Skip to contentSkip to conversation

x86/split_lock: Enable #AC exception for split locked accesses

PATCH Superseded Fenghua Yu 7 participants 65 messages Mar 2, 2019 – Mar 11, 2019

Subsystems (7)

v4 of 9 Superseded Latest is v9 →

lore
Share & tools
lore

Fenghua Yu

author
#0 lore raw

v4 cover x86/split_lock: Enable #AC exception for split locked accesses

==Introduction==

A split lock is any atomic operation whose operand crosses two cache lines. Since the operand spans two cache lines and the operation must be atomic, the system locks the bus while the CPU accesses the two cache lines.

During bus locking, request from other CPUs or bus agents for control of the bus are blocked. Blocking bus access from other CPUs plus overhead of configuring bus locking protocol degrade not only the performance of one CPU but overall system performance.

If operand is cacheable and completely contained in one cache line, atomic operation is optimized by less expensive cache locking on Intel P6 and recent processors. If split lock is detected and the two cache lines in the operand can be merged into one cache line, cache locking instead of more expensive bus locking will be used for atomic operation. Removing split lock can improve overall performance.

Instructions that may cause split lock issue include lock add, lock btc, xchg, lsl, far call, ltr, etc.

More information about split lock, bus locking, and cache locking can be found in the latest Intel 64 and IA-32 Architecture Software Developer's Manual.

==Split lock detection==

Currently we can trace split lock event counter sq_misc.split_lock for debug purpose. But for system deployed in the field, this event counter after the fact is insufficient. We need a mechanism that detects split lock before it happens to ensure that bus lock is never incurred due to split lock.

Intel introduces a mechanism to detect split lock via Alignment Check (#AC) exception before badly aligned atomic instructions might impact whole system performance in Tremont and other future processors.

This capability is critical for real time system designers who build consolidated real time systems. These systems run hard real time code on some cores and run "untrusted" user processes on some other cores. The hard real time cannot afford to have any bus lock from the untrusted processes to hurt real time performance. To date the designers have been unable to deploy these solutions as they have no way to prevent the "untrusted" user code from generating split lock and bus lock to block the hard real time code to access memory during bus locking.

This capability may also find usage in cloud. A user process with split lock running in one guest can block other cores from accessing shared memory during its split locked memory access. That may cause overall system performance degradation.

Split lock may open a security hole where malicious user code may slow down overall system by executing instructions with split lock.

==Enumerate #AC for split lock==

A control bit (bit 29) in TEST_CTL MSR 0x33 will be introduced in future x86 processors. When the bit 29 is set, the processor causes #AC exception for split locked accesses at all CPL.

The split lock detection feature is enumerated through bit 5 in MSR IA32_CORE_CAPABILITY (0xCF). The MSR 0xCF itself is enumerated by CPUID.(EAX=0x7,ECX=0):EDX[30].

The enumeration method is published in the latest Intel Software Developer's Manual.

A few processors don't follow the above enumeration. For those processors, we can enable #AC for split lock based on their FMS numbers. We will enable split lock detection on those processors in the future once their FMS numbers are public.

==Handle Split Lock===

There may be different considerations to handle split lock, e.g. how to handle split lock issue in firmware after kernel enables #AC for split lock.

But we use a simple way to handle split lock which is suggested by Thomas Gleixner and Dave Hansen:

- If split lock happens in kernel, a warning is issued and #AC for split lock is disabled on the local CPU. The split lock issue should be fixed in kernel.

- If split lock happens in user process, the process is killed by SIGBUS. Unless the issue is fixed, the process cannot run in the system.

- If split lock happens in firmware, system may hang in firmware. The issue should be fixed in firmware.

- Enable split lock detection by default once the feature is enumerated.

- Disable split lock detection by "clearcpuid=split_lock_detect" during boot time.

- Disable/enable split lock detection by interface /sys/devices/system/cpu/split_lock_detect during run time.

==Expose to guest==

To expose this feature to guest, we need to
1. report the new CPUID bit to guest.
2. emulate IA32_CORE_CAPABILITIES MSR.
3. emulate TEST_CTL MSR. We must do it. Since this patch series enable split lock detection in host kernel by default, if we do not emualte TEST_CTL MSR for guest, guest will run with the value set by host without knowing that. So guest will run with split lock detection enabled due to the host's value. Thus guest running with buggy firmware and old kernel will fail because they lack the ability to handle #AC for split lock. So we should emulate TEST_CTL MSR and separate its value between host and guest.

==Patches==
Patch 1-4: Fix a few split lock issues.
Patch 5-7: Enumerate features.
Patch 8-12: Extend "clearcpuid=" option.
Patch 13: Handle #AC for split lock
Patch 14: /sys interface to enable/disable split lock detection Patch 15-17: Enable split lock detection in KVM.

==Changelog==
v4:
- Remove "setcpuid=" option
- Enable IA32_CORE_CAPABILITY enumeration for split lock
- Handle CPUID faulting by Peter Zijlstra
- Enable /sys interface to enable/disable split lock detection

v3:
- Handle split lock as suggested by Thomas Gleixner.
- Fix a few potential spit lock issues suggested by Thomas Gleixner.
- Support kernel option "setcpuid=" suggested by Dave Hanson and Thomas Gleixner.
- Support flag string in "clearcpuid=" suggested by Dave Hanson and Thomas Gleixner.

v2:
- Remove code that handles split lock issue in firmware and fix x86_capability issue mainly based on comments from Thomas Gleixner and Peter Zijlstra.

In previous version:
Comments from Dave Hansen:
- Enumerate feature in X86_FEATURE_SPLIT_LOCK_AC
- Separate #AC handler from do_error_trap
- Use CONFIG to configure inherit BIOS setting, enable, or disable split lock. Remove kernel parameter "split_lock_ac="
- Change config interface to debugfs from sysfs
- Fix a few bisectable issues
- Other changes.

Comment from Tony Luck and Dave Hansen:
- Dump right information in #AC handler

Comment from Alan Cox and Dave Hansen:
- Description of split lock in patch 0

Others:
- Remove tracing because we can trace split lock in existing sq_misc.split_lock.
- Add CONFIG to configure either panic or re-execute faulting instruction for split lock in kernel.
- other minor changes.

Fenghua Yu (13):
x86/common: Align cpu_caps_cleared and cpu_caps_set to unsigned long drivers/net/b44: Align pwol_mask to unsigned long for better
performance
wlcore: Align reg_ch_conf_pending and tmp_ch_bitmap to unsigned long
for better performance
x86/split_lock: Align x86_capability to unsigned long to avoid split
locked access
x86/cpufeatures: Enumerate IA32_CORE_CAPABILITIES MSR x86/msr-index: Define IA32_CORE_CAPABILITY MSR and #AC exception for
split lock bit
x86/split_lock: Enumerate #AC for split lock by MSR
IA32_CORE_CAPABILITY
x86/clearcpuid: Support multiple clearcpuid options x86/clearcpuid: Support feature flag string in kernel option
clearcpuid
x86/clearcpuid: Apply cleared feature bits that are forced set before Change document for kernel option clearcpuid x86/split_lock: Handle #AC exception for split lock x86/split_lock: Add a sysfs interface to allow user to enable or
disable split lock detection on all CPUs during run time

Peter Zijlstra (1):
x86/clearcpuid: Handle clearcpuid faulting

Xiaoyao Li (3):
kvm: x86: Report CORE_CAPABILITY on GET_SUPPORTED_CPUID
kvm: x86: Add support IA32_CORE_CAPABILITY MSR
kvm: vmx: Emulate TEST_CTL MSR

 Documentation/admin-guide/kernel-parameters.txt |   6 +-
 arch/x86/include/asm/cmdline.h                  |   3 +
 arch/x86/include/asm/cpu.h                      |   7 +
 arch/x86/include/asm/cpufeature.h               |   5 +
 arch/x86/include/asm/cpufeatures.h              |   2 +
 arch/x86/include/asm/kvm_host.h                 |   1 +
 arch/x86/include/asm/msr-index.h                |   7 +
 arch/x86/include/asm/processor.h                |   4 +-
 arch/x86/kernel/cpu/common.c                    |  13 +-
 arch/x86/kernel/cpu/cpuid-deps.c                |  79 +++++++++++
 arch/x86/kernel/cpu/intel.c                     | 170 +++++++++++++++++++++++-
 arch/x86/kernel/cpu/scattered.c                 |  17 +++
 arch/x86/kernel/fpu/init.c                      |  33 +++--
 arch/x86/kernel/process.c                       |   3 +
 arch/x86/kernel/traps.c                         |  42 ++++++
 arch/x86/kvm/cpuid.c                            |   3 +-
 arch/x86/kvm/vmx/vmx.c                          |  58 ++++++++
 arch/x86/kvm/vmx/vmx.h                          |   2 +
 arch/x86/kvm/x86.c                              |  17 ++-
 arch/x86/lib/cmdline.c                          |  17 ++-
 drivers/net/ethernet/broadcom/b44.c             |   3 +-
 drivers/net/wireless/ti/wlcore/cmd.c            |   3 +-
 drivers/net/wireless/ti/wlcore/wlcore.h         |   6 +-
 23 files changed, 473 insertions(+), 28 deletions(-)
Signature · 1 line
2.7.4
author
#1 lore raw

PATCH v4 1/17 x86/common: Align cpu_caps_cleared and cpu_caps_set to unsigned long

cpu_caps_cleared and cpu_caps_set may not be aligned to unsigned long. Atomic operations (i.e. set_bit and clear_bit) on the bitmaps may access two cache lines (a.k.a. split lock) and lock bus to block all memory accesses from other processors to ensure atomicity.

To avoid the overall performance degradation from the bus locking, align the two variables to unsigned long.

Defining the variables as unsigned long may also fix the issue because they are naturally aligned to unsigned long. But that needs additional code changes. Adding __aligned(unsigned long) are simpler fixes.

---
 arch/x86/kernel/cpu/common.c | 5 +++--
 1 file changed, 3 insertions(+), 2 deletions(-)
arch/x86/kernel/cpu/common.c +3 -2
index cb28e98a0659..51ab37ba5f64 100644
@@ -488,8 +488,9 @@ static const char *table_lookup_model(struct cpuinfo_x86 *c)
 	return NULL;		/* Not found */
 }
 
-__u32 cpu_caps_cleared[NCAPINTS + NBUGINTS];
-__u32 cpu_caps_set[NCAPINTS + NBUGINTS];
+/* Unsigned long alignment to avoid split lock in atomic bitmap ops */
+__u32 cpu_caps_cleared[NCAPINTS + NBUGINTS] __aligned(sizeof(unsigned long));
+__u32 cpu_caps_set[NCAPINTS + NBUGINTS] __aligned(sizeof(unsigned long));
 
 void load_percpu_segment(int cpu)
 {
Signature · 1 line
2.7.4
maintainer
#2 lore raw

Re: [PATCH v4 01/17] x86/common: Align cpu_caps_cleared and cpu_caps_set to unsigned long

Quoted 24 lines · Fenghua Yu

cpu_caps_cleared and cpu_caps_set may not be aligned to unsigned long. Atomic operations (i.e. set_bit and clear_bit) on the bitmaps may access two cache lines (a.k.a. split lock) and lock bus to block all memory accesses from other processors to ensure atomicity.

To avoid the overall performance degradation from the bus locking, align the two variables to unsigned long.

Defining the variables as unsigned long may also fix the issue because they are naturally aligned to unsigned long. But that needs additional code changes. Adding __aligned(unsigned long) are simpler fixes.

---
 arch/x86/kernel/cpu/common.c | 5 +++--
 1 file changed, 3 insertions(+), 2 deletions(-)
arch/x86/kernel/cpu/common.c +0 -0
index cb28e98a0659..51ab37ba5f64 100644
@@ -488,8 +488,9 @@ static const char *table_lookup_model(struct cpuinfo_x86 *c)
 	return NULL;		/* Not found */

arch/x86/kernel/cpu/common.c static const char *table_lookup_model(struct cpuinfo_x86 *c) · line 496

Quoted 10 lines
arch/x86/kernel/cpu/common.c +3 -2
 }
 
-__u32 cpu_caps_cleared[NCAPINTS + NBUGINTS];
-__u32 cpu_caps_set[NCAPINTS + NBUGINTS];
+/* Unsigned long alignment to avoid split lock in atomic bitmap ops */
+__u32 cpu_caps_cleared[NCAPINTS + NBUGINTS] __aligned(sizeof(unsigned long));
+__u32 cpu_caps_set[NCAPINTS + NBUGINTS] __aligned(sizeof(unsigned long));
 
 void load_percpu_segment(int cpu)
 {

(resending including the list)

Why not instead change set_bit/clear_bit to use btsl/btrl instead of btsq/btrq?

Thanks,

Paolo

#3 lore raw

Re: [PATCH v4 01/17] x86/common: Align cpu_caps_cleared and cpu_caps_set to unsigned long

Quoted 2 lines · Paolo Bonzini

Why not instead change set_bit/clear_bit to use btsl/btrl instead of btsq/btrq?

At least one of the faulty users (wireless) is in generic code and needs fixing regardless.

For better or worse; the bitmap stuff is defined to work on unsigned long. Using it on smaller types already relies on small endian; but further enabling just makes it worse I feel. Better have the rules be uniform.

maintainer
#4 lore raw

Re: [PATCH v4 01/17] x86/common: Align cpu_caps_cleared and cpu_caps_set to unsigned long

Quoted 11 lines · Peter Zijlstra
Quoted 2 lines · Paolo Bonzini

Why not instead change set_bit/clear_bit to use btsl/btrl instead of btsq/btrq?

At least one of the faulty users (wireless) is in generic code and needs fixing regardless.

For better or worse; the bitmap stuff is defined to work on unsigned long. Using it on smaller types already relies on small endian; but further enabling just makes it worse I feel. Better have the rules be uniform.

True that. On the other hand btsl/btrl is also one byte smaller if no operand is %r8-%r15.

In any case, /me wonders if we should have a macro like

#define DECLARE_LE_BITMAP(name,bits) \
        u32 name[DIV_ROUND_UP(bits, 32)] __aligned(sizeof(unsigned long))

Paolo

#5 lore raw

Re: [PATCH v4 01/17] x86/common: Align cpu_caps_cleared and cpu_caps_set to unsigned long

Quoted 2 lines · Paolo Bonzini

True that. On the other hand btsl/btrl is also one byte smaller if no operand is %r8-%r15.

Because then we loose the REX prefix, right. Now _that_ might actually be a reason to do that :-)

Quoted 4 lines

In any case, /me wonders if we should have a macro like

#define DECLARE_LE_BITMAP(name,bits) \
        u32 name[DIV_ROUND_UP(bits, 32)] __aligned(sizeof(unsigned long))

s/u32/__le32/

To go in bitops/le.h, sure, if there's enough users.

maintainer
#6 lore raw

Re: [PATCH v4 01/17] x86/common: Align cpu_caps_cleared and cpu_caps_set to unsigned long

Quoted 6 lines · Peter Zijlstra
Quoted 2 lines · Paolo Bonzini

True that. On the other hand btsl/btrl is also one byte smaller if no operand is %r8-%r15.

Because then we loose the REX prefix, right. Now _that_ might actually be a reason to do that :-)

I knew that would be the right way to put it for you. :)

Quoted 8 lines
Quoted 4 lines

In any case, /me wonders if we should have a macro like

#define DECLARE_LE_BITMAP(name,bits) \
        u32 name[DIV_ROUND_UP(bits, 32)] __aligned(sizeof(unsigned long))

s/u32/__le32/

To go in bitops/le.h, sure, if there's enough users.

Hmm... actually that should be "BITS_TO_LONGS(bits) * sizeof(unsigned long) / 4" because bitmap functions may access (or even clear) the last word as 64 bits.

Paolo

author
#7 lore raw

PATCH v4 2/17 drivers/net/b44: Align pwol_mask to unsigned long for better performance

In reply to Fenghua Yu

A bit in pwol_mask is set in b44_magic_pattern automatically by set_bit. set_bit sets the bit in a single unsigned long location. Since pwol_mask may not be aligned to unsigned long, the location may cross two cache lines and accessing the location degradates performance. On x86, accessing two cache lines in locked instruction in set_bit is called split lock and can cause overall performance degradation.

To avoid to impact performance by accessing two cache lines in set_bit, align pwol_mask to unsigned long.

---
 drivers/net/ethernet/broadcom/b44.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)
drivers/net/ethernet/broadcom/b44.c +2 -1
index 97ab0dd25552..bc544b6b9c3a 100644
@@ -1547,7 +1547,8 @@ static void b44_setup_pseudo_magicp(struct b44 *bp)
 	u32 val;
 	int plen0, plen1, plen2;
 	u8 *pwol_pattern;
-	u8 pwol_mask[B44_PMASK_SIZE];
+	/* Align to unsigned long for better performance in set_bit() */
+	u8 pwol_mask[B44_PMASK_SIZE] __aligned(sizeof(unsigned long));
 
 	pwol_pattern = kzalloc(B44_PATTERN_SIZE, GFP_KERNEL);
 	if (!pwol_pattern)
Signature · 1 line
2.7.4
#8 lore raw

Re: [PATCH v4 02/17] drivers/net/b44: Align pwol_mask to unsigned long for better performance

Quoted 20 lines · Fenghua Yu

A bit in pwol_mask is set in b44_magic_pattern automatically by set_bit. set_bit sets the bit in a single unsigned long location. Since pwol_mask may not be aligned to unsigned long, the location may cross two cache lines and accessing the location degradates performance. On x86, accessing two cache lines in locked instruction in set_bit is called split lock and can cause overall performance degradation.

To avoid to impact performance by accessing two cache lines in set_bit, align pwol_mask to unsigned long.

---
 drivers/net/ethernet/broadcom/b44.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)
drivers/net/ethernet/broadcom/b44.c +0 -0
index 97ab0dd25552..bc544b6b9c3a 100644

drivers/net/ethernet/broadcom/b44.c static void b44_setup_pseudo_magicp(struct b44 *bp) · line 1554

Quoted 9 lines
drivers/net/ethernet/broadcom/b44.c +2 -1
@@ -1547,7 +1547,8 @@ static void b44_setup_pseudo_magicp(struct b44 *bp)
 	u32 val;
 	int plen0, plen1, plen2;
 	u8 *pwol_pattern;
-	u8 pwol_mask[B44_PMASK_SIZE];
+	/* Align to unsigned long for better performance in set_bit() */
+	u8 pwol_mask[B44_PMASK_SIZE] __aligned(sizeof(unsigned long));
 
 	pwol_pattern = kzalloc(B44_PATTERN_SIZE, GFP_KERNEL);
 	if (!pwol_pattern)

That is truly horrid code. But afaict pwol_mask is local and never exposed to concurrency, so _why_ does it need atomic bitset in the first place?

Would not the below be a _much_ better solution?

drivers/net/ethernet/broadcom/b44.c +2 -2
index 97ab0dd25552..0b4226b406b1 100644
@@ -1520,7 +1520,7 @@ static int b44_magic_pattern(u8 *macaddr, u8 *ppattern, u8 *pmask, int offset)
 
 	memset(ppattern + offset, 0xff, magicsync);
 	for (j = 0; j < magicsync; j++)
-		set_bit(len++, (unsigned long *) pmask);
+		__set_bit(len++, (unsigned long *) pmask);
 
 	for (j = 0; j < B44_MAX_PATTERNS; j++) {
 		if ((B44_PATTERN_SIZE - len) >= ETH_ALEN)
@@ -1532,7 +1532,7 @@ static int b44_magic_pattern(u8 *macaddr, u8 *ppattern, u8 *pmask, int offset)
 		for (k = 0; k< ethaddr_bytes; k++) {
 			ppattern[offset + magicsync +
 				(j * ETH_ALEN) + k] = macaddr[k];
-			set_bit(len++, (unsigned long *) pmask);
+			__set_bit(len++, (unsigned long *) pmask);
 		}
 	}
 	return len - 1;
author
#9 lore raw

Re: [PATCH v4 02/17] drivers/net/b44: Align pwol_mask to unsigned long for better performance

Quoted 52 lines · Peter Zijlstra
Quoted 20 lines · Fenghua Yu

A bit in pwol_mask is set in b44_magic_pattern automatically by set_bit. set_bit sets the bit in a single unsigned long location. Since pwol_mask may not be aligned to unsigned long, the location may cross two cache lines and accessing the location degradates performance. On x86, accessing two cache lines in locked instruction in set_bit is called split lock and can cause overall performance degradation.

To avoid to impact performance by accessing two cache lines in set_bit, align pwol_mask to unsigned long.

---
 drivers/net/ethernet/broadcom/b44.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)
drivers/net/ethernet/broadcom/b44.c +0 -0
index 97ab0dd25552..bc544b6b9c3a 100644

drivers/net/ethernet/broadcom/b44.c static void b44_setup_pseudo_magicp(struct b44 *bp) · line 1554

Quoted 9 lines
drivers/net/ethernet/broadcom/b44.c +2 -1
@@ -1547,7 +1547,8 @@ static void b44_setup_pseudo_magicp(struct b44 *bp)
 	u32 val;
 	int plen0, plen1, plen2;
 	u8 *pwol_pattern;
-	u8 pwol_mask[B44_PMASK_SIZE];
+	/* Align to unsigned long for better performance in set_bit() */
+	u8 pwol_mask[B44_PMASK_SIZE] __aligned(sizeof(unsigned long));
 
 	pwol_pattern = kzalloc(B44_PATTERN_SIZE, GFP_KERNEL);
 	if (!pwol_pattern)

That is truly horrid code. But afaict pwol_mask is local and never exposed to concurrency, so _why_ does it need atomic bitset in the first place?

Would not the below be a _much_ better solution?

drivers/net/ethernet/broadcom/b44.c +1 -1
index 97ab0dd25552..0b4226b406b1 100644
@@ -1520,7 +1520,7 @@ static int b44_magic_pattern(u8 *macaddr, u8 *ppattern, u8 *pmask, int offset)
 
 	memset(ppattern + offset, 0xff, magicsync);
 	for (j = 0; j < magicsync; j++)
-		set_bit(len++, (unsigned long *) pmask);
+		__set_bit(len++, (unsigned long *) pmask);
 
 	for (j = 0; j < B44_MAX_PATTERNS; j++) {
 		if ((B44_PATTERN_SIZE - len) >= ETH_ALEN)

drivers/net/ethernet/broadcom/b44.c static int b44_magic_pattern(u8 *macaddr, u8 *ppattern, u8 *pmask, int offset) · line 1538

Quoted 8 lines
drivers/net/ethernet/broadcom/b44.c +1 -1
@@ -1532,7 +1532,7 @@ static int b44_magic_pattern(u8 *macaddr, u8 *ppattern, u8 *pmask, int offset)
 		for (k = 0; k< ethaddr_bytes; k++) {
 			ppattern[offset + magicsync +
 				(j * ETH_ALEN) + k] = macaddr[k];
-			set_bit(len++, (unsigned long *) pmask);
+			__set_bit(len++, (unsigned long *) pmask);
 		}
 	}
 	return len - 1;

Sure. I will change the patch to your code. And add

to this patch?

Thanks.

-Fenghua

author
#11 lore raw

PATCH v4 3/17 wlcore: Align reg_ch_conf_pending and tmp_ch_bitmap to unsigned long for better performance

In reply to Fenghua Yu

A bit in reg_ch_conf_pending in wl271 and tmp_ch_bitmap is set atomically by set_bit(). set_bit() sets the bit in a single unsigned long location. If the variables are not aligned to unsigned long, set_bit() accesses two cache lines and thus causes slower performance. On x86, this scenario is called split lock and can cause overall performance degradation due to locked BTSL instruction in set_bit() locks bus.

To avoid performance degradation, the two variables are aligned to unsigned long.

---
 drivers/net/wireless/ti/wlcore/cmd.c    | 3 ++-
 drivers/net/wireless/ti/wlcore/wlcore.h | 6 ++++--
 2 files changed, 6 insertions(+), 3 deletions(-)
drivers/net/wireless/ti/wlcore/cmd.c +2 -1
index 903968735a74..8d15a6307d44 100644
@@ -1707,7 +1707,8 @@ int wlcore_cmd_regdomain_config_locked(struct wl1271 *wl)
 {
 	struct wl12xx_cmd_regdomain_dfs_config *cmd = NULL;
 	int ret = 0, i, b, ch_bit_idx;
-	u32 tmp_ch_bitmap[2];
+	/* Align to unsigned long for better performance in set_bit() */
+	u32 tmp_ch_bitmap[2] __aligned(sizeof(unsigned long));
 	struct wiphy *wiphy = wl->hw->wiphy;
 	struct ieee80211_supported_band *band;
 	bool timeout = false;
drivers/net/wireless/ti/wlcore/wlcore.h +4 -2
index dd14850b0603..92d878f01fa5 100644
@@ -321,8 +321,10 @@ struct wl1271 {
 
 	/* Reg domain last configuration */
 	u32 reg_ch_conf_last[2]  __aligned(8);
-	/* Reg domain pending configuration */
-	u32 reg_ch_conf_pending[2];
+	/* Reg domain pending configuration. Aligned to unsigned long for
+	 * better performane in set_bit().
+	 */
+	u32 reg_ch_conf_pending[2] __aligned(sizeof(unsigned long));
 
 	/* Pointer that holds DMA-friendly block for the mailbox */
 	void *mbox;
Signature · 1 line
2.7.4
#12 lore raw

Re: [PATCH v4 03/17] wlcore: Align reg_ch_conf_pending and tmp_ch_bitmap to unsigned long for better performance

Quoted 37 lines · Fenghua Yu

A bit in reg_ch_conf_pending in wl271 and tmp_ch_bitmap is set atomically by set_bit(). set_bit() sets the bit in a single unsigned long location. If the variables are not aligned to unsigned long, set_bit() accesses two cache lines and thus causes slower performance. On x86, this scenario is called split lock and can cause overall performance degradation due to locked BTSL instruction in set_bit() locks bus.

