@ -248,7 +255,10 @@
         X86_FEATURE_SC_MSR_PV
 .endm
 
-/* Use when exiting to PV guest context. */
+/*
+ * Used when exiting from any entry context, back to PV context.  This
+ * includes from an IST entry which moved onto the primary stack.
+ */
 .macro SPEC_CTRL_EXIT_TO_PV
 /*
  * Requires %rax=spec_ctrl, %rsp=regs/info
@@ -260,7 +270,12 @@
 .endm
 
 /*
- * Use in IST interrupt/exception context.  May interrupt Xen or PV context.
+ * Used after an IST entry (i.e. needs special care, consider to be fully
+ * asynchronous with finding sane state).  May interrupt PV or Xen context,
+ * including other SPEC_CTRL_{ENTRY,EXIT}_* regions with unsanitised state.
+ *
+ * An IST entry which interrupts PV context moves onto the primary stack and
+ * leaves via SPEC_CTRL_EXIT_TO_PV, *not* SPEC_CTRL_EXIT_TO_XEN.
  */
 .macro SPEC_CTRL_ENTRY_FROM_INTR_IST
 /*
@@ -319,7 +334,14 @@ UNLIKELY_DISPATCH_LABEL(\@_serialise):
     UNLIKELY_END(\@_serialise)
 .endm
 
-/* Use when exiting to Xen in IST context. */
+/*
+ * Use when exiting from any entry context, back to Xen context.  This
+ * includes returning to other SPEC_CTRL_{ENTRY,EXIT}_* regions with
+ * unsanitised state.
+ *
+ * Because we might have interrupted Xen beyond SPEC_CTRL_EXIT_TO_$GUEST, we
+ * must treat this as if it were an EXIT_TO_$GUEST case too.
+ */
 .macro SPEC_CTRL_EXIT_TO_XEN
 /*
  * Requires %rbx=stack_end
@@ -344,6 +366,9 @@ UNLIKELY_DISPATCH_LABEL(\@_serialise):
     wrmsr
 
 .L\@_skip_sc_msr:
+
+    /* TODO VERW */
+
 .endm
 
 #endif /* __ASSEMBLY__ */
-- 
2.30.2



  parent reply	other threads:[~2023-09-13 20:28 UTC|newest]

Thread overview: 32+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2023-09-13 20:27 [PATCH 0/8] x86/spec-ctrl: AMD DIV fix, and VERW prerequisite bugfixes Andrew Cooper
2023-09-13 20:27 ` [PATCH 1/8] x86/spec-ctrl: Fix confusion between SPEC_CTRL_EXIT_TO_XEN{,_IST} Andrew Cooper
2023-09-14  6:56   ` Jan Beulich
2023-09-14  9:54     ` Andrew Cooper
2023-09-13 20:27 ` [PATCH 2/8] x86/spec-ctrl: Fold DO_SPEC_CTRL_EXIT_TO_XEN into it's single user Andrew Cooper
2023-09-14  6:59   ` Jan Beulich
2023-09-13 20:27 ` [PATCH 3/8] x86/spec-ctrl: Turn the remaining SPEC_CTRL_{ENTRY,EXIT}_* into asm macros Andrew Cooper
2023-09-14  7:20   ` Jan Beulich
2023-09-13 20:27 ` Andrew Cooper [this message]
2023-09-14  7:58   ` [PATCH 4/8] x86/spec-ctrl: Extend all SPEC_CTRL_{ENTER,EXIT}_* comments Jan Beulich
2023-09-14 19:23     ` Andrew Cooper
2023-09-15  7:07       ` Jan Beulich
2023-09-15  9:27         ` Andrew Cooper
2023-09-13 20:27 ` [PATCH 5/8] x86/entry: Adjust restore_all_xen to hold stack_end in %r14 Andrew Cooper
2023-09-14  8:51   ` Jan Beulich
2023-09-14 19:28     ` Andrew Cooper
2023-09-13 20:27 ` [PATCH 6/8] x86/entry: Track the IST-ness of an entry for the exit paths Andrew Cooper
2023-09-14  9:32   ` Jan Beulich
2023-09-14 19:44     ` Andrew Cooper
2023-09-15  7:13       ` Jan Beulich
2023-09-15  9:30         ` Andrew Cooper
2023-09-13 20:27 ` [PATCH 7/8] x86/spec-ctrl: Issue VERW during IST exit to Xen Andrew Cooper
2023-09-14 10:01   ` Jan Beulich
2023-09-14 19:49     ` Andrew Cooper
2023-09-15  7:15       ` Jan Beulich
2023-09-13 20:27 ` [PATCH 8/8] x86/spec-ctrl: Mitigate the Zen1 DIV leakge Andrew Cooper
2023-09-13 20:43   ` Andrew Cooper
2023-09-13 21:12   ` Marek Marczykowski-Górecki
2023-09-14 10:52   ` Jan Beulich
2023-09-14 20:01     ` Andrew Cooper
2023-09-14 13:12   ` Jason Andryuk
2023-09-14 20:05     ` Andrew Cooper

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=20230913202758.508225-5-andrew.cooper3@citrix.com \
    --to=andrew.cooper3@citrix.com \
    --cc=JBeulich@suse.com \
    --cc=roger.pau@citrix.com \
    --cc=wl@xen.org \
    --cc=xen-devel@lists.xenproject.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 an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.