From: Peter Zijlstra <peterz@infradead.org>
To: tglx@linutronix.de, luto@amacapital.net, peterz@infradead.org
Cc: linux-kernel@vger.kernel.org, x86@kernel.org,
Lai Jiangshan <laijs@linux.alibaba.com>,
sean.j.christopherson@intel.com, andrew.cooper3@citrix.com,
daniel.thompson@linaro.org, a.darwish@linutronix.de,
rostedt@goodmis.org, bigeasy@linutronix.de,
Andy Lutomirski <luto@kernel.org>
Subject: [PATCH 08/14] x86/entry: Optimize local_db_save() for virt
Date: Fri, 29 May 2020 23:27:36 +0200 [thread overview]
Message-ID: <20200529213321.187833200@infradead.org> (raw)
In-Reply-To: 20200529212728.795169701@infradead.org
Because DRn access is 'difficult' with virt; but the DR7 read is
cheaper than a cacheline miss on native, add a virt specific
fast path to local_db_save(), such that when breakpoints are not in
use we avoid touching DRn entirely.
Suggested-by: Andy Lutomirski <luto@kernel.org>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
---
arch/x86/include/asm/debugreg.h | 7 ++++++-
arch/x86/kernel/hw_breakpoint.c | 26 ++++++++++++++++++++++----
arch/x86/kvm/vmx/nested.c | 2 +-
3 files changed, 29 insertions(+), 6 deletions(-)
--- a/arch/x86/include/asm/debugreg.h
+++ b/arch/x86/include/asm/debugreg.h
@@ -85,8 +85,8 @@ static inline void hw_breakpoint_disable
set_debugreg(0UL, 3);
}
-static inline int hw_breakpoint_active(void)
+static inline bool hw_breakpoint_active(void)
{
return __this_cpu_read(cpu_dr7) & DR_GLOBAL_ENABLE_MASK;
}
@@ -117,6 +119,9 @@ static __always_inline unsigned long loc
{
unsigned long dr7;
+ if (static_cpu_has(X86_FEATURE_HYPERVISOR) && !hw_breakpoint_active())
+ return 0;
+
get_debugreg(dr7, 7);
dr7 &= ~0x400; /* architecturally set bit */
if (dr7)
--- a/arch/x86/kernel/hw_breakpoint.c
+++ b/arch/x86/kernel/hw_breakpoint.c
@@ -99,6 +99,8 @@ int arch_install_hw_breakpoint(struct pe
unsigned long *dr7;
int i;
+ lockdep_assert_irqs_disabled();
+
for (i = 0; i < HBP_NUM; i++) {
struct perf_event **slot = this_cpu_ptr(&bp_per_reg[i]);
@@ -117,6 +119,12 @@ int arch_install_hw_breakpoint(struct pe
dr7 = this_cpu_ptr(&cpu_dr7);
*dr7 |= encode_dr7(i, info->len, info->type);
+ /*
+ * Ensure we first write cpu_dr7 before we set the DR7 register.
+ * This ensures an NMI never see cpu_dr7 0 when DR7 is not.
+ */
+ barrier();
+
set_debugreg(*dr7, 7);
if (info->mask)
set_dr_addr_mask(info->mask, i);
@@ -136,9 +144,11 @@ int arch_install_hw_breakpoint(struct pe
void arch_uninstall_hw_breakpoint(struct perf_event *bp)
{
struct arch_hw_breakpoint *info = counter_arch_bp(bp);
- unsigned long *dr7;
+ unsigned long dr7;
int i;
+ lockdep_assert_irqs_disabled();
+
for (i = 0; i < HBP_NUM; i++) {
struct perf_event **slot = this_cpu_ptr(&bp_per_reg[i]);
@@ -151,12 +161,20 @@ void arch_uninstall_hw_breakpoint(struct
if (WARN_ONCE(i == HBP_NUM, "Can't find any breakpoint slot"))
return;
- dr7 = this_cpu_ptr(&cpu_dr7);
- *dr7 &= ~__encode_dr7(i, info->len, info->type);
+ dr7 = this_cpu_read(cpu_dr7);
+ dr7 &= ~__encode_dr7(i, info->len, info->type);
- set_debugreg(*dr7, 7);
+ set_debugreg(dr7, 7);
if (info->mask)
set_dr_addr_mask(0, i);
+
+ /*
+ * Ensure the write to cpu_dr7 is after we've set the DR7 register.
+ * This ensures an NMI never see cpu_dr7 0 when DR7 is not.
+ */
+ barrier();
+
+ this_cpu_write(cpu_dr7, dr7);
}
static int arch_bp_generic_len(int x86_len)
--- a/arch/x86/kvm/vmx/nested.c
+++ b/arch/x86/kvm/vmx/nested.c
@@ -3028,9 +3028,9 @@ static int nested_vmx_check_vmentry_hw(s
/*
* VMExit clears RFLAGS.IF and DR7, even on a consistency check.