To avoid performance degradation, the two variables are aligned to unsigned long.

---
 drivers/net/wireless/ti/wlcore/cmd.c    | 3 ++-
 drivers/net/wireless/ti/wlcore/wlcore.h | 6 ++++--
 2 files changed, 6 insertions(+), 3 deletions(-)
drivers/net/wireless/ti/wlcore/cmd.c +2 -1
index 903968735a74..8d15a6307d44 100644
@@ -1707,7 +1707,8 @@ int wlcore_cmd_regdomain_config_locked(struct wl1271 *wl)
 {
 	struct wl12xx_cmd_regdomain_dfs_config *cmd = NULL;
 	int ret = 0, i, b, ch_bit_idx;
-	u32 tmp_ch_bitmap[2];
+	/* Align to unsigned long for better performance in set_bit() */
+	u32 tmp_ch_bitmap[2] __aligned(sizeof(unsigned long));
 	struct wiphy *wiphy = wl->hw->wiphy;
 	struct ieee80211_supported_band *band;
 	bool timeout = false;
drivers/net/wireless/ti/wlcore/wlcore.h +0 -0
index dd14850b0603..92d878f01fa5 100644
@@ -321,8 +321,10 @@ struct wl1271 {
 
 	/* Reg domain last configuration */

drivers/net/wireless/ti/wlcore/wlcore.h struct wl1271 { · line 330

Quoted 10 lines
drivers/net/wireless/ti/wlcore/wlcore.h +4 -2
 	u32 reg_ch_conf_last[2]  __aligned(8);
-	/* Reg domain pending configuration */
-	u32 reg_ch_conf_pending[2];
+	/* Reg domain pending configuration. Aligned to unsigned long for
+	 * better performane in set_bit().
+	 */
+	u32 reg_ch_conf_pending[2] __aligned(sizeof(unsigned long));
 
 	/* Pointer that holds DMA-friendly block for the mailbox */
 	void *mbox;

This has nothing to do with better performance. This is generic code, not x86 arch code. Many RISC platforms will already refuse unaligned atomic ops.

Also, how is this set_bit() usage endian safe?

And no wireless person on Cc anywhere.

maintainer
#13 lore raw

Re: [PATCH v4 03/17] wlcore: Align reg_ch_conf_pending and tmp_ch_bitmap to unsigned long for better performance

Quoted 17 lines · Peter Zijlstra
Quoted 16 lines · Fenghua Yu

A bit in reg_ch_conf_pending in wl271 and tmp_ch_bitmap is set atomically by set_bit(). set_bit() sets the bit in a single unsigned long location. If the variables are not aligned to unsigned long, set_bit() accesses two cache lines and thus causes slower performance. On x86, this scenario is called split lock and can cause overall performance degradation due to locked BTSL instruction in set_bit() locks bus.

To avoid performance degradation, the two variables are aligned to
unsigned long.

drivers/net/wireless/ti/wlcore/cmd.c | 3 ++-
drivers/net/wireless/ti/wlcore/wlcore.h | 6 ++++-- 2 files changed, 6 insertions(+), 3 deletions(-)

Quoted 10 lines
Quoted 10 lines
(quoted diff) +0 -0

b/drivers/net/wireless/ti/wlcore/cmd.c index 903968735a74..8d15a6307d44 100644 --- a/drivers/net/wireless/ti/wlcore/cmd.c +++ b/drivers/net/wireless/ti/wlcore/cmd.c @@ -1707,7 +1707,8 @@ int wlcore_cmd_regdomain_config_locked(struct wl1271 *wl) { struct wl12xx_cmd_regdomain_dfs_config *cmd = NULL; int ret = 0, i, b,
ch_bit_idx; - u32 tmp_ch_bitmap[2]; + /* Align to unsigned long
for better performance in set_bit() */ + u32 tmp_ch_bitmap[2]
__aligned(sizeof(unsigned long));

This is the only place where an array of u32 is needed, because of

        cmd->ch_bit_map1 = cpu_to_le32(tmp_ch_bitmap[0]);
        cmd->ch_bit_map2 = cpu_to_le32(tmp_ch_bitmap[1]);

All the others should use DECLARE_BITMAP, including reg_ch_conf_last which was already using __aligned. As Peter mentioned they should also use set_bit_le. Actually they do not need locked access at all because the only code paths to the set_bit take a mutex.

There is one other place that is accessing the items of the array, but it is fixed easily. The following patch should do everything:

------------------- 8< --------------------------
From: Paolo Bonzini <[email protected]>
Subject: [PATCH] wlcore: simplify/fix/optimize reg_ch_conf_pending operations

Bitmaps are defined on unsigned longs, so the usage of u32[2] in the wlcore driver is incorrect. As noted by Peter Zijlstra, casting arrays to a bitmap is incorrect for big-endian architectures.

When looking at it I observed that:

- operations on reg_ch_conf_pending is always under the wl_lock mutex, so set_bit is overkill

- the only case where reg_ch_conf_pending is accessed a u32 at a time is unnecessary too.

This patch cleans up everything in this area, and changes tmp_ch_bitmap to have the proper alignment.

drivers/net/wireless/ti/wlcore/cmd.c +4 -7
index 903968735a74..3e093f3a7ec8 100644
@@ -1700,14 +1700,14 @@ void wlcore_set_pending_regdomain_ch(struct wl1271 *wl, u16 channel,
 	ch_bit_idx = wlcore_get_reg_conf_ch_idx(band, channel);
 
 	if (ch_bit_idx >= 0 && ch_bit_idx <= WL1271_MAX_CHANNELS)
-		set_bit(ch_bit_idx, (long *)wl->reg_ch_conf_pending);
+		__set_bit_le(ch_bit_idx, (long *)wl->reg_ch_conf_pending);
 }
 
 int wlcore_cmd_regdomain_config_locked(struct wl1271 *wl)
 {
 	struct wl12xx_cmd_regdomain_dfs_config *cmd = NULL;
 	int ret = 0, i, b, ch_bit_idx;
-	u32 tmp_ch_bitmap[2];
+	u32 tmp_ch_bitmap[2] __aligned(sizeof(unsigned long));
 	struct wiphy *wiphy = wl->hw->wiphy;
 	struct ieee80211_supported_band *band;
 	bool timeout = false;
@@ -1717,7 +1717,7 @@ int wlcore_cmd_regdomain_config_locked(struct wl1271 *wl)
 
 	wl1271_debug(DEBUG_CMD, "cmd reg domain config");
 
-	memset(tmp_ch_bitmap, 0, sizeof(tmp_ch_bitmap));
+	memcpy(tmp_ch_bitmap, wl->reg_ch_conf_pending, sizeof(tmp_ch_bitmap));
 
 	for (b = NL80211_BAND_2GHZ; b <= NL80211_BAND_5GHZ; b++) {
 		band = wiphy->bands[b];
@@ -1738,13 +1738,10 @@ int wlcore_cmd_regdomain_config_locked(struct wl1271 *wl)
 			if (ch_bit_idx < 0)
 				continue;
 
-			set_bit(ch_bit_idx, (long *)tmp_ch_bitmap);
+			__set_bit_le(ch_bit_idx, (long *)tmp_ch_bitmap);
 		}
 	}
 
-	tmp_ch_bitmap[0] |= wl->reg_ch_conf_pending[0];
-	tmp_ch_bitmap[1] |= wl->reg_ch_conf_pending[1];
-
 	if (!memcmp(tmp_ch_bitmap, wl->reg_ch_conf_last, sizeof(tmp_ch_bitmap)))
 		goto out;
 
drivers/net/wireless/ti/wlcore/wlcore.h +2 -2
index dd14850b0603..870eea3e7a27 100644
@@ -320,9 +320,9 @@ struct wl1271 {
 	bool watchdog_recovery;
 
 	/* Reg domain last configuration */
-	u32 reg_ch_conf_last[2]  __aligned(8);
+	DECLARE_BITMAP(reg_ch_conf_last, 64);
 	/* Reg domain pending configuration */
-	u32 reg_ch_conf_pending[2];
+	DECLARE_BITMAP(reg_ch_conf_pending, 64);
 
 	/* Pointer that holds DMA-friendly block for the mailbox */
 	void *mbox;
#14 lore raw

Re: [PATCH v4 03/17] wlcore: Align reg_ch_conf_pending and tmp_ch_bitmap to unsigned long for better performance

Quoted 29 lines · Paolo Bonzini

From: Paolo Bonzini <[email protected]>
Subject: [PATCH] wlcore: simplify/fix/optimize reg_ch_conf_pending operations

Bitmaps are defined on unsigned longs, so the usage of u32[2] in the wlcore driver is incorrect. As noted by Peter Zijlstra, casting arrays to a bitmap is incorrect for big-endian architectures.

When looking at it I observed that:

- operations on reg_ch_conf_pending is always under the wl_lock mutex, so set_bit is overkill

- the only case where reg_ch_conf_pending is accessed a u32 at a time is unnecessary too.

This patch cleans up everything in this area, and changes tmp_ch_bitmap to have the proper alignment.

drivers/net/wireless/ti/wlcore/cmd.c +0 -0
index 903968735a74..3e093f3a7ec8 100644
@@ -1700,14 +1700,14 @@ void wlcore_set_pending_regdomain_ch(struct wl1271 *wl, u16 channel,
 	ch_bit_idx = wlcore_get_reg_conf_ch_idx(band, channel);
 
 	if (ch_bit_idx >= 0 && ch_bit_idx <= WL1271_MAX_CHANNELS)

drivers/net/wireless/ti/wlcore/cmd.c void wlcore_set_pending_regdomain_ch(struct wl1271 *wl, u16 channel, · line 1710

Quoted 10 lines
drivers/net/wireless/ti/wlcore/cmd.c +2 -2
-		set_bit(ch_bit_idx, (long *)wl->reg_ch_conf_pending);
+		__set_bit_le(ch_bit_idx, (long *)wl->reg_ch_conf_pending);
 }
 
 int wlcore_cmd_regdomain_config_locked(struct wl1271 *wl)
 {
 	struct wl12xx_cmd_regdomain_dfs_config *cmd = NULL;
 	int ret = 0, i, b, ch_bit_idx;
-	u32 tmp_ch_bitmap[2];
+	u32 tmp_ch_bitmap[2] __aligned(sizeof(unsigned long));

Also mark it as __le32 ?

drivers/net/wireless/ti/wlcore/cmd.c int wlcore_cmd_regdomain_config_locked(struct wl1271 *wl) · line 1720

Quoted 9 lines
drivers/net/wireless/ti/wlcore/cmd.c +1 -1
 	struct wiphy *wiphy = wl->hw->wiphy;
 	struct ieee80211_supported_band *band;
 	bool timeout = false;
@@ -1717,7 +1717,7 @@ int wlcore_cmd_regdomain_config_locked(struct wl1271 *wl)
 
 	wl1271_debug(DEBUG_CMD, "cmd reg domain config");
 
-	memset(tmp_ch_bitmap, 0, sizeof(tmp_ch_bitmap));
+	memcpy(tmp_ch_bitmap, wl->reg_ch_conf_pending, sizeof(tmp_ch_bitmap));

How about using:

	bitmap_to_arr32(tmp_ch_bitmap, wl->reg_ch_conf_pending, sizeof(tmp_ch_bitmap));
	for (i=0; i<2; i++)
		tmp_ch_bitmap[i] = cpu_to_le32(tmp_ch_bitmap[i]);

(or add bitmap_to_arr32_le ?)

drivers/net/wireless/ti/wlcore/cmd.c int wlcore_cmd_regdomain_config_locked(struct wl1271 *wl) · line 1741

Quoted 8 lines
drivers/net/wireless/ti/wlcore/cmd.c +1 -1
 	for (b = NL80211_BAND_2GHZ; b <= NL80211_BAND_5GHZ; b++) {
 		band = wiphy->bands[b];
@@ -1738,13 +1738,10 @@ int wlcore_cmd_regdomain_config_locked(struct wl1271 *wl)
 			if (ch_bit_idx < 0)
 				continue;
 
-			set_bit(ch_bit_idx, (long *)tmp_ch_bitmap);
+			__set_bit_le(ch_bit_idx, (long *)tmp_ch_bitmap);

But you copied in reg_ch_conf_pending without doing an LE swizzle. With the proposed change, we have two __le32 here and it works again.

drivers/net/wireless/ti/wlcore/cmd.c int wlcore_cmd_regdomain_config_locked(struct wl1271 *wl) · line 1747

Quoted 9 lines
drivers/net/wireless/ti/wlcore/cmd.c +0 -3
 		}
 	}
 
-	tmp_ch_bitmap[0] |= wl->reg_ch_conf_pending[0];
-	tmp_ch_bitmap[1] |= wl->reg_ch_conf_pending[1];
-
 	if (!memcmp(tmp_ch_bitmap, wl->reg_ch_conf_last, sizeof(tmp_ch_bitmap)))
 		goto out;
 

And then remove the cpu_to_le32() on assignment to ch_bit_map*.

drivers/net/wireless/ti/wlcore/wlcore.h struct wl1271 { · line 323

Quoted 10 lines
drivers/net/wireless/ti/wlcore/wlcore.h +1 -1
index dd14850b0603..870eea3e7a27 100644
@@ -320,9 +320,9 @@ struct wl1271 {
 	bool watchdog_recovery;
 
 	/* Reg domain last configuration */
-	u32 reg_ch_conf_last[2]  __aligned(8);
+	DECLARE_BITMAP(reg_ch_conf_last, 64);

Is never actually used as a bitmap but used as opaque storage with memcpy and memcmp against tmp_ch_bitmap.

Quoted 6 lines
drivers/net/wireless/ti/wlcore/wlcore.h +1 -1
 	/* Reg domain pending configuration */
-	u32 reg_ch_conf_pending[2];
+	DECLARE_BITMAP(reg_ch_conf_pending, 64);
 
 	/* Pointer that holds DMA-friendly block for the mailbox */
 	void *mbox;
maintainer
#15 lore raw

Re: [PATCH v4 03/17] wlcore: Align reg_ch_conf_pending and tmp_ch_bitmap to unsigned long for better performance

Quoted 34 lines · Peter Zijlstra
Quoted 33 lines · Paolo Bonzini

From: Paolo Bonzini <[email protected]>
Subject: [PATCH] wlcore: simplify/fix/optimize reg_ch_conf_pending operations

Bitmaps are defined on unsigned longs, so the usage of u32[2] in the wlcore driver is incorrect. As noted by Peter Zijlstra, casting arrays to a bitmap is incorrect for big-endian architectures.

When looking at it I observed that:

- operations on reg_ch_conf_pending is always under the wl_lock mutex, so set_bit is overkill

- the only case where reg_ch_conf_pending is accessed a u32 at a time is unnecessary too.

This patch cleans up everything in this area, and changes tmp_ch_bitmap to have the proper alignment.

drivers/net/wireless/ti/wlcore/cmd.c +1 -1
index 903968735a74..3e093f3a7ec8 100644
@@ -1700,14 +1700,14 @@ void wlcore_set_pending_regdomain_ch(struct wl1271 *wl, u16 channel,
 	ch_bit_idx = wlcore_get_reg_conf_ch_idx(band, channel);
 
 	if (ch_bit_idx >= 0 && ch_bit_idx <= WL1271_MAX_CHANNELS)
-		set_bit(ch_bit_idx, (long *)wl->reg_ch_conf_pending);
+		__set_bit_le(ch_bit_idx, (long *)wl->reg_ch_conf_pending);
 }
 
Quoted 8 lines

drivers/net/wireless/ti/wlcore/cmd.c void wlcore_set_pending_regdomain_ch(struct wl1271 *wl, u16 channel, · line 1710

Quoted 6 lines
drivers/net/wireless/ti/wlcore/cmd.c +1 -1
 int wlcore_cmd_regdomain_config_locked(struct wl1271 *wl)
 {
 	struct wl12xx_cmd_regdomain_dfs_config *cmd = NULL;
 	int ret = 0, i, b, ch_bit_idx;
-	u32 tmp_ch_bitmap[2];
+	u32 tmp_ch_bitmap[2] __aligned(sizeof(unsigned long));

Also mark it as __le32 ?

That would require more changes to mark ch_bit_map1/ch_bit_map2 as __le32 (I think, I don't do much sparse), so I didn't do that.

Quoted 12 lines

drivers/net/wireless/ti/wlcore/cmd.c int wlcore_cmd_regdomain_config_locked(struct wl1271 *wl) · line 1720

Quoted 9 lines
drivers/net/wireless/ti/wlcore/cmd.c +1 -1
 	struct wiphy *wiphy = wl->hw->wiphy;
 	struct ieee80211_supported_band *band;
 	bool timeout = false;
@@ -1717,7 +1717,7 @@ int wlcore_cmd_regdomain_config_locked(struct wl1271 *wl)
 
 	wl1271_debug(DEBUG_CMD, "cmd reg domain config");
 
-	memset(tmp_ch_bitmap, 0, sizeof(tmp_ch_bitmap));
+	memcpy(tmp_ch_bitmap, wl->reg_ch_conf_pending, sizeof(tmp_ch_bitmap));

How about using:

Quoted 5 lines
	bitmap_to_arr32(tmp_ch_bitmap, wl->reg_ch_conf_pending, sizeof(tmp_ch_bitmap));
	for (i=0; i<2; i++)
		tmp_ch_bitmap[i] = cpu_to_le32(tmp_ch_bitmap[i]);

(or add bitmap_to_arr32_le ?)

I've used __set_bit_le when setting reg_ch_conf_pending so no need to swizzle here; OTOH bitmap_to_arr32 doesn't work here that swizzle already swaps halfwords.

Quoted 11 lines

drivers/net/wireless/ti/wlcore/cmd.c int wlcore_cmd_regdomain_config_locked(struct wl1271 *wl) · line 1741

Quoted 8 lines
drivers/net/wireless/ti/wlcore/cmd.c +1 -1
 	for (b = NL80211_BAND_2GHZ; b <= NL80211_BAND_5GHZ; b++) {
 		band = wiphy->bands[b];
@@ -1738,13 +1738,10 @@ int wlcore_cmd_regdomain_config_locked(struct wl1271 *wl)
 			if (ch_bit_idx < 0)
 				continue;
 
-			set_bit(ch_bit_idx, (long *)tmp_ch_bitmap);
+			__set_bit_le(ch_bit_idx, (long *)tmp_ch_bitmap);

But you copied in reg_ch_conf_pending without doing an LE swizzle. With the proposed change, we have two __le32 here and it works again.

(Again there's no need to do an LE swizzle because it's done in wlcore_set_pending_regdomain_ch).

Quoted 11 lines

drivers/net/wireless/ti/wlcore/cmd.c int wlcore_cmd_regdomain_config_locked(struct wl1271 *wl) · line 1747

Quoted 9 lines
drivers/net/wireless/ti/wlcore/cmd.c +0 -3
 		}
 	}
 
-	tmp_ch_bitmap[0] |= wl->reg_ch_conf_pending[0];
-	tmp_ch_bitmap[1] |= wl->reg_ch_conf_pending[1];
-
 	if (!memcmp(tmp_ch_bitmap, wl->reg_ch_conf_last, sizeof(tmp_ch_bitmap)))
 		goto out;
 

And then remove the cpu_to_le32() on assignment to ch_bit_map*.

Yup, I forgot to commit that cpu_to_le32 removal.

Quoted 7 lines
Quoted 7 lines
drivers/net/wireless/ti/wlcore/wlcore.h +0 -0
index dd14850b0603..870eea3e7a27 100644
@@ -320,9 +320,9 @@ struct wl1271 {
 	bool watchdog_recovery;
 
Quoted 6 lines

drivers/net/wireless/ti/wlcore/wlcore.h struct wl1271 { · line 323

Quoted 3 lines
drivers/net/wireless/ti/wlcore/wlcore.h +1 -1
 	/* Reg domain last configuration */
-	u32 reg_ch_conf_last[2]  __aligned(8);
+	DECLARE_BITMAP(reg_ch_conf_last, 64);

Is never actually used as a bitmap but used as opaque storage with memcpy and memcmp against tmp_ch_bitmap.

Yeah, but it is the easiest way to ensure it is the right size as reg_ch_conf_pending. The two are related, it makes sense to declare them the same.

Paolo

#16 lore raw

Re: [PATCH v4 03/17] wlcore: Align reg_ch_conf_pending and tmp_ch_bitmap to unsigned long for better performance

Quoted 11 lines · Paolo Bonzini
Quoted 8 lines

drivers/net/wireless/ti/wlcore/cmd.c void wlcore_set_pending_regdomain_ch(struct wl1271 *wl, u16 channel, · line 1705

Quoted 8 lines
drivers/net/wireless/ti/wlcore/cmd.c +1 -1
@@ -1700,14 +1700,14 @@ void wlcore_set_pending_regdomain_ch(struct wl1271 *wl, u16 channel,
 	ch_bit_idx = wlcore_get_reg_conf_ch_idx(band, channel);
 
 	if (ch_bit_idx >= 0 && ch_bit_idx <= WL1271_MAX_CHANNELS)
-		set_bit(ch_bit_idx, (long *)wl->reg_ch_conf_pending);
+		__set_bit_le(ch_bit_idx, (long *)wl->reg_ch_conf_pending);
 }
 

I've used __set_bit_le when setting reg_ch_conf_pending so no need to swizzle here;

Argh, reading hard. Yes, then it all works.

author
#17 lore raw

Re: [PATCH v4 03/17] wlcore: Align reg_ch_conf_pending and tmp_ch_bitmap to unsigned long for better performance

In reply to Paolo Bonzini
Quoted 72 lines · Paolo Bonzini
Quoted 17 lines · Peter Zijlstra
Quoted 16 lines · Fenghua Yu

A bit in reg_ch_conf_pending in wl271 and tmp_ch_bitmap is set atomically by set_bit(). set_bit() sets the bit in a single unsigned long location. If the variables are not aligned to unsigned long, set_bit() accesses two cache lines and thus causes slower performance. On x86, this scenario is called split lock and can cause overall performance degradation due to locked BTSL instruction in set_bit() locks bus.

To avoid performance degradation, the two variables are aligned to
unsigned long.

drivers/net/wireless/ti/wlcore/cmd.c | 3 ++-
drivers/net/wireless/ti/wlcore/wlcore.h | 6 ++++-- 2 files changed, 6 insertions(+), 3 deletions(-)

Quoted 10 lines
Quoted 10 lines
(quoted diff) +0 -0

b/drivers/net/wireless/ti/wlcore/cmd.c index 903968735a74..8d15a6307d44 100644 --- a/drivers/net/wireless/ti/wlcore/cmd.c +++ b/drivers/net/wireless/ti/wlcore/cmd.c @@ -1707,7 +1707,8 @@ int wlcore_cmd_regdomain_config_locked(struct wl1271 *wl) { struct wl12xx_cmd_regdomain_dfs_config *cmd = NULL; int ret = 0, i, b,
ch_bit_idx; - u32 tmp_ch_bitmap[2]; + /* Align to unsigned long
for better performance in set_bit() */ + u32 tmp_ch_bitmap[2]
__aligned(sizeof(unsigned long));

This is the only place where an array of u32 is needed, because of

        cmd->ch_bit_map1 = cpu_to_le32(tmp_ch_bitmap[0]);
        cmd->ch_bit_map2 = cpu_to_le32(tmp_ch_bitmap[1]);

All the others should use DECLARE_BITMAP, including reg_ch_conf_last which was already using __aligned. As Peter mentioned they should also use set_bit_le. Actually they do not need locked access at all because the only code paths to the set_bit take a mutex.

There is one other place that is accessing the items of the array, but it is fixed easily. The following patch should do everything:

------------------- 8< --------------------------
From: Paolo Bonzini <[email protected]>
Subject: [PATCH] wlcore: simplify/fix/optimize reg_ch_conf_pending operations

Bitmaps are defined on unsigned longs, so the usage of u32[2] in the wlcore driver is incorrect. As noted by Peter Zijlstra, casting arrays to a bitmap is incorrect for big-endian architectures.

When looking at it I observed that:

- operations on reg_ch_conf_pending is always under the wl_lock mutex, so set_bit is overkill

- the only case where reg_ch_conf_pending is accessed a u32 at a time is unnecessary too.

This patch cleans up everything in this area, and changes tmp_ch_bitmap to have the proper alignment.

drivers/net/wireless/ti/wlcore/cmd.c +0 -0
index 903968735a74..3e093f3a7ec8 100644
@@ -1700,14 +1700,14 @@ void wlcore_set_pending_regdomain_ch(struct wl1271 *wl, u16 channel,
 	ch_bit_idx = wlcore_get_reg_conf_ch_idx(band, channel);
 
 	if (ch_bit_idx >= 0 && ch_bit_idx <= WL1271_MAX_CHANNELS)

drivers/net/wireless/ti/wlcore/cmd.c void wlcore_set_pending_regdomain_ch(struct wl1271 *wl, u16 channel, · line 1710

Quoted 10 lines
drivers/net/wireless/ti/wlcore/cmd.c +2 -2
-		set_bit(ch_bit_idx, (long *)wl->reg_ch_conf_pending);
+		__set_bit_le(ch_bit_idx, (long *)wl->reg_ch_conf_pending);
 }
 
 int wlcore_cmd_regdomain_config_locked(struct wl1271 *wl)
 {
 	struct wl12xx_cmd_regdomain_dfs_config *cmd = NULL;
 	int ret = 0, i, b, ch_bit_idx;
-	u32 tmp_ch_bitmap[2];
+	u32 tmp_ch_bitmap[2] __aligned(sizeof(unsigned long));

Now __aligned() is unnecessary because __set_bit_le() handles tmp_ch_bitmap, right?

Quoted 13 lines
drivers/net/wireless/ti/wlcore/cmd.c +1 -1
 	struct wiphy *wiphy = wl->hw->wiphy;
 	struct ieee80211_supported_band *band;
 	bool timeout = false;
@@ -1717,7 +1717,7 @@ int wlcore_cmd_regdomain_config_locked(struct wl1271 *wl)
 
 	wl1271_debug(DEBUG_CMD, "cmd reg domain config");
 
-	memset(tmp_ch_bitmap, 0, sizeof(tmp_ch_bitmap));
+	memcpy(tmp_ch_bitmap, wl->reg_ch_conf_pending, sizeof(tmp_ch_bitmap));
 
 	for (b = NL80211_BAND_2GHZ; b <= NL80211_BAND_5GHZ; b++) {
 		band = wiphy->bands[b];

drivers/net/wireless/ti/wlcore/cmd.c int wlcore_cmd_regdomain_config_locked(struct wl1271 *wl) · line 1741

Quoted 5 lines
drivers/net/wireless/ti/wlcore/cmd.c +1 -1
@@ -1738,13 +1738,10 @@ int wlcore_cmd_regdomain_config_locked(struct wl1271 *wl)
 			if (ch_bit_idx < 0)
 				continue;
 
-			set_bit(ch_bit_idx, (long *)tmp_ch_bitmap);
+			__set_bit_le(ch_bit_idx, (long *)tmp_ch_bitmap);

Is __test_and_set_bit_le() more meaningful to avoid duplicate bit setting ?

Quoted 25 lines
drivers/net/wireless/ti/wlcore/cmd.c +0 -3
 		}
 	}
 
-	tmp_ch_bitmap[0] |= wl->reg_ch_conf_pending[0];
-	tmp_ch_bitmap[1] |= wl->reg_ch_conf_pending[1];
-
 	if (!memcmp(tmp_ch_bitmap, wl->reg_ch_conf_last, sizeof(tmp_ch_bitmap)))
 		goto out;
 
drivers/net/wireless/ti/wlcore/wlcore.h +2 -2
index dd14850b0603..870eea3e7a27 100644
@@ -320,9 +320,9 @@ struct wl1271 {
 	bool watchdog_recovery;
 
 	/* Reg domain last configuration */
-	u32 reg_ch_conf_last[2]  __aligned(8);
+	DECLARE_BITMAP(reg_ch_conf_last, 64);
 	/* Reg domain pending configuration */
-	u32 reg_ch_conf_pending[2];
+	DECLARE_BITMAP(reg_ch_conf_pending, 64);
 
 	/* Pointer that holds DMA-friendly block for the mailbox */
 	void *mbox;
maintainer
#18 lore raw

Re: [PATCH v4 03/17] wlcore: Align reg_ch_conf_pending and tmp_ch_bitmap to unsigned long for better performance

Quoted 4 lines · Fenghua Yu

drivers/net/wireless/ti/wlcore/cmd.c void wlcore_set_pending_regdomain_ch(struct wl1271 *wl, u16 channel, · line 1710

Quoted 1 line
drivers/net/wireless/ti/wlcore/cmd.c +1 -0
+	u32 tmp_ch_bitmap[2] __aligned(sizeof(unsigned long));

Now __aligned() is unnecessary because __set_bit_le() handles tmp_ch_bitmap, right?

It's still needed because, as explained by Peter, bitmap functions should always operate on properly aligned variables.

Quoted 16 lines
Quoted 16 lines
drivers/net/wireless/ti/wlcore/cmd.c +1 -1
 	struct wiphy *wiphy = wl->hw->wiphy;
 	struct ieee80211_supported_band *band;
 	bool timeout = false;
@@ -1717,7 +1717,7 @@ int wlcore_cmd_regdomain_config_locked(struct wl1271 *wl)
 
 	wl1271_debug(DEBUG_CMD, "cmd reg domain config");
 
-	memset(tmp_ch_bitmap, 0, sizeof(tmp_ch_bitmap));
+	memcpy(tmp_ch_bitmap, wl->reg_ch_conf_pending, sizeof(tmp_ch_bitmap));
 
 	for (b = NL80211_BAND_2GHZ; b <= NL80211_BAND_5GHZ; b++) {
 		band = wiphy->bands[b];
@@ -1738,13 +1738,10 @@ int wlcore_cmd_regdomain_config_locked(struct wl1271 *wl)
 			if (ch_bit_idx < 0)
 				continue;
 
Quoted 4 lines

drivers/net/wireless/ti/wlcore/cmd.c int wlcore_cmd_regdomain_config_locked(struct wl1271 *wl) · line 1741

Quoted 2 lines
drivers/net/wireless/ti/wlcore/cmd.c +1 -1
-			set_bit(ch_bit_idx, (long *)tmp_ch_bitmap);
+			__set_bit_le(ch_bit_idx, (long *)tmp_ch_bitmap);

Is __test_and_set_bit_le() more meaningful to avoid duplicate bit setting ?

No, since the return value would be unused.

Note that this patch is missing the removal of cpu_to_le32, as noticed by Peter.

Thanks,

Paolo

Quoted 26 lines
Quoted 26 lines
drivers/net/wireless/ti/wlcore/cmd.c +0 -3
 		}
 	}
 
-	tmp_ch_bitmap[0] |= wl->reg_ch_conf_pending[0];
-	tmp_ch_bitmap[1] |= wl->reg_ch_conf_pending[1];
-
 	if (!memcmp(tmp_ch_bitmap, wl->reg_ch_conf_last, sizeof(tmp_ch_bitmap)))
 		goto out;
 
drivers/net/wireless/ti/wlcore/wlcore.h +2 -2
index dd14850b0603..870eea3e7a27 100644
@@ -320,9 +320,9 @@ struct wl1271 {
 	bool watchdog_recovery;
 
 	/* Reg domain last configuration */
-	u32 reg_ch_conf_last[2]  __aligned(8);
+	DECLARE_BITMAP(reg_ch_conf_last, 64);
 	/* Reg domain pending configuration */
-	u32 reg_ch_conf_pending[2];
+	DECLARE_BITMAP(reg_ch_conf_pending, 64);
 
 	/* Pointer that holds DMA-friendly block for the mailbox */
 	void *mbox;
author
#19 lore raw

PATCH v4 4/17 x86/split_lock: Align x86_capability to unsigned long to avoid split locked access

In reply to Fenghua Yu

set_cpu_cap() calls locked BTS and clear_cpu_cap() calls locked BTR to operate on bitmap defined in x86_capability.

Locked BTS/BTR accesses a single unsigned long location. In 64-bit mode, the location is at:
base address of x86_capability + (bit offset in x86_capability % 64) * 8

Since base address of x86_capability may not aligned to unsigned long, the single unsigned long location may cross two cache lines and accessing the location by locked BTS/BTR introductions will trigger #AC.

To fix the split lock issue, align x86_capability to unsigned long so that the location will be always within one cache line.

Changing x86_capability[]'s type to unsigned long may also fix the issue because x86_capability[] will be naturally aligned to unsigned long. But this needs additional code changes. So we choose the simpler solution by enforcing alignment using __aligned(unsigned long).

---
 arch/x86/include/asm/processor.h | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)
arch/x86/include/asm/processor.h +3 -1
index 33051436c864..eb8ae701ef65 100644
@@ -93,7 +93,9 @@ struct cpuinfo_x86 {
 	__u32			extended_cpuid_level;
 	/* Maximum supported CPUID level, -1=no CPUID: */
 	int			cpuid_level;
-	__u32			x86_capability[NCAPINTS + NBUGINTS];
+	/* Unsigned long alignment to avoid split lock in atomic bitmap ops */
+	__u32			x86_capability[NCAPINTS + NBUGINTS]
+				__aligned(sizeof(unsigned long));
 	char			x86_vendor_id[16];
 	char			x86_model_id[64];
 	/* in KB - valid for CPUS which support this call: */
Signature · 1 line
2.7.4
#20 lore raw

Re: [PATCH v4 04/17] x86/split_lock: Align x86_capability to unsigned long to avoid split locked access

arch/x86/include/asm/processor.h struct cpuinfo_x86 { · line 98

Quoted 12 lines · Fenghua Yu
arch/x86/include/asm/processor.h +3 -1
index 33051436c864..eb8ae701ef65 100644
@@ -93,7 +93,9 @@ struct cpuinfo_x86 {
 	__u32			extended_cpuid_level;
 	/* Maximum supported CPUID level, -1=no CPUID: */
 	int			cpuid_level;
-	__u32			x86_capability[NCAPINTS + NBUGINTS];
+	/* Unsigned long alignment to avoid split lock in atomic bitmap ops */
+	__u32			x86_capability[NCAPINTS + NBUGINTS]
+				__aligned(sizeof(unsigned long));

I think this also warrants a comment in the changelog about the alignment of 'struct cpuinfo_x86'.

author
#21 lore raw

Re: [PATCH v4 04/17] x86/split_lock: Align x86_capability to unsigned long to avoid split locked access

Quoted 16 lines · Dave Hansen

arch/x86/include/asm/processor.h struct cpuinfo_x86 { · line 98

Quoted 12 lines · Fenghua Yu
arch/x86/include/asm/processor.h +3 -1
index 33051436c864..eb8ae701ef65 100644
@@ -93,7 +93,9 @@ struct cpuinfo_x86 {
 	__u32			extended_cpuid_level;
 	/* Maximum supported CPUID level, -1=no CPUID: */
 	int			cpuid_level;
-	__u32			x86_capability[NCAPINTS + NBUGINTS];
+	/* Unsigned long alignment to avoid split lock in atomic bitmap ops */
+	__u32			x86_capability[NCAPINTS + NBUGINTS]
+				__aligned(sizeof(unsigned long));

I think this also warrants a comment in the changelog about the alignment of 'struct cpuinfo_x86'.

How about add "Depending on the starting address where GCC generates for data of struct cpuinfo_x86, x86_capability[] may or may not align to unsigned long...."?

Thanks.

-Fenghua

#22 lore raw

Re: [PATCH v4 04/17] x86/split_lock: Align x86_capability to unsigned long to avoid split locked access

Quoted 16 lines · Fenghua Yu
Quoted 15 lines · Dave Hansen

arch/x86/include/asm/processor.h struct cpuinfo_x86 { · line 98

Quoted 12 lines · Fenghua Yu
arch/x86/include/asm/processor.h +3 -1
index 33051436c864..eb8ae701ef65 100644
@@ -93,7 +93,9 @@ struct cpuinfo_x86 {
 	__u32			extended_cpuid_level;
 	/* Maximum supported CPUID level, -1=no CPUID: */
 	int			cpuid_level;
-	__u32			x86_capability[NCAPINTS + NBUGINTS];
+	/* Unsigned long alignment to avoid split lock in atomic bitmap ops */
+	__u32			x86_capability[NCAPINTS + NBUGINTS]
+				__aligned(sizeof(unsigned long));

I think this also warrants a comment in the changelog about the alignment of 'struct cpuinfo_x86'.

Quoted 3 lines

How about add "Depending on the starting address where GCC generates for data of struct cpuinfo_x86, x86_capability[] may or may not align to unsigned long...."?

If that's the case, then what good is this patch? Sounds like some of the story is missing.

You might want to dig through some of the past discussions about this. I know this exact topic has been broached with Thomas.

#23 lore raw

Re: [PATCH v4 04/17] x86/split_lock: Align x86_capability to unsigned long to avoid split locked access

In reply to Fenghua Yu
Quoted 18 lines · Fenghua Yu
Quoted 16 lines · Dave Hansen

arch/x86/include/asm/processor.h struct cpuinfo_x86 { · line 98

Quoted 12 lines · Fenghua Yu
arch/x86/include/asm/processor.h +3 -1
index 33051436c864..eb8ae701ef65 100644
@@ -93,7 +93,9 @@ struct cpuinfo_x86 {
 	__u32			extended_cpuid_level;
 	/* Maximum supported CPUID level, -1=no CPUID: */
 	int			cpuid_level;
-	__u32			x86_capability[NCAPINTS + NBUGINTS];
+	/* Unsigned long alignment to avoid split lock in atomic bitmap ops */
+	__u32			x86_capability[NCAPINTS + NBUGINTS]
+				__aligned(sizeof(unsigned long));

I think this also warrants a comment in the changelog about the alignment of 'struct cpuinfo_x86'.

Quoted 3 lines

How about add "Depending on the starting address where GCC generates for data of struct cpuinfo_x86, x86_capability[] may or may not align to unsigned long...."?

Composite types inherit the strictest alignment of their members. Without that alignment hints would be utterly useless.

#24 lore raw

Re: [PATCH v4 04/17] x86/split_lock: Align x86_capability to unsigned long to avoid split locked access

In reply to Dave Hansen
Quoted 16 lines · Dave Hansen

arch/x86/include/asm/processor.h struct cpuinfo_x86 { · line 98

Quoted 12 lines · Fenghua Yu
arch/x86/include/asm/processor.h +3 -1
index 33051436c864..eb8ae701ef65 100644
@@ -93,7 +93,9 @@ struct cpuinfo_x86 {
 	__u32			extended_cpuid_level;
 	/* Maximum supported CPUID level, -1=no CPUID: */
 	int			cpuid_level;
-	__u32			x86_capability[NCAPINTS + NBUGINTS];
+	/* Unsigned long alignment to avoid split lock in atomic bitmap ops */
+	__u32			x86_capability[NCAPINTS + NBUGINTS]
+				__aligned(sizeof(unsigned long));

I think this also warrants a comment in the changelog about the alignment of 'struct cpuinfo_x86'.

Nah.

author
#25 lore raw

PATCH v4 5/17 x86/cpufeatures: Enumerate IA32_CORE_CAPABILITIES MSR

In reply to Fenghua Yu

MSR register IA32_CORE_CAPABILITIES (0xCF) contains bits that enumerate some model specific features.

The MSR 0xCF itself is enumerated by CPUID.(EAX=0x7,ECX=0):EDX[30]. When this bit is 1, the MSR 0xCF exists.

Detailed information for the CPUID bit and the MSR can be found in the latest Intel Architecture Instruction Set Extensions and Future Features Programming Reference.

---
 arch/x86/include/asm/cpufeatures.h | 1 +
 1 file changed, 1 insertion(+)
arch/x86/include/asm/cpufeatures.h +1 -0
index 6d6122524711..350eeccd0ce9 100644
@@ -349,6 +349,7 @@
 #define X86_FEATURE_INTEL_STIBP		(18*32+27) /* "" Single Thread Indirect Branch Predictors */
 #define X86_FEATURE_FLUSH_L1D		(18*32+28) /* Flush L1D cache */
 #define X86_FEATURE_ARCH_CAPABILITIES	(18*32+29) /* IA32_ARCH_CAPABILITIES MSR (Intel) */
+#define X86_FEATURE_CORE_CAPABILITY	(18*32+30) /* IA32_CORE_CAPABILITY MSR */
 #define X86_FEATURE_SPEC_CTRL_SSBD	(18*32+31) /* "" Speculative Store Bypass Disable */
 
 /*
Signature · 1 line
2.7.4
#26 lore raw

Re: [PATCH v4 05/17] x86/cpufeatures: Enumerate IA32_CORE_CAPABILITIES MSR

arch/x86/include/asm/cpufeatures.h line 353

Quoted 10 lines · Fenghua Yu
arch/x86/include/asm/cpufeatures.h +1 -0
index 6d6122524711..350eeccd0ce9 100644
@@ -349,6 +349,7 @@
 #define X86_FEATURE_INTEL_STIBP		(18*32+27) /* "" Single Thread Indirect Branch Predictors */
 #define X86_FEATURE_FLUSH_L1D		(18*32+28) /* Flush L1D cache */
 #define X86_FEATURE_ARCH_CAPABILITIES	(18*32+29) /* IA32_ARCH_CAPABILITIES MSR (Intel) */
+#define X86_FEATURE_CORE_CAPABILITY	(18*32+30) /* IA32_CORE_CAPABILITY MSR */
 #define X86_FEATURE_SPEC_CTRL_SSBD	(18*32+31) /* "" Speculative Store Bypass Disable */

What does this feature end up looking like in /proc/cpuinfo?

#27 lore raw

RE: [PATCH v4 05/17] x86/cpufeatures: Enumerate IA32_CORE_CAPABILITIES MSR

Quoted 12 lines

From: Hansen, Dave

Review

Quoted 1 line · Fenghua Yu
(quoted diff) +0 -0

b/arch/x86/include/asm/cpufeatures.h

arch/x86/include/asm/cpufeatures.h line 349

Quoted 5 lines

index 6d6122524711..350eeccd0ce9 100644

arch/x86/include/asm/cpufeatures.h +0 -0
@@ -349,6 +349,7 @@
 #define X86_FEATURE_INTEL_STIBP		(18*32+27) /* "" Single

Thread Indirect Branch Predictors */

arch/x86/include/asm/cpufeatures.h line 350

Quoted 1 line
arch/x86/include/asm/cpufeatures.h +0 -0
 #define X86_FEATURE_FLUSH_L1D		(18*32+28) /* Flush L1D

cache */

Quoted 8 lines

arch/x86/include/asm/cpufeatures.h line 351

Quoted 1 line
arch/x86/include/asm/cpufeatures.h +0 -0
 #define X86_FEATURE_ARCH_CAPABILITIES	(18*32+29) /*

IA32_ARCH_CAPABILITIES MSR (Intel) */

arch/x86/include/asm/cpufeatures.h line 352

Quoted 1 line
arch/x86/include/asm/cpufeatures.h +1 -0
+#define X86_FEATURE_CORE_CAPABILITY	(18*32+30) /*

IA32_CORE_CAPABILITY MSR */

arch/x86/include/asm/cpufeatures.h line 353

Quoted 1 line
arch/x86/include/asm/cpufeatures.h +0 -0
 #define X86_FEATURE_SPEC_CTRL_SSBD	(18*32+31) /* "" Speculative

Store Bypass Disable */

What does this feature end up looking like in /proc/cpuinfo?

The flag string is "core_capability" in /proc/cpuinfo.

Thanks.

-Fenghua

#28 lore raw

Re: [PATCH v4 05/17] x86/cpufeatures: Enumerate IA32_CORE_CAPABILITIES MSR

Quoted 2 lines · Fenghua
Quoted 1 line

What does this feature end up looking like in /proc/cpuinfo?

The flag string is "core_capability" in /proc/cpuinfo.

Ahh, OK, I misread the lack of quotes in the comment:

Quoted 1 line

+#define X86_FEATURE_CORE_CAPABILITY (18*32+30) /* IA32_CORE_CAPABILITY MSR */

author
#29 lore raw

PATCH v4 6/17 x86/msr-index: Define IA32_CORE_CAPABILITY MSR and #AC exception for split lock bit

In reply to Fenghua Yu

A new IA32_CORE_CAPABILITY MSR (0xCF) is defined. Each bit in the MSR enumerates a model specific feature. Currently bit 5 enumerates #AC exception for split locked accesses. When bit 5 is 1, split locked accesses will generate #AC exception. When bit 5 is 0, split locked accesses will not generate #AC exception.

Please check the latest Intel Architecture Instruction Set Extensions and Future Features Programming Reference for more detailed information on the MSR and the split lock bit.

---
 arch/x86/include/asm/msr-index.h | 3 +++
 1 file changed, 3 insertions(+)
arch/x86/include/asm/msr-index.h +3 -0
index 8e40c2446fd1..549e73dcca15 100644
@@ -59,6 +59,9 @@
 #define MSR_PLATFORM_INFO_CPUID_FAULT_BIT	31
 #define MSR_PLATFORM_INFO_CPUID_FAULT		BIT_ULL(MSR_PLATFORM_INFO_CPUID_FAULT_BIT)
 
+#define MSR_IA32_CORE_CAPABILITY	0x000000cf
+#define CORE_CAP_SPLIT_LOCK_DETECT	BIT(5)     /* Detect split lock */
+
 #define MSR_PKG_CST_CONFIG_CONTROL	0x000000e2
 #define NHM_C3_AUTO_DEMOTE		(1UL << 25)
 #define NHM_C1_AUTO_DEMOTE		(1UL << 26)
Signature · 1 line
2.7.4
author
#30 lore raw

PATCH v4 7/17 x86/split_lock: Enumerate #AC for split lock by MSR IA32_CORE_CAPABILITY

In reply to Fenghua Yu

Bits in MSR IA32_CORE_CAPABILITY enumerate features that are not enumerated through CPUID. Currently bit 5 is defined to enumerate feature of #AC for split lock accesses. All other bits are reserved now.

When the bit 5 is 1, the feature is supported and feature bit X86_FEATURE_SPLIT_LOCK_DETECT is set. Otherwise, the feature is not available.

The MSR IA32_CORE_CAPABILITY itself is enumerated by CPUID.(EAX=0x7,ECX=0):EDX[30].

---
 arch/x86/include/asm/cpu.h         |  5 +++++
 arch/x86/include/asm/cpufeatures.h |  1 +
 arch/x86/kernel/cpu/common.c       |  1 +
 arch/x86/kernel/cpu/cpuid-deps.c   |  1 +
 arch/x86/kernel/cpu/intel.c        | 21 +++++++++++++++++++++
 5 files changed, 29 insertions(+)
arch/x86/include/asm/cpu.h +5 -0
index adc6cc86b062..e241abce1a2a 100644
@@ -40,4 +40,9 @@ int mwait_usable(const struct cpuinfo_x86 *);
 unsigned int x86_family(unsigned int sig);
 unsigned int x86_model(unsigned int sig);
 unsigned int x86_stepping(unsigned int sig);
+#ifdef CONFIG_CPU_SUP_INTEL
+void init_core_capability(struct cpuinfo_x86 *c);
+#else
+static inline void init_core_capability(struct cpuinfo_x86 *c) {}
+#endif
 #endif /* _ASM_X86_CPU_H */
arch/x86/include/asm/cpufeatures.h +1 -0
index 350eeccd0ce9..54c73e74213d 100644
@@ -221,6 +221,7 @@
 #define X86_FEATURE_ZEN			( 7*32+28) /* "" CPU is AMD family 0x17 (Zen) */
 #define X86_FEATURE_L1TF_PTEINV		( 7*32+29) /* "" L1TF workaround PTE inversion */
 #define X86_FEATURE_IBRS_ENHANCED	( 7*32+30) /* Enhanced IBRS */
+#define X86_FEATURE_SPLIT_LOCK_DETECT	( 7*32+31) /* #AC for split lock */
 
 /* Virtualization flags: Linux defined, word 8 */
 #define X86_FEATURE_TPR_SHADOW		( 8*32+ 0) /* Intel TPR Shadow */
arch/x86/kernel/cpu/common.c +1 -0
index 51ab37ba5f64..79e7cc0c4c85 100644
@@ -897,6 +897,7 @@ void get_cpu_cap(struct cpuinfo_x86 *c)
 
 	init_scattered_cpuid_features(c);
 	init_speculation_control(c);
+	init_core_capability(c);
 
 	/*
 	 * Clear/Set all flags overridden by options, after probe.
arch/x86/kernel/cpu/cpuid-deps.c +1 -0
index 2c0bd38a44ab..5ba11ce99f92 100644
@@ -59,6 +59,7 @@ static const struct cpuid_dep cpuid_deps[] = {
 	{ X86_FEATURE_AVX512_4VNNIW,	X86_FEATURE_AVX512F   },
 	{ X86_FEATURE_AVX512_4FMAPS,	X86_FEATURE_AVX512F   },
 	{ X86_FEATURE_AVX512_VPOPCNTDQ, X86_FEATURE_AVX512F   },
+	{ X86_FEATURE_SPLIT_LOCK_DETECT, X86_FEATURE_CORE_CAPABILITY},
 	{}
 };
 
arch/x86/kernel/cpu/intel.c +21 -0
index fc3c07fe7df5..0c44c49f6005 100644
@@ -1029,3 +1029,24 @@ static const struct cpu_dev intel_cpu_dev = {
 
 cpu_dev_register(intel_cpu_dev);
 
+/**
+ * init_core_capability - enumerate features supported in IA32_CORE_CAPABILITY
+ * @c: pointer to cpuinfo_x86
+ *
+ * Return: void
+ */
+void init_core_capability(struct cpuinfo_x86 *c)
+{
+	/*
+	 * If MSR_IA32_CORE_CAPABILITY exists, enumerate features that are
+	 * reported in the MSR.
+	 */
+	if (c == &boot_cpu_data && cpu_has(c, X86_FEATURE_CORE_CAPABILITY)) {
+		u64 val;
+
+		rdmsrl(MSR_IA32_CORE_CAPABILITY, val);
+
+		if (val & CORE_CAP_SPLIT_LOCK_DETECT)
+			setup_force_cpu_cap(X86_FEATURE_SPLIT_LOCK_DETECT);
+	}
+}
Signature · 1 line
2.7.4
#31 lore raw

Re: [PATCH v4 07/17] x86/split_lock: Enumerate #AC for split lock by MSR IA32_CORE_CAPABILITY

arch/x86/kernel/cpu/cpuid-deps.c static const struct cpuid_dep cpuid_deps[] = { · line 64

Quoted 12 lines
arch/x86/kernel/cpu/common.c +0 -0
 	 * Clear/Set all flags overridden by options, after probe.
arch/x86/kernel/cpu/cpuid-deps.c +1 -0
index 2c0bd38a44ab..5ba11ce99f92 100644
@@ -59,6 +59,7 @@ static const struct cpuid_dep cpuid_deps[] = {
 	{ X86_FEATURE_AVX512_4VNNIW,	X86_FEATURE_AVX512F   },
 	{ X86_FEATURE_AVX512_4FMAPS,	X86_FEATURE_AVX512F   },
 	{ X86_FEATURE_AVX512_VPOPCNTDQ, X86_FEATURE_AVX512F   },
+	{ X86_FEATURE_SPLIT_LOCK_DETECT, X86_FEATURE_CORE_CAPABILITY},
 	{}
 };

Please reindent the rest of the structure whenever you break the record for name length.