*/
- local_irq_enable();
if (hw_breakpoint_active())
set_debugreg(__this_cpu_read(cpu_dr7), 7);
+ local_irq_enable();
preempt_enable();
/*
next prev parent reply other threads:[~2020-05-29 21:37 UTC|newest]
Thread overview: 35+ messages / expand[flat|nested] mbox.gz Atom feed top
2020-05-29 21:27 [PATCH 00/14] x86/entry: disallow #DB more and x86/entry lockdep/nmi Peter Zijlstra
2020-05-29 21:27 ` [PATCH 01/14] x86/hw_breakpoint: Add within_area() to check data breakpoints Peter Zijlstra
2020-05-29 21:27 ` [PATCH 02/14] x86/hw_breakpoint: Prevent data breakpoints on direct GDT Peter Zijlstra
2020-05-30 12:45 ` Andrew Cooper
2020-05-30 15:15 ` Lai Jiangshan
2020-05-29 21:27 ` [PATCH 03/14] x86/hw_breakpoint: Prevent data breakpoints on per_cpu cpu_tss_rw Peter Zijlstra
2020-05-29 21:27 ` [PATCH 04/14] x86/hw_breakpoint: Prevent data breakpoints on user_pcid_flush_mask Peter Zijlstra
2020-05-29 21:27 ` [PATCH 05/14] x86/entry: Introduce local_db_{save,restore}() Peter Zijlstra
2020-05-30 9:57 ` [tip: x86/entry] " tip-bot2 for Peter Zijlstra
2020-05-29 21:27 ` [PATCH 06/14] x86/entry, nmi: Disable #DB Peter Zijlstra
2020-05-30 9:57 ` [tip: x86/entry] " tip-bot2 for Peter Zijlstra
2020-05-29 21:27 ` [PATCH 07/14] x86/entry, mce: Disallow #DB during #MC Peter Zijlstra
2020-05-30 9:57 ` [tip: x86/entry] " tip-bot2 for Peter Zijlstra
2020-05-29 21:27 ` Peter Zijlstra [this message]
2020-05-30 9:57 ` [tip: x86/entry] x86/entry: Optimize local_db_save() for virt tip-bot2 for Peter Zijlstra
2020-06-03 1:17 ` [PATCH 08/14] " Sean Christopherson
2020-05-29 21:27 ` [PATCH 09/14] x86/entry: Remove debug IDT frobbing Peter Zijlstra
2020-05-30 9:57 ` [tip: x86/entry] " tip-bot2 for Peter Zijlstra
2020-05-29 21:27 ` [PATCH 10/14] x86/entry: Remove DBn stacks Peter Zijlstra
2020-05-30 9:57 ` [tip: x86/entry] " tip-bot2 for Peter Zijlstra
2020-05-29 21:27 ` [PATCH 11/14] x86/entry: Clarify irq_{enter,exit}_rcu() Peter Zijlstra
2020-05-30 9:57 ` [tip: x86/entry] " tip-bot2 for Peter Zijlstra
2020-06-02 14:42 ` [PATCH 11/14] " Qian Cai
2020-06-02 15:05 ` Peter Zijlstra
2020-06-02 18:47 ` Qian Cai
2020-06-03 17:50 ` [tip: x86/entry] x86/entry: Use __irq_exit_rcu() in irq_exit() tip-bot2 for Peter Zijlstra
2020-05-29 21:27 ` [PATCH 12/14] x86/entry: Rename trace_hardirqs_off_prepare() Peter Zijlstra
2020-05-30 9:57 ` [tip: x86/entry] " tip-bot2 for Peter Zijlstra
2020-05-29 21:27 ` [PATCH 13/14] lockdep: Prepare for NMI IRQ state tracking Peter Zijlstra
2020-05-29 22:14 ` Steven Rostedt
2020-05-29 22:25 ` Peter Zijlstra
2020-05-29 22:28 ` Steven Rostedt
2020-05-29 22:33 ` Peter Zijlstra
2020-06-02 20:00 ` Peter Zijlstra
2020-05-29 21:27 ` [PATCH 14/14] x86/entry: Fix NMI vs " Peter Zijlstra
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=20200529213321.187833200@infradead.org \
--to=peterz@infradead.org \
--cc=a.darwish@linutronix.de \
--cc=andrew.cooper3@citrix.com \
--cc=bigeasy@linutronix.de \
--cc=daniel.thompson@linaro.org \
--cc=laijs@linux.alibaba.com \
--cc=linux-kernel@vger.kernel.org \
--cc=luto@amacapital.net \
--cc=luto@kernel.org \
--cc=rostedt@goodmis.org \
--cc=sean.j.christopherson@intel.com \
--cc=tglx@linutronix.de \
--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).