Quoted 11 lines
arch/x86/kernel/cpu/intel.c +3 -0
index fc3c07fe7df5..0c44c49f6005 100644
@@ -1029,3 +1029,24 @@ static const struct cpu_dev intel_cpu_dev = {
 
 cpu_dev_register(intel_cpu_dev);
 
+/**
+ * init_core_capability - enumerate features supported in IA32_CORE_CAPABILITY
+ * @c: pointer to cpuinfo_x86

arch/x86/kernel/cpu/intel.c static const struct cpu_dev intel_cpu_dev = { · line 1044

Quoted 10 lines
arch/x86/kernel/cpu/intel.c +10 -0
+ *
+ * Return: void
+ */
+void init_core_capability(struct cpuinfo_x86 *c)
+{
+	/*
+	 * If MSR_IA32_CORE_CAPABILITY exists, enumerate features that are
+	 * reported in the MSR.
+	 */
+	if (c == &boot_cpu_data && cpu_has(c, X86_FEATURE_CORE_CAPABILITY)) {

First of all, please endeavor to leave the main flow of the function at the first indent.

*If* you were to do this, it would look like:

	if (c != &boot_cpu_data)
		return;
	if (!cpu_has(c, X86_FEATURE_CORE_CAPABILITY))
		return;

	rdmsrl(...);

But, it goes unmentioned why you do the boot-cpu-only restriction here.
It doesn't match the behavior of the two functions called before
init_core_capability():

        init_scattered_cpuid_features(c);
        init_speculation_control(c);

So why is this new function special?

Quoted 9 lines
arch/x86/kernel/cpu/intel.c +8 -0
+		u64 val;
+
+		rdmsrl(MSR_IA32_CORE_CAPABILITY, val);
+
+		if (val & CORE_CAP_SPLIT_LOCK_DETECT)
+			setup_force_cpu_cap(X86_FEATURE_SPLIT_LOCK_DETECT);
+	}
+}
author
#32 lore raw

Re: [PATCH v4 07/17] x86/split_lock: Enumerate #AC for split lock by MSR IA32_CORE_CAPABILITY

Quoted 7 lines · Dave Hansen
Quoted 7 lines
arch/x86/kernel/cpu/common.c +0 -0
 	 * Clear/Set all flags overridden by options, after probe.
arch/x86/kernel/cpu/cpuid-deps.c +0 -0
index 2c0bd38a44ab..5ba11ce99f92 100644
@@ -59,6 +59,7 @@ static const struct cpuid_dep cpuid_deps[] = {
 	{ X86_FEATURE_AVX512_4VNNIW,	X86_FEATURE_AVX512F   },
Quoted 8 lines

arch/x86/kernel/cpu/cpuid-deps.c static const struct cpuid_dep cpuid_deps[] = { · line 64

Quoted 5 lines
arch/x86/kernel/cpu/cpuid-deps.c +1 -0
 	{ X86_FEATURE_AVX512_4FMAPS,	X86_FEATURE_AVX512F   },
 	{ X86_FEATURE_AVX512_VPOPCNTDQ, X86_FEATURE_AVX512F   },
+	{ X86_FEATURE_SPLIT_LOCK_DETECT, X86_FEATURE_CORE_CAPABILITY},
 	{}
 };

Please reindent the rest of the structure whenever you break the record for name length.

Ok. Will do it.

Quoted 29 lines
Quoted 12 lines
arch/x86/kernel/cpu/intel.c +3 -0
index fc3c07fe7df5..0c44c49f6005 100644
@@ -1029,3 +1029,24 @@ static const struct cpu_dev intel_cpu_dev = {
 
 cpu_dev_register(intel_cpu_dev);
 
+/**
+ * init_core_capability - enumerate features supported in IA32_CORE_CAPABILITY
+ * @c: pointer to cpuinfo_x86

arch/x86/kernel/cpu/intel.c static const struct cpu_dev intel_cpu_dev = { · line 1044

Quoted 10 lines
arch/x86/kernel/cpu/intel.c +10 -0
+ *
+ * Return: void
+ */
+void init_core_capability(struct cpuinfo_x86 *c)
+{
+	/*
+	 * If MSR_IA32_CORE_CAPABILITY exists, enumerate features that are
+	 * reported in the MSR.
+	 */
+	if (c == &boot_cpu_data && cpu_has(c, X86_FEATURE_CORE_CAPABILITY)) {

First of all, please endeavor to leave the main flow of the function at the first indent.

*If* you were to do this, it would look like:

Quoted 6 lines
	if (c != &boot_cpu_data)
		return;
	if (!cpu_has(c, X86_FEATURE_CORE_CAPABILITY))
		return;

	rdmsrl(...);

Ok. Will do it.

Quoted 9 lines

But, it goes unmentioned why you do the boot-cpu-only restriction here.
It doesn't match the behavior of the two functions called before
init_core_capability():

        init_scattered_cpuid_features(c);
        init_speculation_control(c);

So why is this new function special?

The function sets the split_lock_detect feature which needs to be applied before BSP calls apply_enforced_caps() in get_cpu_cap(). And I want to enable split lock detection in earlier phase to detect split lock earlier.

Thanks.

-Fenghua

#33 lore raw

Re: [PATCH v4 07/17] x86/split_lock: Enumerate #AC for split lock by MSR IA32_CORE_CAPABILITY

Quoted 10 lines · Fenghua Yu
Quoted 8 lines

But, it goes unmentioned why you do the boot-cpu-only restriction here.
It doesn't match the behavior of the two functions called before
init_core_capability():

        init_scattered_cpuid_features(c);
        init_speculation_control(c);

So why is this new function special?

The function sets the split_lock_detect feature which needs to be applied before BSP calls apply_enforced_caps() in get_cpu_cap().

Why does it need to be applied that way?

Quoted 2 lines

And I want to enable split lock detection in earlier phase to detect split lock earlier.

Why do you want to do this? Why is this *required* as part of this implementation?

... and how is this different than the other features that are processed for cpus other than the boot CPU?

author
#34 lore raw

PATCH v4 8/17 x86/clearcpuid: Support multiple clearcpuid options

In reply to Fenghua Yu

Currently only one kernel option "clearcpuid=" can be picked up by kernel during boot time.

In some cases, user may want to clear a few cpu caps. This may be useful to replace endless (new) kernel options like nosmep, nosmap, etc.

We add support of multiple clearcpuid options to allow user to clear multiple cpu caps.

---
 arch/x86/include/asm/cmdline.h |  3 +++
 arch/x86/kernel/fpu/init.c     | 30 ++++++++++++++++++++----------
 arch/x86/lib/cmdline.c         | 17 +++++++++++++++--
 3 files changed, 38 insertions(+), 12 deletions(-)
arch/x86/include/asm/cmdline.h +3 -0
index 6faaf27e8899..059e29558bb3 100644
@@ -5,5 +5,8 @@
 int cmdline_find_option_bool(const char *cmdline_ptr, const char *option);
 int cmdline_find_option(const char *cmdline_ptr, const char *option,
 			char *buffer, int bufsize);
+int cmdline_find_option_in_range(const char *cmdline_ptr, int cmdline_size,
+				 const char *option, char *buffer, int bufsize,
+				 char **arg_pos);
 
 #endif /* _ASM_X86_CMDLINE_H */
arch/x86/kernel/fpu/init.c +20 -10
index 6abd83572b01..88bbba7ee96a 100644
@@ -243,16 +243,31 @@ static void __init fpu__init_system_ctx_switch(void)
 	WARN_ON_FPU(current->thread.fpu.initialized);
 }
 
+static void __init clear_cpuid(void)
+{
+	char arg[32], *argptr, *option_pos, clearcpuid_option[] = "clearcpuid";
+	int bit, cmdline_size;
+
+	/* Find each option in boot_command_line and clear specified cpu cap. */
+	cmdline_size = COMMAND_LINE_SIZE;
+	while (cmdline_find_option_in_range(boot_command_line, cmdline_size,
+					    clearcpuid_option, arg,
+					    sizeof(arg), &option_pos) >= 0) {
+		/* Chang command line range for next search. */
+		cmdline_size = option_pos - boot_command_line + 1;
+		argptr = arg;
+		if (get_option(&argptr, &bit) &&
+		    bit >= 0 && bit < NCAPINTS * 32)
+			setup_clear_cpu_cap(bit);
+	}
+}
+
 /*
  * We parse fpu parameters early because fpu__init_system() is executed
  * before parse_early_param().
  */
 static void __init fpu__init_parse_early_param(void)
 {
-	char arg[32];
-	char *argptr = arg;
-	int bit;
-
 	if (cmdline_find_option_bool(boot_command_line, "no387"))
 		setup_clear_cpu_cap(X86_FEATURE_FPU);
 
@@ -271,12 +286,7 @@ static void __init fpu__init_parse_early_param(void)
 	if (cmdline_find_option_bool(boot_command_line, "noxsaves"))
 		setup_clear_cpu_cap(X86_FEATURE_XSAVES);
 
-	if (cmdline_find_option(boot_command_line, "clearcpuid", arg,
-				sizeof(arg)) &&
-	    get_option(&argptr, &bit) &&
-	    bit >= 0 &&
-	    bit < NCAPINTS * 32)
-		setup_clear_cpu_cap(bit);
+	clear_cpuid();
 }
 
 /*
arch/x86/lib/cmdline.c +15 -2
index 3261abb21ef4..9cf1a0773877 100644
@@ -114,13 +114,15 @@ __cmdline_find_option_bool(const char *cmdline, int max_cmdline_size,
  * @option: option string to look for
  * @buffer: memory buffer to return the option argument
  * @bufsize: size of the supplied memory buffer
+ * @option_pos: pointer to the option if found
  *
  * Returns the length of the argument (regardless of if it was
  * truncated to fit in the buffer), or -1 on not found.
  */
 static int
 __cmdline_find_option(const char *cmdline, int max_cmdline_size,
-		      const char *option, char *buffer, int bufsize)
+		      const char *option, char *buffer, int bufsize,
+		      char **arg_pos)
 {
 	char c;
 	int pos = 0, len = -1;
@@ -164,6 +166,9 @@ __cmdline_find_option(const char *cmdline, int max_cmdline_size,
 				len = 0;
 				bufptr = buffer;
 				state = st_bufcpy;
+				if (arg_pos)
+					*arg_pos = (char *)cmdline -
+						   strlen(option) - 1;
 				break;
 			} else if (c == *opptr++) {
 				/*
@@ -211,5 +216,13 @@ int cmdline_find_option(const char *cmdline, const char *option, char *buffer,
 			int bufsize)
 {
 	return __cmdline_find_option(cmdline, COMMAND_LINE_SIZE, option,
-				     buffer, bufsize);
+				     buffer, bufsize, NULL);
+}
+
+int cmdline_find_option_in_range(const char *cmdline, int cmdline_size,
+				 char *option, char *buffer, int bufsize,
+				 char **arg_pos)
+{
+	return __cmdline_find_option(cmdline, cmdline_size, option, buffer,
+				     bufsize, arg_pos);
 }
Signature · 1 line
2.7.4
author
#35 lore raw

PATCH v4 9/17 x86/clearcpuid: Support feature flag string in kernel option clearcpuid

In reply to Fenghua Yu

The kernel option clearcpuid currently only takes feature bit which can be changed from kernel to kernel.

Extend clearcpuid to use cap flag string, which is defined in x86_cap_flags[] and won't be changed from kernel to kernel. And user can easily get the cap flag string from /proc/cpuinfo.

---
 arch/x86/include/asm/cpufeature.h |  1 +
 arch/x86/kernel/cpu/cpuid-deps.c  | 26 ++++++++++++++++++++++++++
 arch/x86/kernel/fpu/init.c        |  3 ++-
 3 files changed, 29 insertions(+), 1 deletion(-)
arch/x86/include/asm/cpufeature.h +1 -0
index ce95b8cbd229..6792088525e3 100644
@@ -132,6 +132,7 @@ extern const char * const x86_bug_flags[NBUGINTS*32];
 
 extern void setup_clear_cpu_cap(unsigned int bit);
 extern void clear_cpu_cap(struct cpuinfo_x86 *c, unsigned int bit);
+bool find_cpu_cap(char *cap_flag, unsigned int *pfeature);
 
 #define setup_force_cpu_cap(bit) do { \
 	set_cpu_cap(&boot_cpu_data, bit);	\
arch/x86/kernel/cpu/cpuid-deps.c +26 -0
index 5ba11ce99f92..b84b133d0ebd 100644
@@ -120,3 +120,29 @@ void setup_clear_cpu_cap(unsigned int feature)
 {
 	do_clear_cpu_cap(NULL, feature);
 }
+
+/**
+ * find_cpu_cap - Given a cap flag string, find its corresponding feature bit.
+ * @cap_flag:	cap flag string as defined in x86_cap_flags[]
+ * @pfeature:	feature bit
+ *
+ * Return: true if the feature is found. false if not found
+ */
+bool find_cpu_cap(char *cap_flag, unsigned int *pfeature)
+{
+#ifdef CONFIG_X86_FEATURE_NAMES
+	unsigned int feature;
+
+	for (feature = 0; feature < NCAPINTS * 32; feature++) {
+		if (!x86_cap_flags[feature])
+			continue;
+
+		if (strcmp(cap_flag, x86_cap_flags[feature]) == 0) {
+			*pfeature = feature;
+
+			return true;
+		}
+	}
+#endif
+	return false;
+}
arch/x86/kernel/fpu/init.c +2 -1
index 88bbba7ee96a..99b895eea166 100644
@@ -256,7 +256,8 @@ static void __init clear_cpuid(void)
 		/* Chang command line range for next search. */
 		cmdline_size = option_pos - boot_command_line + 1;
 		argptr = arg;
-		if (get_option(&argptr, &bit) &&
+		/* cpu cap can be specified by either feature bit or string */
+		if ((get_option(&argptr, &bit) || find_cpu_cap(arg, &bit)) &&
 		    bit >= 0 && bit < NCAPINTS * 32)
 			setup_clear_cpu_cap(bit);
 	}
Signature · 1 line
2.7.4
author
#36 lore raw

PATCH v4 10/17 x86/clearcpuid: Apply cleared feature bits that are forced set before

In reply to Fenghua Yu

Some CPU feature bits are forced set and stored in cpuinfo_x86 before handling clearcpuid options. To clear those bits from cpuinfo_x86, apply_forced_cap() is called after handling the options.

Please note, apply_forced_cap() is called twice on boot CPU. But this code is simple and there is no functionality issue.

---
 arch/x86/include/asm/cpu.h   | 2 ++
 arch/x86/kernel/cpu/common.c | 5 +++--
 arch/x86/kernel/fpu/init.c   | 2 ++
 3 files changed, 7 insertions(+), 2 deletions(-)
arch/x86/include/asm/cpu.h +2 -0
index e241abce1a2a..487f32ecea0b 100644
@@ -26,6 +26,8 @@ struct x86_cpu {
 	struct cpu cpu;
 };
 
+void apply_forced_caps(struct cpuinfo_x86 *c);
+
 #ifdef CONFIG_HOTPLUG_CPU
 extern int arch_register_cpu(int num);
 extern void arch_unregister_cpu(int);
arch/x86/kernel/cpu/common.c +3 -2
index 79e7cc0c4c85..26723ea322fa 100644
@@ -758,13 +758,14 @@ void cpu_detect(struct cpuinfo_x86 *c)
 	}
 }
 
-static void apply_forced_caps(struct cpuinfo_x86 *c)
+void apply_forced_caps(struct cpuinfo_x86 *c)
 {
 	int i;
 
 	for (i = 0; i < NCAPINTS + NBUGINTS; i++) {
-		c->x86_capability[i] &= ~cpu_caps_cleared[i];
+		/* Bits may be cleared after they are set. */
 		c->x86_capability[i] |= cpu_caps_set[i];
+		c->x86_capability[i] &= ~cpu_caps_cleared[i];
 	}
 }
 
arch/x86/kernel/fpu/init.c +2 -0
index 99b895eea166..9c2801b605e3 100644
@@ -5,6 +5,7 @@
 #include <asm/tlbflush.h>
 #include <asm/setup.h>
 #include <asm/cmdline.h>
+#include <asm/cpu.h>
 
 #include <linux/sched.h>
 #include <linux/sched/task.h>
@@ -261,6 +262,7 @@ static void __init clear_cpuid(void)
 		    bit >= 0 && bit < NCAPINTS * 32)
 			setup_clear_cpu_cap(bit);
 	}
+	apply_forced_caps(&boot_cpu_data);
 }
 
 /*
Signature · 1 line
2.7.4
author
#37 lore raw

PATCH v4 11/17 x86/clearcpuid: Clear CPUID bit in CPUID faulting

In reply to Fenghua Yu

From: Peter Zijlstra <[email protected]>

After kernel clears a CPUID bit through clearcpuid or other kernel options, CPUID instruction executed from user space should see the same value for the bit. The CPUID faulting handler returns the cleared bit to user.

---
 arch/x86/include/asm/cpufeature.h |  4 +++
 arch/x86/kernel/cpu/common.c      |  2 ++
 arch/x86/kernel/cpu/cpuid-deps.c  | 52 ++++++++++++++++++++++++++++++++++++
 arch/x86/kernel/cpu/intel.c       | 55 ++++++++++++++++++++++++++++++++++++---
 arch/x86/kernel/cpu/scattered.c   | 17 ++++++++++++
 arch/x86/kernel/process.c         |  3 +++
 arch/x86/kernel/traps.c           | 11 ++++++++
 7 files changed, 141 insertions(+), 3 deletions(-)
arch/x86/include/asm/cpufeature.h +4 -0
index 6792088525e3..c4ac787d9b85 100644
@@ -227,5 +227,9 @@ static __always_inline __pure bool _static_cpu_has(u16 bit)
 #define CPU_FEATURE_TYPEVAL		boot_cpu_data.x86_vendor, boot_cpu_data.x86, \
 					boot_cpu_data.x86_model
 
+extern int cpuid_fault;
+u32 scattered_cpuid_mask(u32 leaf, u32 count, enum cpuid_regs_idx reg);
+u32 cpuid_cap_mask(u32 leaf, u32 count, enum cpuid_regs_idx reg);
+
 #endif /* defined(__KERNEL__) && !defined(__ASSEMBLY__) */
 #endif /* _ASM_X86_CPUFEATURE_H */
arch/x86/kernel/cpu/common.c +2 -0
index 26723ea322fa..aa2658136fe0 100644
@@ -1502,6 +1502,8 @@ void print_cpu_info(struct cpuinfo_x86 *c)
 		pr_cont(")\n");
 }
 
+int cpuid_fault;
+
 /*
  * clearcpuid= was already parsed in fpu__init_parse_early_param.
  * But we need to keep a dummy __setup around otherwise it would
arch/x86/kernel/cpu/cpuid-deps.c +52 -0

61 lines not on this page · raw message

arch/x86/kernel/cpu/intel.c +52 -3

71 lines not on this page · raw message

arch/x86/kernel/cpu/scattered.c +17 -0
index 94aa1c72ca98..353756c27096 100644
@@ -62,3 +62,20 @@ void init_scattered_cpuid_features(struct cpuinfo_x86 *c)
 			set_cpu_cap(c, cb->feature);
 	}
 }
+
+u32 scattered_cpuid_mask(u32 leaf, u32 count, enum cpuid_regs_idx reg)
+{
+	const struct cpuid_bit *cb;
+	u32 mask = ~0U;
+
+	for (cb = cpuid_bits; cb->feature; cb++) {
+		if (cb->level == leaf && cb->sub_leaf == count &&
+		    cb->reg == reg) {
+			if (test_bit(cb->feature,
+				     (unsigned long *)cpu_caps_cleared))
+				mask &= ~BIT(cb->bit);
+		}
+	}
+
+	return mask;
+}
arch/x86/kernel/process.c +3 -0
index 90ae0ca51083..1bba1a3c0b01 100644
@@ -196,6 +196,9 @@ static void set_cpuid_faulting(bool on)
 {
 	u64 msrval;
 
+	if (cpuid_fault)
+		return;
+
 	msrval = this_cpu_read(msr_misc_features_shadow);
 	msrval &= ~MSR_MISC_FEATURES_ENABLES_CPUID_FAULT;
 	msrval |= (on << MSR_MISC_FEATURES_ENABLES_CPUID_FAULT_BIT);
arch/x86/kernel/traps.c +11 -0
index 9b7c4ca8f0a7..c43c2608332e 100644
@@ -517,6 +517,12 @@ dotraplinkage void do_bounds(struct pt_regs *regs, long error_code)
 	do_trap(X86_TRAP_BR, SIGSEGV, "bounds", regs, error_code, 0, NULL);
 }
 
+#ifdef CONFIG_CPU_SUP_INTEL
+extern bool fixup_cpuid_exception(struct pt_regs *regs);
+#else
+static inline bool fixup_cpuid_exception(struct pt_regs *regs) { return false; }
+#endif
+
 dotraplinkage void
 do_general_protection(struct pt_regs *regs, long error_code)
 {
@@ -531,6 +537,11 @@ do_general_protection(struct pt_regs *regs, long error_code)
 			return;
 	}
 
+	if (static_cpu_has(X86_FEATURE_CPUID_FAULT)) {
+		if (user_mode(regs) && fixup_cpuid_exception(regs))
+			return;
+	}
+
 	if (v8086_mode(regs)) {
 		local_irq_enable();
 		handle_vm86_fault((struct kernel_vm86_regs *) regs, error_code);
Signature · 1 line
2.7.4
#38 lore raw

Re: [PATCH v4 11/17] x86/clearcpuid: Clear CPUID bit in CPUID faulting

I love your patch! Perhaps something to improve:

Hi Peter,

I love your patch! Perhaps something to improve:

[auto build test WARNING on kvm/linux-next]
[also build test WARNING on v5.0 next-20190304]
[if your patch is applied to the wrong git tree, please drop us a note to help improve the system]

url:    https://github.com/0day-ci/linux/commits/Fenghua-Yu/x86-split_lock-Enable-AC-exception-for-split-locked-accesses/20190305-044643
base:   https://git.kernel.org/pub/scm/virt/kvm/kvm.git linux-next
config: x86_64-randconfig-x017-201909 (attached as .config)
compiler: gcc-8 (Debian 8.2.0-21) 8.2.0
reproduce:
        # save the attached .config to linux build tree
        make ARCH=x86_64 

All warnings (new ones prefixed by >>):

In file included from arch/x86/kernel/cpu/intel.c:22:

>> arch/x86/include/asm/insn-eval.h:18:39: warning: 'struct insn' declared inside parameter list will not be visible outside of this definition or declaration
    void __user *insn_get_addr_ref(struct insn *insn, struct pt_regs *regs);
                                          ^~~~
   arch/x86/include/asm/insn-eval.h:19:34: warning: 'struct insn' declared inside parameter list will not be visible outside of this definition or declaration
    int insn_get_modrm_rm_off(struct insn *insn, struct pt_regs *regs);
                                     ^~~~

vim +18 arch/x86/include/asm/insn-eval.h

4efea85f Ricardo Neri    2017-10-27  17  
32542ee2 Ricardo Neri    2017-10-27 @18  void __user *insn_get_addr_ref(struct insn *insn, struct pt_regs *regs);
e5e45f11 Ricardo Neri    2017-10-27  19  int insn_get_modrm_rm_off(struct insn *insn, struct pt_regs *regs);
bd5a410a Ricardo Neri    2017-10-27  20  unsigned long insn_get_seg_base(struct pt_regs *regs, int seg_reg_idx);

e2a5dca7 Borislav Petkov 2017-11-23 21 int insn_get_code_seg_params(struct pt_regs *regs);
32542ee2 Ricardo Neri 2017-10-27 22

:::::: The code at line 18 was first introduced by commit
:::::: 32542ee295bec38e5e1608f8c9d6d28e5a7e6112 x86/mpx, x86/insn: Relocate insn util functions to a new insn-eval file

:::::: TO: Ricardo Neri <[email protected]> :::::: CC: Thomas Gleixner <[email protected]>

---
0-DAY kernel test infrastructure Open Source Technology Center
https://lists.01.org/pipermail/kbuild-all Intel Corporation

#39 lore raw

Re: [PATCH v4 11/17] x86/clearcpuid: Clear CPUID bit in CPUID faulting

In reply to Fenghua Yu
I love your patch! Perhaps something to improve:

Hi Peter,

I love your patch! Perhaps something to improve:

[auto build test WARNING on kvm/linux-next]
[also build test WARNING on v5.0 next-20190304]
[if your patch is applied to the wrong git tree, please drop us a note to help improve the system]

url:    https://github.com/0day-ci/linux/commits/Fenghua-Yu/x86-split_lock-Enable-AC-exception-for-split-locked-accesses/20190305-044643
base:   https://git.kernel.org/pub/scm/virt/kvm/kvm.git linux-next
config: i386-randconfig-a1-201909 (attached as .config)
compiler: gcc-4.9 (Debian 4.9.4-2) 4.9.4
reproduce:
        # save the attached .config to linux build tree
        make ARCH=i386 

All warnings (new ones prefixed by >>):

In file included from arch/x86/kernel/cpu/intel.c:22:0:

>> arch/x86/include/asm/insn-eval.h:18:58: warning: 'struct insn' declared inside parameter list
    void __user *insn_get_addr_ref(struct insn *insn, struct pt_regs *regs);
                                                             ^
>> arch/x86/include/asm/insn-eval.h:18:58: warning: its scope is only this definition or declaration, which is probably not what you want
   arch/x86/include/asm/insn-eval.h:19:53: warning: 'struct insn' declared inside parameter list
    int insn_get_modrm_rm_off(struct insn *insn, struct pt_regs *regs);
                                                        ^

Cyclomatic Complexity 5 include/linux/compiler.h:__read_once_size
Cyclomatic Complexity 1 include/linux/kasan-checks.h:kasan_check_read Cyclomatic Complexity 2 arch/x86/include/asm/bitops.h:set_bit
Cyclomatic Complexity 1 arch/x86/include/asm/bitops.h:constant_test_bit Cyclomatic Complexity 1 arch/x86/include/asm/bitops.h:variable_test_bit Cyclomatic Complexity 1 arch/x86/include/asm/bitops.h:fls
Cyclomatic Complexity 2 include/linux/bitops.h:get_count_order
Cyclomatic Complexity 1 include/linux/string.h:strnlen
Cyclomatic Complexity 4 include/linux/string.h:strlen
Cyclomatic Complexity 4 include/linux/string.h:memcpy
Cyclomatic Complexity 2 include/linux/string.h:strcpy
Cyclomatic Complexity 1 arch/x86/include/asm/atomic.h:arch_atomic_read Cyclomatic Complexity 1 include/asm-generic/atomic-instrumented.h:atomic_read Cyclomatic Complexity 1 arch/x86/include/asm/current.h:get_current Cyclomatic Complexity 1 arch/x86/include/asm/irqflags.h:native_save_fl Cyclomatic Complexity 1 arch/x86/include/asm/irqflags.h:native_restore_fl Cyclomatic Complexity 1 arch/x86/include/asm/irqflags.h:native_irq_disable Cyclomatic Complexity 1 arch/x86/include/asm/irqflags.h:arch_local_save_flags Cyclomatic Complexity 1 arch/x86/include/asm/irqflags.h:arch_local_irq_restore Cyclomatic Complexity 1 arch/x86/include/asm/irqflags.h:arch_local_irq_disable Cyclomatic Complexity 1 arch/x86/include/asm/irqflags.h:arch_local_irq_save Cyclomatic Complexity 1 include/linux/jump_label.h:static_key_count Cyclomatic Complexity 2 include/linux/jump_label.h:static_key_false Cyclomatic Complexity 1 arch/x86/include/asm/ptrace.h:user_64bit_mode Cyclomatic Complexity 1 arch/x86/include/asm/msr.h:__rdmsr
Cyclomatic Complexity 1 arch/x86/include/asm/msr.h:__wrmsr
Cyclomatic Complexity 2 arch/x86/include/asm/msr.h:native_read_msr Cyclomatic Complexity 2 arch/x86/include/asm/msr.h:native_read_msr_safe Cyclomatic Complexity 2 arch/x86/include/asm/msr.h:native_write_msr Cyclomatic Complexity 1 arch/x86/include/asm/msr.h:wrmsrl
Cyclomatic Complexity 1 arch/x86/include/asm/msr.h:rdmsrl_safe
Cyclomatic Complexity 1 arch/x86/include/asm/processor.h:native_cpuid Cyclomatic Complexity 1 arch/x86/include/asm/processor.h:native_cpuid_eax Cyclomatic Complexity 1 arch/x86/include/asm/processor.h:cpuid
Cyclomatic Complexity 1 arch/x86/include/asm/processor.h:cpuid_count Cyclomatic Complexity 1 arch/x86/include/asm/processor.h:cpuid_eax Cyclomatic Complexity 2 include/linux/thread_info.h:test_ti_thread_flag Cyclomatic Complexity 2 include/linux/thread_info.h:check_object_size Cyclomatic Complexity 2 include/linux/thread_info.h:copy_overflow
Cyclomatic Complexity 4 include/linux/thread_info.h:check_copy_size Cyclomatic Complexity 2 include/linux/uaccess.h:copy_from_user
Cyclomatic Complexity 1 arch/x86/include/asm/microcode_intel.h:intel_get_microcode_revision Cyclomatic Complexity 1 arch/x86/kernel/cpu/intel.c:forcempx_setup Cyclomatic Complexity 1 arch/x86/kernel/cpu/intel.c:ring3mwait_disable Cyclomatic Complexity 5 arch/x86/kernel/cpu/intel.c:probe_xeon_phi_r3mwait Cyclomatic Complexity 6 arch/x86/kernel/cpu/intel.c:bad_spectre_microcode Cyclomatic Complexity 1 arch/x86/kernel/cpu/intel.c:forcepae_setup Cyclomatic Complexity 1 arch/x86/kernel/cpu/intel.c:srat_detect_node Cyclomatic Complexity 4 arch/x86/kernel/cpu/intel.c:init_cpuid_fault Cyclomatic Complexity 2 arch/x86/kernel/cpu/intel.c:init_intel_misc_features Cyclomatic Complexity 4 arch/x86/kernel/cpu/intel.c:intel_size_cache Cyclomatic Complexity 38 arch/x86/kernel/cpu/intel.c:intel_tlb_lookup Cyclomatic Complexity 6 arch/x86/kernel/cpu/intel.c:intel_detect_tlb Cyclomatic Complexity 5 arch/x86/kernel/cpu/intel.c:init_intel_energy_perf Cyclomatic Complexity 1 arch/x86/kernel/cpu/intel.c:intel_bsp_resume Cyclomatic Complexity 16 arch/x86/kernel/cpu/intel.c:detect_tme
Cyclomatic Complexity 9 arch/x86/kernel/cpu/intel.c:intel_smp_check Cyclomatic Complexity 9 arch/x86/kernel/cpu/intel.c:detect_vmx_virtcap Cyclomatic Complexity 13 arch/x86/kernel/cpu/intel.c:intel_workarounds Cyclomatic Complexity 1 arch/x86/include/asm/microcode.h:exit_amd_microcode Cyclomatic Complexity 4 arch/x86/kernel/cpu/intel.c:check_mpx_erratum Cyclomatic Complexity 37 arch/x86/kernel/cpu/intel.c:early_init_intel Cyclomatic Complexity 30 arch/x86/kernel/cpu/intel.c:init_intel
Cyclomatic Complexity 3 arch/x86/kernel/cpu/intel.c:ppro_with_ram_bug Cyclomatic Complexity 8 arch/x86/kernel/cpu/intel.c:fixup_cpuid_exception Cyclomatic Complexity 4 arch/x86/kernel/cpu/intel.c:init_core_capability
--
In file included from arch/x86//kernel/cpu/intel.c:22:0:

>> arch/x86/include/asm/insn-eval.h:18:58: warning: 'struct insn' declared inside parameter list
    void __user *insn_get_addr_ref(struct insn *insn, struct pt_regs *regs);
                                                             ^
>> arch/x86/include/asm/insn-eval.h:18:58: warning: its scope is only this definition or declaration, which is probably not what you want
   arch/x86/include/asm/insn-eval.h:19:53: warning: 'struct insn' declared inside parameter list
    int insn_get_modrm_rm_off(struct insn *insn, struct pt_regs *regs);
                                                        ^

Cyclomatic Complexity 5 include/linux/compiler.h:__read_once_size
Cyclomatic Complexity 1 include/linux/kasan-checks.h:kasan_check_read Cyclomatic Complexity 2 arch/x86/include/asm/bitops.h:set_bit
Cyclomatic Complexity 1 arch/x86/include/asm/bitops.h:constant_test_bit Cyclomatic Complexity 1 arch/x86/include/asm/bitops.h:variable_test_bit Cyclomatic Complexity 1 arch/x86/include/asm/bitops.h:fls
Cyclomatic Complexity 2 include/linux/bitops.h:get_count_order
Cyclomatic Complexity 1 include/linux/string.h:strnlen
Cyclomatic Complexity 4 include/linux/string.h:strlen
Cyclomatic Complexity 4 include/linux/string.h:memcpy
Cyclomatic Complexity 2 include/linux/string.h:strcpy
Cyclomatic Complexity 1 arch/x86/include/asm/atomic.h:arch_atomic_read Cyclomatic Complexity 1 include/asm-generic/atomic-instrumented.h:atomic_read Cyclomatic Complexity 1 arch/x86/include/asm/current.h:get_current Cyclomatic Complexity 1 arch/x86/include/asm/irqflags.h:native_save_fl Cyclomatic Complexity 1 arch/x86/include/asm/irqflags.h:native_restore_fl Cyclomatic Complexity 1 arch/x86/include/asm/irqflags.h:native_irq_disable Cyclomatic Complexity 1 arch/x86/include/asm/irqflags.h:arch_local_save_flags Cyclomatic Complexity 1 arch/x86/include/asm/irqflags.h:arch_local_irq_restore Cyclomatic Complexity 1 arch/x86/include/asm/irqflags.h:arch_local_irq_disable Cyclomatic Complexity 1 arch/x86/include/asm/irqflags.h:arch_local_irq_save Cyclomatic Complexity 1 include/linux/jump_label.h:static_key_count Cyclomatic Complexity 2 include/linux/jump_label.h:static_key_false Cyclomatic Complexity 1 arch/x86/include/asm/ptrace.h:user_64bit_mode Cyclomatic Complexity 1 arch/x86/include/asm/msr.h:__rdmsr
Cyclomatic Complexity 1 arch/x86/include/asm/msr.h:__wrmsr
Cyclomatic Complexity 2 arch/x86/include/asm/msr.h:native_read_msr Cyclomatic Complexity 2 arch/x86/include/asm/msr.h:native_read_msr_safe Cyclomatic Complexity 2 arch/x86/include/asm/msr.h:native_write_msr Cyclomatic Complexity 1 arch/x86/include/asm/msr.h:wrmsrl
Cyclomatic Complexity 1 arch/x86/include/asm/msr.h:rdmsrl_safe
Cyclomatic Complexity 1 arch/x86/include/asm/processor.h:native_cpuid Cyclomatic Complexity 1 arch/x86/include/asm/processor.h:native_cpuid_eax Cyclomatic Complexity 1 arch/x86/include/asm/processor.h:cpuid
Cyclomatic Complexity 1 arch/x86/include/asm/processor.h:cpuid_count Cyclomatic Complexity 1 arch/x86/include/asm/processor.h:cpuid_eax Cyclomatic Complexity 2 include/linux/thread_info.h:test_ti_thread_flag Cyclomatic Complexity 2 include/linux/thread_info.h:check_object_size Cyclomatic Complexity 2 include/linux/thread_info.h:copy_overflow
Cyclomatic Complexity 4 include/linux/thread_info.h:check_copy_size Cyclomatic Complexity 2 include/linux/uaccess.h:copy_from_user
Cyclomatic Complexity 1 arch/x86/include/asm/microcode_intel.h:intel_get_microcode_revision Cyclomatic Complexity 1 arch/x86//kernel/cpu/intel.c:forcempx_setup Cyclomatic Complexity 1 arch/x86//kernel/cpu/intel.c:ring3mwait_disable Cyclomatic Complexity 5 arch/x86//kernel/cpu/intel.c:probe_xeon_phi_r3mwait Cyclomatic Complexity 6 arch/x86//kernel/cpu/intel.c:bad_spectre_microcode Cyclomatic Complexity 1 arch/x86//kernel/cpu/intel.c:forcepae_setup Cyclomatic Complexity 1 arch/x86//kernel/cpu/intel.c:srat_detect_node Cyclomatic Complexity 4 arch/x86//kernel/cpu/intel.c:init_cpuid_fault Cyclomatic Complexity 2 arch/x86//kernel/cpu/intel.c:init_intel_misc_features Cyclomatic Complexity 4 arch/x86//kernel/cpu/intel.c:intel_size_cache Cyclomatic Complexity 38 arch/x86//kernel/cpu/intel.c:intel_tlb_lookup Cyclomatic Complexity 6 arch/x86//kernel/cpu/intel.c:intel_detect_tlb Cyclomatic Complexity 5 arch/x86//kernel/cpu/intel.c:init_intel_energy_perf Cyclomatic Complexity 1 arch/x86//kernel/cpu/intel.c:intel_bsp_resume Cyclomatic Complexity 16 arch/x86//kernel/cpu/intel.c:detect_tme
Cyclomatic Complexity 9 arch/x86//kernel/cpu/intel.c:intel_smp_check Cyclomatic Complexity 9 arch/x86//kernel/cpu/intel.c:detect_vmx_virtcap Cyclomatic Complexity 13 arch/x86//kernel/cpu/intel.c:intel_workarounds Cyclomatic Complexity 1 arch/x86/include/asm/microcode.h:exit_amd_microcode Cyclomatic Complexity 4 arch/x86//kernel/cpu/intel.c:check_mpx_erratum Cyclomatic Complexity 37 arch/x86//kernel/cpu/intel.c:early_init_intel Cyclomatic Complexity 30 arch/x86//kernel/cpu/intel.c:init_intel
Cyclomatic Complexity 3 arch/x86//kernel/cpu/intel.c:ppro_with_ram_bug Cyclomatic Complexity 8 arch/x86//kernel/cpu/intel.c:fixup_cpuid_exception Cyclomatic Complexity 4 arch/x86//kernel/cpu/intel.c:init_core_capability

vim +18 arch/x86/include/asm/insn-eval.h

4efea85f Ricardo Neri    2017-10-27  17  
32542ee2 Ricardo Neri    2017-10-27 @18  void __user *insn_get_addr_ref(struct insn *insn, struct pt_regs *regs);
e5e45f11 Ricardo Neri    2017-10-27  19  int insn_get_modrm_rm_off(struct insn *insn, struct pt_regs *regs);
bd5a410a Ricardo Neri    2017-10-27  20  unsigned long insn_get_seg_base(struct pt_regs *regs, int seg_reg_idx);

e2a5dca7 Borislav Petkov 2017-11-23 21 int insn_get_code_seg_params(struct pt_regs *regs);
32542ee2 Ricardo Neri 2017-10-27 22

:::::: The code at line 18 was first introduced by commit
:::::: 32542ee295bec38e5e1608f8c9d6d28e5a7e6112 x86/mpx, x86/insn: Relocate insn util functions to a new insn-eval file

:::::: TO: Ricardo Neri <[email protected]> :::::: CC: Thomas Gleixner <[email protected]>

---
0-DAY kernel test infrastructure Open Source Technology Center
https://lists.01.org/pipermail/kbuild-all Intel Corporation

author
#40 lore raw

PATCH v4 12/17 Change document for kernel option clearcpuid

In reply to Fenghua Yu

Since kernel option clearcpuid now supports multiple options and CPU capability flags, the document needs to be changed.

---
 Documentation/admin-guide/kernel-parameters.txt | 6 ++++--
 1 file changed, 4 insertions(+), 2 deletions(-)
Documentation/admin-guide/kernel-parameters.txt +4 -2
index 858b6c0b9a15..1a195f93d3cc 100644
@@ -558,10 +558,12 @@
 			loops can be debugged more effectively on production
 			systems.
 
-	clearcpuid=BITNUM [X86]
+	clearcpuid=BITNUM | FLAG [X86]
 			Disable CPUID feature X for the kernel. See
 			arch/x86/include/asm/cpufeatures.h for the valid bit
-			numbers. Note the Linux specific bits are not necessarily
+			numbers or /proc/cpuinfo for valid CPU flags.
+			Multiple options can be used to disable a few features.
+			Note the Linux specific bits are not necessarily
 			stable over kernel options, but the vendor specific
 			ones should be.
 			Also note that user programs calling CPUID directly
Signature · 1 line
2.7.4
author
#41 lore raw

PATCH v4 13/17 x86/split_lock: Handle #AC exception for split lock

In reply to Fenghua Yu

There may be different considerations on how to handle #AC for split lock, e.g. how to handle system hang caused by split lock issue in firmware, how to emulate faulting instruction, etc. We use a simple method to handle user and kernel split lock and may extend the method in the future.

When #AC exception for split lock is triggered from user process, the process is killed by SIGBUS. To execute the process properly, user application developer needs to fix the split lock issue.

When #AC exception for split lock is triggered from a kernel instruction, disable #AC for split lock on local CPU and warn the split lock issue. After the exception, the faulting instruction will be executed and kernel execution continues. #AC for split lock is only disabled on the local CPU not globally. It will be re-enabled if the CPU is offline and then online.

Kernel developer should check the warning, which contains helpful faulting address, context, and callstack info, and fix the split lock issue one by one. Then further split lock may be captured and fixed.

After bit 29 in MSR_TEST_CTL is set as one in kernel, firmware inherits the setting when firmware is executed in S4, S5, run time services, SMI, etc. Split lock issue in firmware triggers #AC and may hang the system depending on how firmware handles the #AC. It's up to firmware developer to fix the split lock issues in firmware.

---
 arch/x86/include/asm/msr-index.h |  4 ++++
 arch/x86/kernel/traps.c          | 31 +++++++++++++++++++++++++++++++
 2 files changed, 35 insertions(+)
arch/x86/include/asm/msr-index.h +4 -0
index 549e73dcca15..d2d43ebcb8ab 100644
@@ -39,6 +39,10 @@
 
 /* Intel MSRs. Some also available on other CPUs */
 
+#define MSR_TEST_CTL					0x00000033
+#define TEST_CTL_ENABLE_SPLIT_LOCK_DETECT_SHIFT	29
+#define TEST_CTL_ENABLE_SPLIT_LOCK_DETECT		BIT(29)
+
 #define MSR_IA32_SPEC_CTRL		0x00000048 /* Speculation Control */
 #define SPEC_CTRL_IBRS			(1 << 0)   /* Indirect Branch Restricted Speculation */
 #define SPEC_CTRL_STIBP_SHIFT		1	   /* Single Thread Indirect Branch Predictor (STIBP) bit */
arch/x86/kernel/traps.c +31 -0
index c43c2608332e..63a4dc985340 100644
@@ -61,6 +61,7 @@
 #include <asm/mpx.h>
 #include <asm/vm86.h>
 #include <asm/umip.h>
+#include <asm/cpu.h>
 
 #ifdef CONFIG_X86_64
 #include <asm/x86_init.h>
@@ -292,7 +293,37 @@ DO_ERROR(X86_TRAP_OLD_MF, SIGFPE,           0, NULL, "coprocessor segment overru
 DO_ERROR(X86_TRAP_TS,     SIGSEGV,          0, NULL, "invalid TSS",         invalid_TSS)
 DO_ERROR(X86_TRAP_NP,     SIGBUS,           0, NULL, "segment not present", segment_not_present)
 DO_ERROR(X86_TRAP_SS,     SIGBUS,           0, NULL, "stack segment",       stack_segment)
+#ifndef CONFIG_CPU_SUP_INTEL
 DO_ERROR(X86_TRAP_AC,     SIGBUS,  BUS_ADRALN, NULL, "alignment check",     alignment_check)
+#else
+dotraplinkage void do_alignment_check(struct pt_regs *regs, long error_code)
+{
+	unsigned int trapnr = X86_TRAP_AC;
+	char str[] = "alignment check";
+	int signr = SIGBUS;
+
+	RCU_LOCKDEP_WARN(!rcu_is_watching(), "entry code didn't wake RCU");
+
+	if (notify_die(DIE_TRAP, str, regs, error_code, trapnr, signr) !=
+			NOTIFY_STOP) {
+		cond_local_irq_enable(regs);
+		if (!user_mode(regs)) {
+			/*
+			 * Only split lock can generate #AC from kernel. Warn
+			 * and disable #AC for split lock on current CPU.
+			 */
+			msr_clear_bit(MSR_TEST_CTL,
+				      TEST_CTL_ENABLE_SPLIT_LOCK_DETECT_SHIFT);
+			WARN_ONCE(1, "A split lock issue is detected.\n");
+
+			return;
+		}
+		/* Handle #AC generated from user code. */
+		do_trap(X86_TRAP_AC, SIGBUS, "alignment check", regs,
+			error_code, BUS_ADRALN, NULL);
+	}
+}
+#endif
 #undef IP
 
 #ifdef CONFIG_VMAP_STACK
Signature · 1 line
2.7.4
author
#42 lore raw

PATCH v4 14/17 x86/split_lock: Add a sysfs interface to allow user to enable or disable split lock detection on all CPUs during run time

In reply to Fenghua Yu

The interface /sys/device/system/cpu/split_lock_detect is added to allow user to control split lock detection and show current split lock detection setting.

Writing 1 to the file enables split lock detection and writing 0 disables split lock detection. Split lock detection is enabled or disabled on all CPUs.

Reading the file shows current global split lock detection setting:
disabled when 0 and enabled when 1.

Please note the interface only shows global setting. During run time, split lock detection on one CPU may be disabled if split lock in kernel code happens on the CPU. The interface doesn't show split lock detection status on individual CPU. User can run rdmsr on 0x33 to check split lock detection on each CPU.

---
 arch/x86/kernel/cpu/intel.c | 96 ++++++++++++++++++++++++++++++++++++++++++++-
 1 file changed, 95 insertions(+), 1 deletion(-)
arch/x86/kernel/cpu/intel.c +95 -1
index 1c1ec413a9a9..1512e96287b8 100644
@@ -33,6 +33,12 @@
 #include <asm/apic.h>
 #endif
 
+#define DISABLE_SPLIT_LOCK_DETECT 0
+#define ENABLE_SPLIT_LOCK_DETECT  1
+
+static DEFINE_MUTEX(split_lock_detect_mutex);
+static int split_lock_detect_val;
+
 /*
  * Just in case our CPU detection goes bad, or you have a weird system,
  * allow a way to override the automatic disabling of MPX.
@@ -163,10 +169,35 @@ static bool bad_spectre_microcode(struct cpuinfo_x86 *c)
 	return false;
 }
 
+static u32 new_sp_test_ctl_val(u32 test_ctl_val)
+{
+	/* Change the split lock setting. */
+	if (split_lock_detect_val == DISABLE_SPLIT_LOCK_DETECT)
+		test_ctl_val &= ~TEST_CTL_ENABLE_SPLIT_LOCK_DETECT;
+	else
+		test_ctl_val |= TEST_CTL_ENABLE_SPLIT_LOCK_DETECT;
+
+	return test_ctl_val;
+}
+
+static void init_split_lock_detect(struct cpuinfo_x86 *c)
+{
+	if (cpu_has(c, X86_FEATURE_SPLIT_LOCK_DETECT)) {
+		u32 l, h;
+
+		rdmsr(MSR_TEST_CTL, l, h);
+		l = new_sp_test_ctl_val(l);
+		wrmsr(MSR_TEST_CTL, l, h);
+		pr_info_once("#AC for split lock is enabled\n");
+	}
+}
+
 static void early_init_intel(struct cpuinfo_x86 *c)
 {
 	u64 misc_enable;
 
+	init_split_lock_detect(c);
+
 	/* Unmask CPUID levels if masked: */
 	if (c->x86 > 6 || (c->x86 == 6 && c->x86_model >= 0xd)) {
 		if (msr_clear_bit(MSR_IA32_MISC_ENABLE,
@@ -1095,7 +1126,70 @@ void init_core_capability(struct cpuinfo_x86 *c)
 
 		rdmsrl(MSR_IA32_CORE_CAPABILITY, val);
 
-		if (val & CORE_CAP_SPLIT_LOCK_DETECT)
+		if (val & CORE_CAP_SPLIT_LOCK_DETECT) {
 			setup_force_cpu_cap(X86_FEATURE_SPLIT_LOCK_DETECT);
+			split_lock_detect_val = 1;
+		}
 	}
 }
+
+static ssize_t
+split_lock_detect_show(struct device *dev, struct device_attribute *attr,
+		       char *buf)
+{
+	return sprintf(buf, "%u\n", READ_ONCE(split_lock_detect_val));
+}
+
+static ssize_t
+split_lock_detect_store(struct device *dev, struct device_attribute *attr,
+			const char *buf, size_t count)
+{
+	u32 val, l, h;
+	int cpu, ret;
+
+	ret = kstrtou32(buf, 10, &val);
+	if (ret)
+		return ret;
+
+	if (val != DISABLE_SPLIT_LOCK_DETECT && val != ENABLE_SPLIT_LOCK_DETECT)
+		return -EINVAL;
+
+	/* No need to update MSR if new setting is the same as old one. */
+	if (val == split_lock_detect_val)
+		return count;
+
+	/* Only one write can happen at the same time. */
+	mutex_lock(&split_lock_detect_mutex);
+
+	WRITE_ONCE(split_lock_detect_val, val);
+
+	/* Get MSR_TEST_CTL on this CPU, assuming all CPUs have same value. */
+	rdmsr(MSR_TEST_CTL, l, h);
+	l = new_sp_test_ctl_val(l);
+	/* Update the split lock detection setting on all online CPUs. */
+	for_each_online_cpu(cpu)
+		wrmsr_on_cpu(cpu, MSR_TEST_CTL, l, h);
+
+	mutex_unlock(&split_lock_detect_mutex);
+
+	return count;
+}
+
+static DEVICE_ATTR_RW(split_lock_detect);
+
+static int __init split_lock_init(void)
+{
+	int ret;
+
+	if (!boot_cpu_has(X86_FEATURE_SPLIT_LOCK_DETECT))
+		return -ENODEV;
+
+	ret = device_create_file(cpu_subsys.dev_root,
+				 &dev_attr_split_lock_detect);
+	if (ret)
+		return ret;
+
+	return 0;
+}
+
+subsys_initcall(split_lock_init);
Signature · 1 line
2.7.4
author
#43 lore raw

PATCH v4 15/17 kvm: x86: Report CORE_CAPABILITY on GET_SUPPORTED_CPUID

In reply to Fenghua Yu

From: Xiaoyao Li <[email protected]>

In the latest Intel SDM, CPUID.(EAX=7H,ECX=0):EDX[30] will enumerate the presence of the IA32_CORE_CAPABILITY MSR.

Update GET_SUPPORTED_CPUID to expose this feature bit to user space, so that user space know this bit can be enabled in CPUID.

---
 arch/x86/kvm/cpuid.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)
arch/x86/kvm/cpuid.c +2 -1
index c07958b59f50..e0e17b9c65da 100644
@@ -410,7 +410,8 @@ static inline int __do_cpuid_ent(struct kvm_cpuid_entry2 *entry, u32 function,
 	/* cpuid 7.0.edx*/
 	const u32 kvm_cpuid_7_0_edx_x86_features =
 		F(AVX512_4VNNIW) | F(AVX512_4FMAPS) | F(SPEC_CTRL) |
-		F(SPEC_CTRL_SSBD) | F(ARCH_CAPABILITIES) | F(INTEL_STIBP);
+		F(SPEC_CTRL_SSBD) | F(ARCH_CAPABILITIES) | F(CORE_CAPABILITY) |
+		F(INTEL_STIBP);
 
 	/* all calls to cpuid_count() should be made on the same cpu */
 	get_cpu();
Signature · 1 line
2.7.4
maintainer
#44 lore raw

Re: [PATCH v4 15/17] kvm: x86: Report CORE_CAPABILITY on GET_SUPPORTED_CPUID

Quoted 18 lines · Fenghua Yu

From: Xiaoyao Li <[email protected]>

In the latest Intel SDM, CPUID.(EAX=7H,ECX=0):EDX[30] will enumerate the presence of the IA32_CORE_CAPABILITY MSR.

Update GET_SUPPORTED_CPUID to expose this feature bit to user space, so that user space know this bit can be enabled in CPUID.

---
 arch/x86/kvm/cpuid.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)
arch/x86/kvm/cpuid.c +0 -0
index c07958b59f50..e0e17b9c65da 100644

arch/x86/kvm/cpuid.c static inline int __do_cpuid_ent(struct kvm_cpuid_entry2 *entry, u32 function, · line 414

Quoted 6 lines
arch/x86/kvm/cpuid.c +2 -1
@@ -410,7 +410,8 @@ static inline int __do_cpuid_ent(struct kvm_cpuid_entry2 *entry, u32 function,
 	/* cpuid 7.0.edx*/
 	const u32 kvm_cpuid_7_0_edx_x86_features =
 		F(AVX512_4VNNIW) | F(AVX512_4FMAPS) | F(SPEC_CTRL) |
-		F(SPEC_CTRL_SSBD) | F(ARCH_CAPABILITIES) | F(INTEL_STIBP);
+		F(SPEC_CTRL_SSBD) | F(ARCH_CAPABILITIES) | F(CORE_CAPABILITY) |
+		F(INTEL_STIBP);

This should be enabled always if boot_cpu_has(X86_FEATURE_SPLIT_LOCK_DETECT), since the MSR is emulated. This way, guests can always rely on IA32_CORE_CAPABILITY_MSR and it won't have to rely on the FMS (which means nothing inside a guest).

Paolo

Quoted 3 lines
arch/x86/kvm/cpuid.c +0 -0
 	/* all calls to cpuid_count() should be made on the same cpu */
 	get_cpu();
 
#45 lore raw

Re: [PATCH v4 15/17] kvm: x86: Report CORE_CAPABILITY on GET_SUPPORTED_CPUID

Quoted 32 lines · Paolo Bonzini
Quoted 18 lines · Fenghua Yu

From: Xiaoyao Li <[email protected]>

In the latest Intel SDM, CPUID.(EAX=7H,ECX=0):EDX[30] will enumerate the presence of the IA32_CORE_CAPABILITY MSR.

Update GET_SUPPORTED_CPUID to expose this feature bit to user space, so that user space know this bit can be enabled in CPUID.

---
 arch/x86/kvm/cpuid.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)
arch/x86/kvm/cpuid.c +0 -0
index c07958b59f50..e0e17b9c65da 100644

arch/x86/kvm/cpuid.c static inline int __do_cpuid_ent(struct kvm_cpuid_entry2 · line 415

Quoted 7 lines
arch/x86/kvm/cpuid.c +2 -1
@@ -410,7 +410,8 @@ static inline int __do_cpuid_ent(struct kvm_cpuid_entry2
 *entry, u32 function,
 	/* cpuid 7.0.edx*/
 	const u32 kvm_cpuid_7_0_edx_x86_features =
 		F(AVX512_4VNNIW) | F(AVX512_4FMAPS) | F(SPEC_CTRL) |
-		F(SPEC_CTRL_SSBD) | F(ARCH_CAPABILITIES) | F(INTEL_STIBP);
+		F(SPEC_CTRL_SSBD) | F(ARCH_CAPABILITIES) | F(CORE_CAPABILITY) |
+		F(INTEL_STIBP);

This should be enabled always if boot_cpu_has(X86_FEATURE_SPLIT_LOCK_DETECT), since the MSR is emulated. This way, guests can always rely on IA32_CORE_CAPABILITY_MSR and it won't have to rely on the FMS (which means nothing inside a guest).

Quoted 1 line

Paolo

Hi, Paolo
Do you mean that we don't need this here, but to add the handling below?

static inline int __do_cpuid_ent(struct kvm_cpuid_entry2 * entry, u32 function,
...

	switch (function) {
	...
	case 7: {
		...
		if (index ==0) {
			...
			if(boot_cpu_has(X86_FEATURE_SPLIT_LOCK_DETECT))
				entry->edx |= F(CORE_CAPABILITY);
		}
		...
	...
	}

...
}

Quoted 5 lines
Quoted 3 lines
arch/x86/kvm/cpuid.c +0 -0
 	/* all calls to cpuid_count() should be made on the same cpu */
 	get_cpu();
maintainer
#46 lore raw

Re: [PATCH v4 15/17] kvm: x86: Report CORE_CAPABILITY on GET_SUPPORTED_CPUID

Quoted 52 lines · Xiaoyao Li
Quoted 32 lines · Paolo Bonzini
Quoted 18 lines · Fenghua Yu

From: Xiaoyao Li <[email protected]>

In the latest Intel SDM, CPUID.(EAX=7H,ECX=0):EDX[30] will enumerate the presence of the IA32_CORE_CAPABILITY MSR.

Update GET_SUPPORTED_CPUID to expose this feature bit to user space, so that user space know this bit can be enabled in CPUID.

---
 arch/x86/kvm/cpuid.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)
arch/x86/kvm/cpuid.c +0 -0
index c07958b59f50..e0e17b9c65da 100644

arch/x86/kvm/cpuid.c static inline int __do_cpuid_ent(struct kvm_cpuid_entry2 · line 415

Quoted 7 lines
arch/x86/kvm/cpuid.c +2 -1
@@ -410,7 +410,8 @@ static inline int __do_cpuid_ent(struct kvm_cpuid_entry2
 *entry, u32 function,
 	/* cpuid 7.0.edx*/
 	const u32 kvm_cpuid_7_0_edx_x86_features =
 		F(AVX512_4VNNIW) | F(AVX512_4FMAPS) | F(SPEC_CTRL) |
-		F(SPEC_CTRL_SSBD) | F(ARCH_CAPABILITIES) | F(INTEL_STIBP);
+		F(SPEC_CTRL_SSBD) | F(ARCH_CAPABILITIES) | F(CORE_CAPABILITY) |
+		F(INTEL_STIBP);

This should be enabled always if boot_cpu_has(X86_FEATURE_SPLIT_LOCK_DETECT), since the MSR is emulated. This way, guests can always rely on IA32_CORE_CAPABILITY_MSR and it won't have to rely on the FMS (which means nothing inside a guest).

Quoted 1 line

Paolo

Hi, Paolo
Do you mean that we don't need this here, but to add the handling below?

static inline int __do_cpuid_ent(struct kvm_cpuid_entry2 * entry, u32 function,
...

	switch (function) {
	...
	case 7: {
		...
		if (index ==0) {
			...
			if(boot_cpu_has(X86_FEATURE_SPLIT_LOCK_DETECT))
				entry->edx |= F(CORE_CAPABILITY);
		}
		...
	...
	}
Quoted 2 lines

...
}

Yes, exactly.

Paolo

Quoted 6 lines
Quoted 5 lines
Quoted 3 lines
arch/x86/kvm/cpuid.c +0 -0
 	/* all calls to cpuid_count() should be made on the same cpu */
 	get_cpu();
#47 lore raw

Re: [PATCH v4 15/17] kvm: x86: Report CORE_CAPABILITY on GET_SUPPORTED_CPUID

Quoted 61 lines · Paolo Bonzini
Quoted 57 lines · Xiaoyao Li
Quoted 36 lines · Paolo Bonzini
Quoted 27 lines · Fenghua Yu

From: Xiaoyao Li <[email protected]>

In the latest Intel SDM, CPUID.(EAX=7H,ECX=0):EDX[30] will enumerate the presence of the IA32_CORE_CAPABILITY MSR.

Update GET_SUPPORTED_CPUID to expose this feature bit to user space, so that user space know this bit can be enabled in CPUID.

---
 arch/x86/kvm/cpuid.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)
arch/x86/kvm/cpuid.c +0 -0
index c07958b59f50..e0e17b9c65da 100644
@@ -410,7 +410,8 @@ static inline int __do_cpuid_ent(struct

kvm_cpuid_entry2
*entry, u32 function,

arch/x86/kvm/cpuid.c +0 -1
 	/* cpuid 7.0.edx*/
 	const u32 kvm_cpuid_7_0_edx_x86_features =
 		F(AVX512_4VNNIW) | F(AVX512_4FMAPS) | F(SPEC_CTRL) |
-		F(SPEC_CTRL_SSBD) | F(ARCH_CAPABILITIES) |

F(INTEL_STIBP);

arch/x86/kvm/cpuid.c +1 -0
+		F(SPEC_CTRL_SSBD) | F(ARCH_CAPABILITIES) |

F(CORE_CAPABILITY) |

arch/x86/kvm/cpuid.c static inline int __do_cpuid_ent(struct kvm_cpuid_entry2 *entry, u32 function, · line 414

Quoted 1 line
arch/x86/kvm/cpuid.c +1 -0
+		F(INTEL_STIBP);

This should be enabled always if boot_cpu_has(X86_FEATURE_SPLIT_LOCK_DETECT),
since the MSR is emulated. This way, guests can always rely on IA32_CORE_CAPABILITY_MSR and it won't have to rely on the FMS (which means nothing inside a guest).

Quoted 1 line

Paolo

Hi, Paolo
Do you mean that we don't need this here, but to add the handling below?

static inline int __do_cpuid_ent(struct kvm_cpuid_entry2 * entry, u32 function,
...

	switch (function) {
	...
	case 7: {
		...
		if (index ==0) {
			...
			if(boot_cpu_has(X86_FEATURE_SPLIT_LOCK_DETECT))
				entry->edx |= F(CORE_CAPABILITY);
		}
		...
	...
	}
Quoted 2 lines

...
}

Quoted 3 lines

Yes, exactly.

Paolo

Like you said before, I think we don't need the condition judgment "if(boot_cpu_has(X86_FEATURE_SPLIT_LOCK_DETECT))", but to set F(CORE_CAPABILITY) always for guest since MSR_IA32_CORE_CAPABILITY is emulated.

And we should set the right emulated value of MSR_IA32_CORE_CAPABILITY for guest in function kvm_get_core_capability() based on whether boot_cpu_has(X86_FEATURE_SPLIT_LOCK_DETECT) just as you commented in the next patch.

Xiaoyao

Quoted 7 lines
Quoted 5 lines
Quoted 5 lines
Quoted 3 lines
arch/x86/kvm/cpuid.c +0 -0
 	/* all calls to cpuid_count() should be made on the same cpu */
 	get_cpu();
maintainer
#48 lore raw

Re: [PATCH v4 15/17] kvm: x86: Report CORE_CAPABILITY on GET_SUPPORTED_CPUID

Quoted 8 lines · Xiaoyao Li

Like you said before, I think we don't need the condition judgment "if(boot_cpu_has(X86_FEATURE_SPLIT_LOCK_DETECT))", but to set F(CORE_CAPABILITY) always for guest since MSR_IA32_CORE_CAPABILITY is emulated.

And we should set the right emulated value of MSR_IA32_CORE_CAPABILITY for guest in function kvm_get_core_capability() based on whether boot_cpu_has(X86_FEATURE_SPLIT_LOCK_DETECT) just as you commented in the next patch.

Yes, that would certainly be better. However, you'd also have to move MSR_IA32_CORE_CAPABILITY handling to x86.c, because you'd have to enable X86_FEATURE_CORE_CAPABILITY for AMD.

Paolo

#49 lore raw

Re: [PATCH v4 15/17] kvm: x86: Report CORE_CAPABILITY on GET_SUPPORTED_CPUID

Quoted 13 lines · Paolo Bonzini
Quoted 11 lines · Xiaoyao Li

Like you said before, I think we don't need the condition judgment "if(boot_cpu_has(X86_FEATURE_SPLIT_LOCK_DETECT))", but to set F(CORE_CAPABILITY)
always for guest since MSR_IA32_CORE_CAPABILITY is emulated.

And we should set the right emulated value of MSR_IA32_CORE_CAPABILITY for guest
in function kvm_get_core_capability() based on whether boot_cpu_has(X86_FEATURE_SPLIT_LOCK_DETECT) just as you commented in the next
patch.

Quoted 5 lines

Yes, that would certainly be better. However, you'd also have to move MSR_IA32_CORE_CAPABILITY handling to x86.c, because you'd have to enable X86_FEATURE_CORE_CAPABILITY for AMD.

Paolo

Thanks for your comments and advises.
I'll do it in next version.

Xiaoyao

#50 lore raw

Re: [PATCH v4 15/17] kvm: x86: Report CORE_CAPABILITY on GET_SUPPORTED_CPUID

In reply to Paolo Bonzini
Quoted 13 lines · Paolo Bonzini
Quoted 11 lines · Xiaoyao Li

Like you said before, I think we don't need the condition judgment "if(boot_cpu_has(X86_FEATURE_SPLIT_LOCK_DETECT))", but to set F(CORE_CAPABILITY)
always for guest since MSR_IA32_CORE_CAPABILITY is emulated.

And we should set the right emulated value of MSR_IA32_CORE_CAPABILITY for guest
in function kvm_get_core_capability() based on whether boot_cpu_has(X86_FEATURE_SPLIT_LOCK_DETECT) just as you commented in the next
patch.

Quoted 5 lines

Yes, that would certainly be better. However, you'd also have to move MSR_IA32_CORE_CAPABILITY handling to x86.c, because you'd have to enable X86_FEATURE_CORE_CAPABILITY for AMD.

Paolo

Hi, Paolo

I just notice that F(ARCH_CAPABILITIES) is set unconditionally. However the handling of MSR_IA32_ARCH_CAPABILITIES only exists with vmx, and the emulation of this MSR is in vmx->arch_capabilities.
These will cause #GP when guest kernel rdmsr(MSR_IA32_ARCH_CAPABILITES) with AMD CPU since there is handling for svm.
Maybe what I think is not correct due to my limit knowledge of MSR_IA32_ARCH_CAPABILITIES and how kernel handles its related features.

If what I said above is true and it's indeed an issue. So based on the fact that both MSR_IA32_ARCH_CAPABILITIES and MSR_IA32_CORE_CAPABILITY are feature- enumerating MSR and we emulate them in KVM, there are 2 choices for us to handle it:
1. we unconditionally set F(ARCH_CAPABILITIES) and F(CORE_CAPABILITY) for guest, move the emulation of these 2 MSRs to vcpu->arch.***, and move all the handling of these 2 MSRs to x86.c.

2. we conditionally set F(ARCH_CAPABILITIES) and F(CORE_CAPABILITY) only if it is intel CPU. So we just need to emulate these 2 MSRs in vmx->*** for intel CPU.

I prefer option 2 personally for CORE_CAPABILITY since it makes no sense to expose MSR_IA32_CORE_CAPABILITY to other x86 vendors.
About ARCH_CAPABILITIES, it seems that we emulate it for generic x86 cpus that !x86_match_cpu(cpu_no_speculation). So we should choose option 1, to move the emulation and handling to x86.c?

Xiaoyao

author
#51 lore raw

PATCH v4 16/17 kvm: x86: Add support IA32_CORE_CAPABILITY MSR

In reply to Fenghua Yu

From: Xiaoyao Li <[email protected]>

MSR IA32_CORE_CAPABILITY is a feature-enumerating MSR, bit 5 of which reports the capability of enabling detection of split locks (will be supported on future processors based on Tremont microarchitecture and later).

Please check the latest Intel Architecture Instruction Set Extensions and Future Features Programming Reference for more detailed information on the MSR and the split lock bit.

1. Expose it to user space as a feature-enumerating MSR, so that user space can query it.

2. Emualte MSR_IA32_CORE_CAPABILITY with vmx->core_capability. And add the get and set handler of MSR_IA32_CORE_CAPABILITY.
For uesrspace, it can set this MSR when customizing features of guest,
also it can read the value of this MSR of guest.
For guest, as it's a feature-enumerating MSR, guest only can read it.

---
 arch/x86/include/asm/kvm_host.h |  1 +
 arch/x86/kvm/vmx/vmx.c          | 23 +++++++++++++++++++++++
 arch/x86/kvm/vmx/vmx.h          |  1 +
 arch/x86/kvm/x86.c              | 17 ++++++++++++++++-
 4 files changed, 41 insertions(+), 1 deletion(-)
arch/x86/include/asm/kvm_host.h +1 -0
index 180373360e34..208f15570d17 100644
@@ -1527,6 +1527,7 @@ int kvm_pv_send_ipi(struct kvm *kvm, unsigned long ipi_bitmap_low,
 		    unsigned long icr, int op_64_bit);
 
 u64 kvm_get_arch_capabilities(void);
+u64 kvm_get_core_capability(void);
 void kvm_define_shared_msr(unsigned index, u32 msr);
 int kvm_set_shared_msr(unsigned index, u64 val, u64 mask);
 
arch/x86/kvm/vmx/vmx.c +23 -0
index 30a6bcd735ec..3e03c6e1e558 100644
@@ -1679,6 +1679,12 @@ static int vmx_get_msr(struct kvm_vcpu *vcpu, struct msr_data *msr_info)
 
 		msr_info->data = to_vmx(vcpu)->spec_ctrl;
 		break;
+	case MSR_IA32_CORE_CAPABILITY:
+		if (!msr_info->host_initiated &&
+		    !guest_cpuid_has(vcpu, X86_FEATURE_CORE_CAPABILITY))
+			return 1;
+		msr_info->data = vmx->core_capability;
+		break;
 	case MSR_IA32_ARCH_CAPABILITIES:
 		if (!msr_info->host_initiated &&
 		    !guest_cpuid_has(vcpu, X86_FEATURE_ARCH_CAPABILITIES))
@@ -1891,6 +1897,21 @@ static int vmx_set_msr(struct kvm_vcpu *vcpu, struct msr_data *msr_info)
 		vmx_disable_intercept_for_msr(vmx->vmcs01.msr_bitmap, MSR_IA32_PRED_CMD,
 					      MSR_TYPE_W);
 		break;
+	case MSR_IA32_CORE_CAPABILITY:
+		if (!msr_info->host_initiated)
+			return 1;
+		if (data & ~CORE_CAP_SPLIT_LOCK_DETECT)
+			return 1;
+
+		/*
+		 * Since AC split lock is a hardware feature, and there is no
+		 * software emulation yet, we cannot enable it for guest if
+		 * host hardware doesn't support it.
+		 */
+		if (!boot_cpu_has(X86_FEATURE_SPLIT_LOCK_DETECT))
+			data &= ~CORE_CAP_SPLIT_LOCK_DETECT;
+		vmx->core_capability = data;
+		break;
 	case MSR_IA32_ARCH_CAPABILITIES:
 		if (!msr_info->host_initiated)
 			return 1;
@@ -4083,6 +4104,8 @@ static void vmx_vcpu_setup(struct vcpu_vmx *vmx)
 		++vmx->nmsrs;
 	}
 
+	vmx->core_capability = kvm_get_core_capability();
+
 	vmx->arch_capabilities = kvm_get_arch_capabilities();
 
 	vm_exit_controls_init(vmx, vmx_vmexit_ctrl());
arch/x86/kvm/vmx/vmx.h +1 -0
index 0ac0a64c7790..cc22379991f3 100644
@@ -191,6 +191,7 @@ struct vcpu_vmx {
 	u64		      msr_guest_kernel_gs_base;
 #endif
 
+	u64		      core_capability;
 	u64		      arch_capabilities;
 	u64		      spec_ctrl;
 
arch/x86/kvm/x86.c +16 -1
index 941f932373d0..c3c9e3f2d08a 100644
@@ -1125,7 +1125,8 @@ static u32 msrs_to_save[] = {
 #endif
 	MSR_IA32_TSC, MSR_IA32_CR_PAT, MSR_VM_HSAVE_PA,
 	MSR_IA32_FEATURE_CONTROL, MSR_IA32_BNDCFGS, MSR_TSC_AUX,
-	MSR_IA32_SPEC_CTRL, MSR_IA32_ARCH_CAPABILITIES,
+	MSR_IA32_SPEC_CTRL, MSR_IA32_CORE_CAPABILITY,
+	MSR_IA32_ARCH_CAPABILITIES,
 	MSR_IA32_RTIT_CTL, MSR_IA32_RTIT_STATUS, MSR_IA32_RTIT_CR3_MATCH,
 	MSR_IA32_RTIT_OUTPUT_BASE, MSR_IA32_RTIT_OUTPUT_MASK,
 	MSR_IA32_RTIT_ADDR0_A, MSR_IA32_RTIT_ADDR0_B,
@@ -1197,11 +1198,22 @@ static u32 msr_based_features[] = {
 
 	MSR_F10H_DECFG,
 	MSR_IA32_UCODE_REV,
+	MSR_IA32_CORE_CAPABILITY,
 	MSR_IA32_ARCH_CAPABILITIES,
 };
 
 static unsigned int num_msr_based_features;
 
+u64 kvm_get_core_capability(void)
+{
+	u64 data;
+
+	rdmsrl_safe(MSR_IA32_CORE_CAPABILITY, &data);
+
+	return data;
+}
+EXPORT_SYMBOL_GPL(kvm_get_core_capability);
+
 u64 kvm_get_arch_capabilities(void)
 {
 	u64 data;
@@ -1227,6 +1239,9 @@ EXPORT_SYMBOL_GPL(kvm_get_arch_capabilities);
 static int kvm_get_msr_feature(struct kvm_msr_entry *msr)
 {
 	switch (msr->index) {
+	case MSR_IA32_CORE_CAPABILITY:
+		msr->data = kvm_get_core_capability();
+		break;
 	case MSR_IA32_ARCH_CAPABILITIES:
 		msr->data = kvm_get_arch_capabilities();
 		break;
Signature · 1 line
2.7.4
maintainer
#52 lore raw

Re: [PATCH v4 16/17] kvm: x86: Add support IA32_CORE_CAPABILITY MSR

Quoted 120 lines · Fenghua Yu

From: Xiaoyao Li <[email protected]>

MSR IA32_CORE_CAPABILITY is a feature-enumerating MSR, bit 5 of which reports the capability of enabling detection of split locks (will be supported on future processors based on Tremont microarchitecture and later).

Please check the latest Intel Architecture Instruction Set Extensions and Future Features Programming Reference for more detailed information on the MSR and the split lock bit.

1. Expose it to user space as a feature-enumerating MSR, so that user space can query it.

2. Emualte MSR_IA32_CORE_CAPABILITY with vmx->core_capability. And add the get and set handler of MSR_IA32_CORE_CAPABILITY.
For uesrspace, it can set this MSR when customizing features of guest,
also it can read the value of this MSR of guest.
For guest, as it's a feature-enumerating MSR, guest only can read it.

---
 arch/x86/include/asm/kvm_host.h |  1 +
 arch/x86/kvm/vmx/vmx.c          | 23 +++++++++++++++++++++++
 arch/x86/kvm/vmx/vmx.h          |  1 +
 arch/x86/kvm/x86.c              | 17 ++++++++++++++++-
 4 files changed, 41 insertions(+), 1 deletion(-)
arch/x86/include/asm/kvm_host.h +1 -0
index 180373360e34..208f15570d17 100644
@@ -1527,6 +1527,7 @@ int kvm_pv_send_ipi(struct kvm *kvm, unsigned long ipi_bitmap_low,
 		    unsigned long icr, int op_64_bit);
 
 u64 kvm_get_arch_capabilities(void);
+u64 kvm_get_core_capability(void);
 void kvm_define_shared_msr(unsigned index, u32 msr);
 int kvm_set_shared_msr(unsigned index, u64 val, u64 mask);
 
arch/x86/kvm/vmx/vmx.c +23 -0
index 30a6bcd735ec..3e03c6e1e558 100644
@@ -1679,6 +1679,12 @@ static int vmx_get_msr(struct kvm_vcpu *vcpu, struct msr_data *msr_info)
 
 		msr_info->data = to_vmx(vcpu)->spec_ctrl;
 		break;
+	case MSR_IA32_CORE_CAPABILITY:
+		if (!msr_info->host_initiated &&
+		    !guest_cpuid_has(vcpu, X86_FEATURE_CORE_CAPABILITY))
+			return 1;
+		msr_info->data = vmx->core_capability;
+		break;
 	case MSR_IA32_ARCH_CAPABILITIES:
 		if (!msr_info->host_initiated &&
 		    !guest_cpuid_has(vcpu, X86_FEATURE_ARCH_CAPABILITIES))
@@ -1891,6 +1897,21 @@ static int vmx_set_msr(struct kvm_vcpu *vcpu, struct msr_data *msr_info)
 		vmx_disable_intercept_for_msr(vmx->vmcs01.msr_bitmap, MSR_IA32_PRED_CMD,
 					      MSR_TYPE_W);
 		break;
+	case MSR_IA32_CORE_CAPABILITY:
+		if (!msr_info->host_initiated)
+			return 1;
+		if (data & ~CORE_CAP_SPLIT_LOCK_DETECT)
+			return 1;
+
+		/*
+		 * Since AC split lock is a hardware feature, and there is no
+		 * software emulation yet, we cannot enable it for guest if
+		 * host hardware doesn't support it.
+		 */
+		if (!boot_cpu_has(X86_FEATURE_SPLIT_LOCK_DETECT))
+			data &= ~CORE_CAP_SPLIT_LOCK_DETECT;
+		vmx->core_capability = data;
+		break;
 	case MSR_IA32_ARCH_CAPABILITIES:
 		if (!msr_info->host_initiated)
 			return 1;
@@ -4083,6 +4104,8 @@ static void vmx_vcpu_setup(struct vcpu_vmx *vmx)
 		++vmx->nmsrs;
 	}
 
+	vmx->core_capability = kvm_get_core_capability();
+
 	vmx->arch_capabilities = kvm_get_arch_capabilities();
 
 	vm_exit_controls_init(vmx, vmx_vmexit_ctrl());
arch/x86/kvm/vmx/vmx.h +1 -0
index 0ac0a64c7790..cc22379991f3 100644
@@ -191,6 +191,7 @@ struct vcpu_vmx {
 	u64		      msr_guest_kernel_gs_base;
 #endif
 
+	u64		      core_capability;
 	u64		      arch_capabilities;
 	u64		      spec_ctrl;
 
arch/x86/kvm/x86.c +3 -1
index 941f932373d0..c3c9e3f2d08a 100644
@@ -1125,7 +1125,8 @@ static u32 msrs_to_save[] = {
 #endif
 	MSR_IA32_TSC, MSR_IA32_CR_PAT, MSR_VM_HSAVE_PA,
 	MSR_IA32_FEATURE_CONTROL, MSR_IA32_BNDCFGS, MSR_TSC_AUX,
-	MSR_IA32_SPEC_CTRL, MSR_IA32_ARCH_CAPABILITIES,
+	MSR_IA32_SPEC_CTRL, MSR_IA32_CORE_CAPABILITY,
+	MSR_IA32_ARCH_CAPABILITIES,
 	MSR_IA32_RTIT_CTL, MSR_IA32_RTIT_STATUS, MSR_IA32_RTIT_CR3_MATCH,
 	MSR_IA32_RTIT_OUTPUT_BASE, MSR_IA32_RTIT_OUTPUT_MASK,
 	MSR_IA32_RTIT_ADDR0_A, MSR_IA32_RTIT_ADDR0_B,
@@ -1197,11 +1198,22 @@ static u32 msr_based_features[] = {
 
 	MSR_F10H_DECFG,
 	MSR_IA32_UCODE_REV,
+	MSR_IA32_CORE_CAPABILITY,

arch/x86/kvm/x86.c static u32 msr_based_features[] = { · line 1211

Quoted 10 lines
arch/x86/kvm/x86.c +5 -0
 	MSR_IA32_ARCH_CAPABILITIES,
 };
 
 static unsigned int num_msr_based_features;
 
+u64 kvm_get_core_capability(void)
+{
+	u64 data;
+
+	rdmsrl_safe(MSR_IA32_CORE_CAPABILITY, &data);

This patch should be merged with the previous patch. Also here you should add:

	data &= CORE_CAP_SPLIT_LOCK_DETECT;

so that non-virtualizable features are hidden and

	if (boot_cpu_has(X86_FEATURE_SPLIT_LOCK_DETECT))
		data |= CORE_CAP_SPLIT_LOCK_DETECT;

so that userspace gets "for free" the FMS list that will be added later to the kernel.

Thanks,

Paolo

Quoted 19 lines
arch/x86/kvm/x86.c +8 -0
+
+	return data;
+}
+EXPORT_SYMBOL_GPL(kvm_get_core_capability);
+
 u64 kvm_get_arch_capabilities(void)
 {
 	u64 data;
@@ -1227,6 +1239,9 @@ EXPORT_SYMBOL_GPL(kvm_get_arch_capabilities);
 static int kvm_get_msr_feature(struct kvm_msr_entry *msr)
 {
 	switch (msr->index) {
+	case MSR_IA32_CORE_CAPABILITY:
+		msr->data = kvm_get_core_capability();
+		break;
 	case MSR_IA32_ARCH_CAPABILITIES:
 		msr->data = kvm_get_arch_capabilities();
 		break;
#53 lore raw

Re: [PATCH v4 16/17] kvm: x86: Add support IA32_CORE_CAPABILITY MSR

Quoted 147 lines · Paolo Bonzini
Quoted 125 lines · Fenghua Yu

From: Xiaoyao Li <[email protected]>

MSR IA32_CORE_CAPABILITY is a feature-enumerating MSR, bit 5 of which reports the capability of enabling detection of split locks (will be supported on future processors based on Tremont microarchitecture and later).

Please check the latest Intel Architecture Instruction Set Extensions and Future Features Programming Reference for more detailed information on the MSR and the split lock bit.

1. Expose it to user space as a feature-enumerating MSR, so that user space can query it.

2. Emualte MSR_IA32_CORE_CAPABILITY with vmx->core_capability. And add the get and set handler of MSR_IA32_CORE_CAPABILITY.
For uesrspace, it can set this MSR when customizing features of guest,
also it can read the value of this MSR of guest.
For guest, as it's a feature-enumerating MSR, guest only can read it.

---
 arch/x86/include/asm/kvm_host.h |  1 +
 arch/x86/kvm/vmx/vmx.c          | 23 +++++++++++++++++++++++
 arch/x86/kvm/vmx/vmx.h          |  1 +
 arch/x86/kvm/x86.c              | 17 ++++++++++++++++-
 4 files changed, 41 insertions(+), 1 deletion(-)
arch/x86/include/asm/kvm_host.h +1 -0
index 180373360e34..208f15570d17 100644
@@ -1527,6 +1527,7 @@ int kvm_pv_send_ipi(struct kvm *kvm, unsigned long
 ipi_bitmap_low,
 		    unsigned long icr, int op_64_bit);
 
 u64 kvm_get_arch_capabilities(void);
+u64 kvm_get_core_capability(void);
 void kvm_define_shared_msr(unsigned index, u32 msr);
 int kvm_set_shared_msr(unsigned index, u64 val, u64 mask);
arch/x86/kvm/vmx/vmx.c +6 -0
index 30a6bcd735ec..3e03c6e1e558 100644
@@ -1679,6 +1679,12 @@ static int vmx_get_msr(struct kvm_vcpu *vcpu, struct
 msr_data *msr_info)
 
 		msr_info->data = to_vmx(vcpu)->spec_ctrl;
 		break;
+	case MSR_IA32_CORE_CAPABILITY:
+		if (!msr_info->host_initiated &&
+		    !guest_cpuid_has(vcpu, X86_FEATURE_CORE_CAPABILITY))
+			return 1;
+		msr_info->data = vmx->core_capability;
+		break;
 	case MSR_IA32_ARCH_CAPABILITIES:
 		if (!msr_info->host_initiated &&
arch/x86/kvm/vmx/vmx.c +15 -0
 		    !guest_cpuid_has(vcpu, X86_FEATURE_ARCH_CAPABILITIES))
@@ -1891,6 +1897,21 @@ static int vmx_set_msr(struct kvm_vcpu *vcpu, struct
 msr_data *msr_info)
 		vmx_disable_intercept_for_msr(vmx->vmcs01.msr_bitmap,
 MSR_IA32_PRED_CMD,
 					      MSR_TYPE_W);
 		break;
+	case MSR_IA32_CORE_CAPABILITY:
+		if (!msr_info->host_initiated)
+			return 1;
+		if (data & ~CORE_CAP_SPLIT_LOCK_DETECT)
+			return 1;
+
+		/*
+		 * Since AC split lock is a hardware feature, and there is no
+		 * software emulation yet, we cannot enable it for guest if
+		 * host hardware doesn't support it.
+		 */
+		if (!boot_cpu_has(X86_FEATURE_SPLIT_LOCK_DETECT))
+			data &= ~CORE_CAP_SPLIT_LOCK_DETECT;
+		vmx->core_capability = data;
+		break;
 	case MSR_IA32_ARCH_CAPABILITIES:
arch/x86/kvm/vmx/vmx.c +2 -0
 		if (!msr_info->host_initiated)
 			return 1;
@@ -4083,6 +4104,8 @@ static void vmx_vcpu_setup(struct vcpu_vmx *vmx)
 		++vmx->nmsrs;
 	}
 
+	vmx->core_capability = kvm_get_core_capability();
+
 	vmx->arch_capabilities = kvm_get_arch_capabilities();
 
 	vm_exit_controls_init(vmx, vmx_vmexit_ctrl());
arch/x86/kvm/vmx/vmx.h +1 -0
index 0ac0a64c7790..cc22379991f3 100644
@@ -191,6 +191,7 @@ struct vcpu_vmx {
 	u64		      msr_guest_kernel_gs_base;
 #endif
 
+	u64		      core_capability;
 	u64		      arch_capabilities;
 	u64		      spec_ctrl;
 
arch/x86/kvm/x86.c +3 -1
index 941f932373d0..c3c9e3f2d08a 100644
@@ -1125,7 +1125,8 @@ static u32 msrs_to_save[] = {
 #endif
 	MSR_IA32_TSC, MSR_IA32_CR_PAT, MSR_VM_HSAVE_PA,
 	MSR_IA32_FEATURE_CONTROL, MSR_IA32_BNDCFGS, MSR_TSC_AUX,
-	MSR_IA32_SPEC_CTRL, MSR_IA32_ARCH_CAPABILITIES,
+	MSR_IA32_SPEC_CTRL, MSR_IA32_CORE_CAPABILITY,
+	MSR_IA32_ARCH_CAPABILITIES,
 	MSR_IA32_RTIT_CTL, MSR_IA32_RTIT_STATUS, MSR_IA32_RTIT_CR3_MATCH,
 	MSR_IA32_RTIT_OUTPUT_BASE, MSR_IA32_RTIT_OUTPUT_MASK,
 	MSR_IA32_RTIT_ADDR0_A, MSR_IA32_RTIT_ADDR0_B,
@@ -1197,11 +1198,22 @@ static u32 msr_based_features[] = {
 
 	MSR_F10H_DECFG,
 	MSR_IA32_UCODE_REV,
+	MSR_IA32_CORE_CAPABILITY,

arch/x86/kvm/x86.c static u32 msr_based_features[] = { · line 1211

Quoted 10 lines
arch/x86/kvm/x86.c +5 -0
 	MSR_IA32_ARCH_CAPABILITIES,
 };
 
 static unsigned int num_msr_based_features;
 
+u64 kvm_get_core_capability(void)
+{
+	u64 data;
+
+	rdmsrl_safe(MSR_IA32_CORE_CAPABILITY, &data);

This patch should be merged with the previous patch. Also here you should add:

	data &= CORE_CAP_SPLIT_LOCK_DETECT;

so that non-virtualizable features are hidden and

	if (boot_cpu_has(X86_FEATURE_SPLIT_LOCK_DETECT))
		data |= CORE_CAP_SPLIT_LOCK_DETECT;
Quoted 6 lines

so that userspace gets "for free" the FMS list that will be added later to the kernel.

Thanks,

Paolo

You are right. I'll correct it in next version.

Quoted 21 lines
Quoted 19 lines
arch/x86/kvm/x86.c +8 -0
+
+	return data;
+}
+EXPORT_SYMBOL_GPL(kvm_get_core_capability);
+
 u64 kvm_get_arch_capabilities(void)
 {
 	u64 data;
@@ -1227,6 +1239,9 @@ EXPORT_SYMBOL_GPL(kvm_get_arch_capabilities);
 static int kvm_get_msr_feature(struct kvm_msr_entry *msr)
 {
 	switch (msr->index) {
+	case MSR_IA32_CORE_CAPABILITY:
+		msr->data = kvm_get_core_capability();
+		break;
 	case MSR_IA32_ARCH_CAPABILITIES:
 		msr->data = kvm_get_arch_capabilities();
 		break;
#54 lore raw

Re: [PATCH v4 16/17] kvm: x86: Add support IA32_CORE_CAPABILITY MSR

In reply to Paolo Bonzini

Hi, Paolo

Quoted 140 lines · Paolo Bonzini
Quoted 125 lines · Fenghua Yu

From: Xiaoyao Li <[email protected]>

MSR IA32_CORE_CAPABILITY is a feature-enumerating MSR, bit 5 of which reports the capability of enabling detection of split locks (will be supported on future processors based on Tremont microarchitecture and later).

Please check the latest Intel Architecture Instruction Set Extensions and Future Features Programming Reference for more detailed information on the MSR and the split lock bit.

1. Expose it to user space as a feature-enumerating MSR, so that user space can query it.

2. Emualte MSR_IA32_CORE_CAPABILITY with vmx->core_capability. And add the get and set handler of MSR_IA32_CORE_CAPABILITY.
For uesrspace, it can set this MSR when customizing features of guest,
also it can read the value of this MSR of guest.
For guest, as it's a feature-enumerating MSR, guest only can read it.

---
 arch/x86/include/asm/kvm_host.h |  1 +
 arch/x86/kvm/vmx/vmx.c          | 23 +++++++++++++++++++++++
 arch/x86/kvm/vmx/vmx.h          |  1 +
 arch/x86/kvm/x86.c              | 17 ++++++++++++++++-
 4 files changed, 41 insertions(+), 1 deletion(-)
arch/x86/include/asm/kvm_host.h +1 -0
index 180373360e34..208f15570d17 100644
@@ -1527,6 +1527,7 @@ int kvm_pv_send_ipi(struct kvm *kvm, unsigned long
 ipi_bitmap_low,
 		    unsigned long icr, int op_64_bit);
 
 u64 kvm_get_arch_capabilities(void);
+u64 kvm_get_core_capability(void);
 void kvm_define_shared_msr(unsigned index, u32 msr);
 int kvm_set_shared_msr(unsigned index, u64 val, u64 mask);
arch/x86/kvm/vmx/vmx.c +6 -0
index 30a6bcd735ec..3e03c6e1e558 100644
@@ -1679,6 +1679,12 @@ static int vmx_get_msr(struct kvm_vcpu *vcpu, struct
 msr_data *msr_info)
 
 		msr_info->data = to_vmx(vcpu)->spec_ctrl;
 		break;
+	case MSR_IA32_CORE_CAPABILITY:
+		if (!msr_info->host_initiated &&
+		    !guest_cpuid_has(vcpu, X86_FEATURE_CORE_CAPABILITY))
+			return 1;
+		msr_info->data = vmx->core_capability;
+		break;
 	case MSR_IA32_ARCH_CAPABILITIES:
 		if (!msr_info->host_initiated &&
arch/x86/kvm/vmx/vmx.c +15 -0
 		    !guest_cpuid_has(vcpu, X86_FEATURE_ARCH_CAPABILITIES))
@@ -1891,6 +1897,21 @@ static int vmx_set_msr(struct kvm_vcpu *vcpu, struct
 msr_data *msr_info)
 		vmx_disable_intercept_for_msr(vmx->vmcs01.msr_bitmap,
 MSR_IA32_PRED_CMD,
 					      MSR_TYPE_W);
 		break;
+	case MSR_IA32_CORE_CAPABILITY:
+		if (!msr_info->host_initiated)
+			return 1;
+		if (data & ~CORE_CAP_SPLIT_LOCK_DETECT)
+			return 1;
+
+		/*
+		 * Since AC split lock is a hardware feature, and there is no
+		 * software emulation yet, we cannot enable it for guest if
+		 * host hardware doesn't support it.
+		 */
+		if (!boot_cpu_has(X86_FEATURE_SPLIT_LOCK_DETECT))
+			data &= ~CORE_CAP_SPLIT_LOCK_DETECT;
+		vmx->core_capability = data;
+		break;
 	case MSR_IA32_ARCH_CAPABILITIES:
arch/x86/kvm/vmx/vmx.c +2 -0
 		if (!msr_info->host_initiated)
 			return 1;
@@ -4083,6 +4104,8 @@ static void vmx_vcpu_setup(struct vcpu_vmx *vmx)
 		++vmx->nmsrs;
 	}
 
+	vmx->core_capability = kvm_get_core_capability();
+
 	vmx->arch_capabilities = kvm_get_arch_capabilities();
 
 	vm_exit_controls_init(vmx, vmx_vmexit_ctrl());
arch/x86/kvm/vmx/vmx.h +1 -0
index 0ac0a64c7790..cc22379991f3 100644
@@ -191,6 +191,7 @@ struct vcpu_vmx {
 	u64		      msr_guest_kernel_gs_base;
 #endif
 
+	u64		      core_capability;
 	u64		      arch_capabilities;
 	u64		      spec_ctrl;
 
arch/x86/kvm/x86.c +3 -1
index 941f932373d0..c3c9e3f2d08a 100644
@@ -1125,7 +1125,8 @@ static u32 msrs_to_save[] = {
 #endif
 	MSR_IA32_TSC, MSR_IA32_CR_PAT, MSR_VM_HSAVE_PA,
 	MSR_IA32_FEATURE_CONTROL, MSR_IA32_BNDCFGS, MSR_TSC_AUX,
-	MSR_IA32_SPEC_CTRL, MSR_IA32_ARCH_CAPABILITIES,
+	MSR_IA32_SPEC_CTRL, MSR_IA32_CORE_CAPABILITY,
+	MSR_IA32_ARCH_CAPABILITIES,
 	MSR_IA32_RTIT_CTL, MSR_IA32_RTIT_STATUS, MSR_IA32_RTIT_CR3_MATCH,
 	MSR_IA32_RTIT_OUTPUT_BASE, MSR_IA32_RTIT_OUTPUT_MASK,
 	MSR_IA32_RTIT_ADDR0_A, MSR_IA32_RTIT_ADDR0_B,
@@ -1197,11 +1198,22 @@ static u32 msr_based_features[] = {
 
 	MSR_F10H_DECFG,
 	MSR_IA32_UCODE_REV,
+	MSR_IA32_CORE_CAPABILITY,

arch/x86/kvm/x86.c static u32 msr_based_features[] = { · line 1211

Quoted 10 lines
arch/x86/kvm/x86.c +5 -0
 	MSR_IA32_ARCH_CAPABILITIES,
 };
 
 static unsigned int num_msr_based_features;
 
+u64 kvm_get_core_capability(void)
+{
+	u64 data;
+
+	rdmsrl_safe(MSR_IA32_CORE_CAPABILITY, &data);

This patch should be merged with the previous patch. Also here you should add:

Quoted 1 line
	data &= CORE_CAP_SPLIT_LOCK_DETECT;

I agree with this.

Quoted 7 lines

so that non-virtualizable features are hidden and

	if (boot_cpu_has(X86_FEATURE_SPLIT_LOCK_DETECT))
		data |= CORE_CAP_SPLIT_LOCK_DETECT;

so that userspace gets "for free" the FMS list that will be added later to the kernel.

I think it's redundant. Because there is no case that rdmsrl_safe(MSR_IA32_CORE_CAPABILITY, &data) shows no split lock detection while boot_cpu_has(X86_FEATURE_SPLIT_LOCK_DETECT).

Xiaoyao

Quoted 25 lines

Thanks,

Paolo

Quoted 19 lines
arch/x86/kvm/x86.c +8 -0
+
+	return data;
+}
+EXPORT_SYMBOL_GPL(kvm_get_core_capability);
+
 u64 kvm_get_arch_capabilities(void)
 {
 	u64 data;
@@ -1227,6 +1239,9 @@ EXPORT_SYMBOL_GPL(kvm_get_arch_capabilities);
 static int kvm_get_msr_feature(struct kvm_msr_entry *msr)
 {
 	switch (msr->index) {
+	case MSR_IA32_CORE_CAPABILITY:
+		msr->data = kvm_get_core_capability();
+		break;
 	case MSR_IA32_ARCH_CAPABILITIES:
 		msr->data = kvm_get_arch_capabilities();
 		break;
maintainer
#55 lore raw

Re: [PATCH v4 16/17] kvm: x86: Add support IA32_CORE_CAPABILITY MSR

Quoted 10 lines · Xiaoyao Li
Quoted 7 lines

so that non-virtualizable features are hidden and

	if (boot_cpu_has(X86_FEATURE_SPLIT_LOCK_DETECT))
		data |= CORE_CAP_SPLIT_LOCK_DETECT;

so that userspace gets "for free" the FMS list that will be added later to the kernel.

I think it's redundant. Because there is no case that rdmsrl_safe(MSR_IA32_CORE_CAPABILITY, &data) shows no split lock detection while boot_cpu_has(X86_FEATURE_SPLIT_LOCK_DETECT).

There will be when the kernel will add a list of FMS values that have split lock detection but lack the core capabilities MSR. Or at least that is what Fenghua said in the cover letter.

Paolo

#56 lore raw

Re: [PATCH v4 16/17] kvm: x86: Add support IA32_CORE_CAPABILITY MSR

Quoted 14 lines · Paolo Bonzini
Quoted 5 lines · Xiaoyao Li
Quoted 5 lines

so that non-virtualizable features are hidden and

	if (boot_cpu_has(X86_FEATURE_SPLIT_LOCK_DETECT))
		data |= CORE_CAP_SPLIT_LOCK_DETECT;
Quoted 7 lines
Quoted 2 lines

so that userspace gets "for free" the FMS list that will be added later to the kernel.

I think it's redundant. Because there is no case that rdmsrl_safe(MSR_IA32_CORE_CAPABILITY, &data) shows no split lock detection while
boot_cpu_has(X86_FEATURE_SPLIT_LOCK_DETECT).

Quoted 5 lines

There will be when the kernel will add a list of FMS values that have split lock detection but lack the core capabilities MSR. Or at least that is what Fenghua said in the cover letter.

Paolo

Got it.
Thanks for your explanation.

author
#57 lore raw

PATCH v4 17/17 kvm: vmx: Emulate TEST_CTL MSR

In reply to Fenghua Yu

From: Xiaoyao Li <[email protected]>

A control bit (bit 29) in TEST_CTL MSR 0x33 will be introduced in future x86 processors. When bit 29 is set, the processor causes #AC exception for split locked accesses at all CPL.

Please check the latest Intel Software Developer's Manual for more detailed information on the MSR and the split lock bit.

1. Since the kernel chooses to enable AC split lock by default, which means if we don't emulate TEST_CTL MSR for guest, guest will run with this feature enable while does not known it. Thus existing guests with buggy firmware (like OVMF) and old kernels having the cross cache line issues will fail the boot due to #AC.

So we should emulate TEST_CTL MSR, and set it zero to disable AC split lock by default. Whether and when to enable it is left to guest firmware and guest kernel.

2. Host and guest can enable AC split lock independently, so using msr autoload to switch it during VM entry/exit.

---
 arch/x86/kvm/vmx/vmx.c | 35 +++++++++++++++++++++++++++++++++++
 arch/x86/kvm/vmx/vmx.h |  1 +
 2 files changed, 36 insertions(+)
arch/x86/kvm/vmx/vmx.c +35 -0
index 3e03c6e1e558..c0c5e8621afa 100644
@@ -1659,6 +1659,12 @@ static int vmx_get_msr(struct kvm_vcpu *vcpu, struct msr_data *msr_info)
 	u32 index;
 
 	switch (msr_info->index) {
+	case MSR_TEST_CTL:
+		if (!msr_info->host_initiated &&
+		    !(vmx->core_capability & CORE_CAP_SPLIT_LOCK_DETECT))
+			return 1;
+		msr_info->data = vmx->msr_test_ctl;
+		break;
 #ifdef CONFIG_X86_64
 	case MSR_FS_BASE:
 		msr_info->data = vmcs_readl(GUEST_FS_BASE);
@@ -1805,6 +1811,14 @@ static int vmx_set_msr(struct kvm_vcpu *vcpu, struct msr_data *msr_info)
 	u32 index;
 
 	switch (msr_index) {
+	case MSR_TEST_CTL:
+		if (!(vmx->core_capability & CORE_CAP_SPLIT_LOCK_DETECT))
+			return 1;
+
+		if (data & ~TEST_CTL_ENABLE_SPLIT_LOCK_DETECT)
+			return 1;
+		vmx->msr_test_ctl = data;
+		break;
 	case MSR_EFER:
 		ret = kvm_set_msr_common(vcpu, msr_info);
 		break;
@@ -4108,6 +4122,9 @@ static void vmx_vcpu_setup(struct vcpu_vmx *vmx)
 
 	vmx->arch_capabilities = kvm_get_arch_capabilities();
 
+	/* disable AC split lock by default */
+	vmx->msr_test_ctl = 0;
+
 	vm_exit_controls_init(vmx, vmx_vmexit_ctrl());
 
 	/* 22.2.1, 20.8.1 */
@@ -4145,6 +4162,7 @@ static void vmx_vcpu_reset(struct kvm_vcpu *vcpu, bool init_event)
 
 	vmx->rmode.vm86_active = 0;
 	vmx->spec_ctrl = 0;
+	vmx->msr_test_ctl = 0;
 
 	vcpu->arch.microcode_version = 0x100000000ULL;
 	vmx->vcpu.arch.regs[VCPU_REGS_RDX] = get_rdx_init_val();
@@ -6344,6 +6362,21 @@ static void atomic_switch_perf_msrs(struct vcpu_vmx *vmx)
 					msrs[i].host, false);
 }
 
+static void atomic_switch_msr_test_ctl(struct vcpu_vmx *vmx)
+{
+	u64 host_msr_test_ctl;
+
+	if (!boot_cpu_has(X86_FEATURE_SPLIT_LOCK_DETECT))
+		return;
+
+	rdmsrl(MSR_TEST_CTL, host_msr_test_ctl);
+	if (host_msr_test_ctl == vmx->msr_test_ctl)
+		clear_atomic_switch_msr(vmx, MSR_TEST_CTL);
+	else
+		add_atomic_switch_msr(vmx, MSR_TEST_CTL, vmx->msr_test_ctl,
+				      host_msr_test_ctl, false);
+}
+
 static void vmx_arm_hv_timer(struct vcpu_vmx *vmx, u32 val)
 {
 	vmcs_write32(VMX_PREEMPTION_TIMER_VALUE, val);
@@ -6585,6 +6618,8 @@ static void vmx_vcpu_run(struct kvm_vcpu *vcpu)
 
 	atomic_switch_perf_msrs(vmx);
 
+	atomic_switch_msr_test_ctl(vmx);
+
 	vmx_update_hv_timer(vcpu);
 
 	/*
arch/x86/kvm/vmx/vmx.h +1 -0
index cc22379991f3..e8831609c6c3 100644
@@ -191,6 +191,7 @@ struct vcpu_vmx {
 	u64		      msr_guest_kernel_gs_base;
 #endif
 
+	u64		      msr_test_ctl;
 	u64		      core_capability;
 	u64		      arch_capabilities;
 	u64		      spec_ctrl;
Signature · 1 line
2.7.4
#58 lore raw

Re: [PATCH v4 17/17] kvm: vmx: Emulate TEST_CTL MSR

Hi, Paolo,

Do you have any comments on this patch?

We are preparing v5 patches for split lock detection, if you have any comments about this one, please let me know.

Thanks,
Xiaoyao

Quoted 126 lines · Fenghua Yu

From: Xiaoyao Li <[email protected]>

A control bit (bit 29) in TEST_CTL MSR 0x33 will be introduced in future x86 processors. When bit 29 is set, the processor causes #AC exception for split locked accesses at all CPL.

Please check the latest Intel Software Developer's Manual for more detailed information on the MSR and the split lock bit.

1. Since the kernel chooses to enable AC split lock by default, which means if we don't emulate TEST_CTL MSR for guest, guest will run with this feature enable while does not known it. Thus existing guests with buggy firmware (like OVMF) and old kernels having the cross cache line issues will fail the boot due to #AC.

So we should emulate TEST_CTL MSR, and set it zero to disable AC split lock by default. Whether and when to enable it is left to guest firmware and guest kernel.

2. Host and guest can enable AC split lock independently, so using msr autoload to switch it during VM entry/exit.

---
 arch/x86/kvm/vmx/vmx.c | 35 +++++++++++++++++++++++++++++++++++
 arch/x86/kvm/vmx/vmx.h |  1 +
 2 files changed, 36 insertions(+)
arch/x86/kvm/vmx/vmx.c +6 -0
index 3e03c6e1e558..c0c5e8621afa 100644
@@ -1659,6 +1659,12 @@ static int vmx_get_msr(struct kvm_vcpu *vcpu, struct
 msr_data *msr_info)
 	u32 index;
 
 	switch (msr_info->index) {
+	case MSR_TEST_CTL:
+		if (!msr_info->host_initiated &&
+		    !(vmx->core_capability & CORE_CAP_SPLIT_LOCK_DETECT))
+			return 1;
+		msr_info->data = vmx->msr_test_ctl;
+		break;
 #ifdef CONFIG_X86_64
 	case MSR_FS_BASE:
arch/x86/kvm/vmx/vmx.c +8 -0
 		msr_info->data = vmcs_readl(GUEST_FS_BASE);
@@ -1805,6 +1811,14 @@ static int vmx_set_msr(struct kvm_vcpu *vcpu, struct
 msr_data *msr_info)
 	u32 index;
 
 	switch (msr_index) {
+	case MSR_TEST_CTL:
+		if (!(vmx->core_capability & CORE_CAP_SPLIT_LOCK_DETECT))
+			return 1;
+
+		if (data & ~TEST_CTL_ENABLE_SPLIT_LOCK_DETECT)
+			return 1;
+		vmx->msr_test_ctl = data;
+		break;
 	case MSR_EFER:
 		ret = kvm_set_msr_common(vcpu, msr_info);
arch/x86/kvm/vmx/vmx.c +4 -0
 		break;
@@ -4108,6 +4122,9 @@ static void vmx_vcpu_setup(struct vcpu_vmx *vmx)
 
 	vmx->arch_capabilities = kvm_get_arch_capabilities();
 
+	/* disable AC split lock by default */
+	vmx->msr_test_ctl = 0;
+
 	vm_exit_controls_init(vmx, vmx_vmexit_ctrl());
 
 	/* 22.2.1, 20.8.1 */
@@ -4145,6 +4162,7 @@ static void vmx_vcpu_reset(struct kvm_vcpu *vcpu, bool
 init_event)
 
 	vmx->rmode.vm86_active = 0;
 	vmx->spec_ctrl = 0;
+	vmx->msr_test_ctl = 0;
 
 	vcpu->arch.microcode_version = 0x100000000ULL;
arch/x86/kvm/vmx/vmx.c +15 -0
 	vmx->vcpu.arch.regs[VCPU_REGS_RDX] = get_rdx_init_val();
@@ -6344,6 +6362,21 @@ static void atomic_switch_perf_msrs(struct vcpu_vmx
 *vmx)
 					msrs[i].host, false);
 }
 
+static void atomic_switch_msr_test_ctl(struct vcpu_vmx *vmx)
+{
+	u64 host_msr_test_ctl;
+
+	if (!boot_cpu_has(X86_FEATURE_SPLIT_LOCK_DETECT))
+		return;
+
+	rdmsrl(MSR_TEST_CTL, host_msr_test_ctl);
+	if (host_msr_test_ctl == vmx->msr_test_ctl)
+		clear_atomic_switch_msr(vmx, MSR_TEST_CTL);
+	else
+		add_atomic_switch_msr(vmx, MSR_TEST_CTL, vmx->msr_test_ctl,
+				      host_msr_test_ctl, false);
+}
+
 static void vmx_arm_hv_timer(struct vcpu_vmx *vmx, u32 val)
 {
arch/x86/kvm/vmx/vmx.c +2 -0
 	vmcs_write32(VMX_PREEMPTION_TIMER_VALUE, val);
@@ -6585,6 +6618,8 @@ static void vmx_vcpu_run(struct kvm_vcpu *vcpu)
 
 	atomic_switch_perf_msrs(vmx);
 
+	atomic_switch_msr_test_ctl(vmx);
+
 	vmx_update_hv_timer(vcpu);
 
 	/*
arch/x86/kvm/vmx/vmx.h +1 -0
index cc22379991f3..e8831609c6c3 100644
@@ -191,6 +191,7 @@ struct vcpu_vmx {
 	u64		      msr_guest_kernel_gs_base;
 #endif
 
+	u64		      msr_test_ctl;
 	u64		      core_capability;
 	u64		      arch_capabilities;
 	u64		      spec_ctrl;
maintainer
#59 lore raw

Re: [PATCH v4 17/17] kvm: vmx: Emulate TEST_CTL MSR

Quoted 6 lines · Xiaoyao Li

Hi, Paolo,

Do you have any comments on this patch?

We are preparing v5 patches for split lock detection, if you have any comments about this one, please let me know.

No, my only comment is that it should be placed _before_ the other two for bisectability. I think I have already sent that small remark.

Thanks,

Paolo

Quoted 131 lines

Thanks,
Xiaoyao

Quoted 126 lines · Fenghua Yu

From: Xiaoyao Li <[email protected]>

A control bit (bit 29) in TEST_CTL MSR 0x33 will be introduced in future x86 processors. When bit 29 is set, the processor causes #AC exception for split locked accesses at all CPL.

Please check the latest Intel Software Developer's Manual for more detailed information on the MSR and the split lock bit.

1. Since the kernel chooses to enable AC split lock by default, which means if we don't emulate TEST_CTL MSR for guest, guest will run with this feature enable while does not known it. Thus existing guests with buggy firmware (like OVMF) and old kernels having the cross cache line issues will fail the boot due to #AC.

So we should emulate TEST_CTL MSR, and set it zero to disable AC split lock by default. Whether and when to enable it is left to guest firmware and guest kernel.

2. Host and guest can enable AC split lock independently, so using msr autoload to switch it during VM entry/exit.

---
 arch/x86/kvm/vmx/vmx.c | 35 +++++++++++++++++++++++++++++++++++
 arch/x86/kvm/vmx/vmx.h |  1 +
 2 files changed, 36 insertions(+)
arch/x86/kvm/vmx/vmx.c +6 -0
index 3e03c6e1e558..c0c5e8621afa 100644
@@ -1659,6 +1659,12 @@ static int vmx_get_msr(struct kvm_vcpu *vcpu, struct
 msr_data *msr_info)
 	u32 index;
 
 	switch (msr_info->index) {
+	case MSR_TEST_CTL:
+		if (!msr_info->host_initiated &&
+		    !(vmx->core_capability & CORE_CAP_SPLIT_LOCK_DETECT))
+			return 1;
+		msr_info->data = vmx->msr_test_ctl;
+		break;
 #ifdef CONFIG_X86_64
 	case MSR_FS_BASE:
arch/x86/kvm/vmx/vmx.c +8 -0
 		msr_info->data = vmcs_readl(GUEST_FS_BASE);
@@ -1805,6 +1811,14 @@ static int vmx_set_msr(struct kvm_vcpu *vcpu, struct
 msr_data *msr_info)
 	u32 index;
 
 	switch (msr_index) {
+	case MSR_TEST_CTL:
+		if (!(vmx->core_capability & CORE_CAP_SPLIT_LOCK_DETECT))
+			return 1;
+
+		if (data & ~TEST_CTL_ENABLE_SPLIT_LOCK_DETECT)
+			return 1;
+		vmx->msr_test_ctl = data;
+		break;
 	case MSR_EFER:
 		ret = kvm_set_msr_common(vcpu, msr_info);
arch/x86/kvm/vmx/vmx.c +4 -0
 		break;
@@ -4108,6 +4122,9 @@ static void vmx_vcpu_setup(struct vcpu_vmx *vmx)
 
 	vmx->arch_capabilities = kvm_get_arch_capabilities();
 
+	/* disable AC split lock by default */
+	vmx->msr_test_ctl = 0;
+
 	vm_exit_controls_init(vmx, vmx_vmexit_ctrl());
 
 	/* 22.2.1, 20.8.1 */
@@ -4145,6 +4162,7 @@ static void vmx_vcpu_reset(struct kvm_vcpu *vcpu, bool
 init_event)
 
 	vmx->rmode.vm86_active = 0;
 	vmx->spec_ctrl = 0;
+	vmx->msr_test_ctl = 0;
 
 	vcpu->arch.microcode_version = 0x100000000ULL;
arch/x86/kvm/vmx/vmx.c +15 -0
 	vmx->vcpu.arch.regs[VCPU_REGS_RDX] = get_rdx_init_val();
@@ -6344,6 +6362,21 @@ static void atomic_switch_perf_msrs(struct vcpu_vmx
 *vmx)
 					msrs[i].host, false);
 }
 
+static void atomic_switch_msr_test_ctl(struct vcpu_vmx *vmx)
+{
+	u64 host_msr_test_ctl;
+
+	if (!boot_cpu_has(X86_FEATURE_SPLIT_LOCK_DETECT))
+		return;
+
+	rdmsrl(MSR_TEST_CTL, host_msr_test_ctl);
+	if (host_msr_test_ctl == vmx->msr_test_ctl)
+		clear_atomic_switch_msr(vmx, MSR_TEST_CTL);
+	else
+		add_atomic_switch_msr(vmx, MSR_TEST_CTL, vmx->msr_test_ctl,
+				      host_msr_test_ctl, false);
+}
+
 static void vmx_arm_hv_timer(struct vcpu_vmx *vmx, u32 val)
 {
arch/x86/kvm/vmx/vmx.c +2 -0
 	vmcs_write32(VMX_PREEMPTION_TIMER_VALUE, val);
@@ -6585,6 +6618,8 @@ static void vmx_vcpu_run(struct kvm_vcpu *vcpu)
 
 	atomic_switch_perf_msrs(vmx);
 
+	atomic_switch_msr_test_ctl(vmx);
+
 	vmx_update_hv_timer(vcpu);
 
 	/*
arch/x86/kvm/vmx/vmx.h +1 -0
index cc22379991f3..e8831609c6c3 100644
@@ -191,6 +191,7 @@ struct vcpu_vmx {
 	u64		      msr_guest_kernel_gs_base;
 #endif
 
+	u64		      msr_test_ctl;
 	u64		      core_capability;
 	u64		      arch_capabilities;
 	u64		      spec_ctrl;
#60 lore raw

Re: [PATCH v4 17/17] kvm: vmx: Emulate TEST_CTL MSR

Quoted 9 lines · Paolo Bonzini
Quoted 7 lines · Xiaoyao Li

Hi, Paolo,

Do you have any comments on this patch?

We are preparing v5 patches for split lock detection, if you have any comments
about this one, please let me know.

Quoted 6 lines

No, my only comment is that it should be placed _before_ the other two for bisectability. I think I have already sent that small remark.

Thanks,

Paolo

I cannot find the small remark you sent before. Maybe I missed something. But I'am confused why it should be placed _before_ the other two. This patch just use the vmx->core_capability that defined it the previous patch.

Quoted 135 lines
Quoted 133 lines

Thanks,
Xiaoyao

Quoted 129 lines · Fenghua Yu

From: Xiaoyao Li <[email protected]>

A control bit (bit 29) in TEST_CTL MSR 0x33 will be introduced in future x86 processors. When bit 29 is set, the processor causes #AC exception for split locked accesses at all CPL.

Please check the latest Intel Software Developer's Manual for more detailed information on the MSR and the split lock bit.

1. Since the kernel chooses to enable AC split lock by default, which means if we don't emulate TEST_CTL MSR for guest, guest will run with this feature enable while does not known it. Thus existing guests with buggy firmware (like OVMF) and old kernels having the cross cache line issues will fail the boot due to #AC.

So we should emulate TEST_CTL MSR, and set it zero to disable AC split lock by default. Whether and when to enable it is left to guest firmware and guest kernel.

2. Host and guest can enable AC split lock independently, so using msr autoload to switch it during VM entry/exit.

---
 arch/x86/kvm/vmx/vmx.c | 35 +++++++++++++++++++++++++++++++++++
 arch/x86/kvm/vmx/vmx.h |  1 +
 2 files changed, 36 insertions(+)
arch/x86/kvm/vmx/vmx.c +0 -0
index 3e03c6e1e558..c0c5e8621afa 100644
@@ -1659,6 +1659,12 @@ static int vmx_get_msr(struct kvm_vcpu *vcpu,

struct
msr_data *msr_info)

arch/x86/kvm/vmx/vmx.c +6 -0
 	u32 index;
 
 	switch (msr_info->index) {
+	case MSR_TEST_CTL:
+		if (!msr_info->host_initiated &&
+		    !(vmx->core_capability & CORE_CAP_SPLIT_LOCK_DETECT))
+			return 1;
+		msr_info->data = vmx->msr_test_ctl;
+		break;
 #ifdef CONFIG_X86_64
 	case MSR_FS_BASE:
 		msr_info->data = vmcs_readl(GUEST_FS_BASE);
@@ -1805,6 +1811,14 @@ static int vmx_set_msr(struct kvm_vcpu *vcpu,

struct
msr_data *msr_info)

arch/x86/kvm/vmx/vmx.c +11 -0
 	u32 index;
 
 	switch (msr_index) {
+	case MSR_TEST_CTL:
+		if (!(vmx->core_capability & CORE_CAP_SPLIT_LOCK_DETECT))
+			return 1;
+
+		if (data & ~TEST_CTL_ENABLE_SPLIT_LOCK_DETECT)
+			return 1;
+		vmx->msr_test_ctl = data;
+		break;
 	case MSR_EFER:
 		ret = kvm_set_msr_common(vcpu, msr_info);
 		break;
@@ -4108,6 +4122,9 @@ static void vmx_vcpu_setup(struct vcpu_vmx *vmx)
 
 	vmx->arch_capabilities = kvm_get_arch_capabilities();
 
+	/* disable AC split lock by default */
+	vmx->msr_test_ctl = 0;
+
 	vm_exit_controls_init(vmx, vmx_vmexit_ctrl());
 
 	/* 22.2.1, 20.8.1 */
@@ -4145,6 +4162,7 @@ static void vmx_vcpu_reset(struct kvm_vcpu *vcpu,

bool
init_event)

arch/x86/kvm/vmx/vmx.c +16 -0
 	vmx->rmode.vm86_active = 0;
 	vmx->spec_ctrl = 0;
+	vmx->msr_test_ctl = 0;
 
 	vcpu->arch.microcode_version = 0x100000000ULL;
 	vmx->vcpu.arch.regs[VCPU_REGS_RDX] = get_rdx_init_val();
@@ -6344,6 +6362,21 @@ static void atomic_switch_perf_msrs(struct vcpu_vmx
 *vmx)
 					msrs[i].host, false);
 }
 
+static void atomic_switch_msr_test_ctl(struct vcpu_vmx *vmx)
+{
+	u64 host_msr_test_ctl;
+
+	if (!boot_cpu_has(X86_FEATURE_SPLIT_LOCK_DETECT))
+		return;
+
+	rdmsrl(MSR_TEST_CTL, host_msr_test_ctl);
+	if (host_msr_test_ctl == vmx->msr_test_ctl)
+		clear_atomic_switch_msr(vmx, MSR_TEST_CTL);
+	else
+		add_atomic_switch_msr(vmx, MSR_TEST_CTL, vmx->msr_test_ctl,
+				      host_msr_test_ctl, false);
+}
+
 static void vmx_arm_hv_timer(struct vcpu_vmx *vmx, u32 val)
 {
arch/x86/kvm/vmx/vmx.c +2 -0
 	vmcs_write32(VMX_PREEMPTION_TIMER_VALUE, val);
@@ -6585,6 +6618,8 @@ static void vmx_vcpu_run(struct kvm_vcpu *vcpu)
 
 	atomic_switch_perf_msrs(vmx);
 
+	atomic_switch_msr_test_ctl(vmx);
+
 	vmx_update_hv_timer(vcpu);
 
 	/*
arch/x86/kvm/vmx/vmx.h +1 -0
index cc22379991f3..e8831609c6c3 100644
@@ -191,6 +191,7 @@ struct vcpu_vmx {
 	u64		      msr_guest_kernel_gs_base;
 #endif
 
+	u64		      msr_test_ctl;
 	u64		      core_capability;
 	u64		      arch_capabilities;
 	u64		      spec_ctrl;
maintainer
#61 lore raw

Re: [PATCH v4 17/17] kvm: vmx: Emulate TEST_CTL MSR

Quoted 13 lines · Xiaoyao Li
Quoted 12 lines · Paolo Bonzini
Quoted 7 lines · Xiaoyao Li

Hi, Paolo,

Do you have any comments on this patch?

We are preparing v5 patches for split lock detection, if you have any comments
about this one, please let me know.

No, my only comment is that it should be placed _before_ the other two for bisectability. I think I have already sent that small remark.

Quoted 7 lines
Quoted 3 lines

Thanks,

Paolo

I cannot find the small remark you sent before. Maybe I missed something. But I'am confused why it should be placed _before_ the other two. This patch just use the vmx->core_capability that defined it the previous patch.

Because otherwise the guest can see core_capability != 0 and will #GP when trying to use split lock detection.

But you are right, this patch must be the last. Instead, kvm-get_core_capability() should always return 0 until the previous patch. Then in this patch you add the rdmsr and boot_cpu_has() tests.

Paolo

Quoted 77 lines
Quoted 76 lines
Quoted 74 lines
Quoted 74 lines
arch/x86/kvm/vmx/vmx.c +10 -0
+		if (!(vmx->core_capability & CORE_CAP_SPLIT_LOCK_DETECT))
+			return 1;
+
+		if (data & ~TEST_CTL_ENABLE_SPLIT_LOCK_DETECT)
+			return 1;
+		vmx->msr_test_ctl = data;
+		break;
 	case MSR_EFER:
 		ret = kvm_set_msr_common(vcpu, msr_info);
 		break;
@@ -4108,6 +4122,9 @@ static void vmx_vcpu_setup(struct vcpu_vmx *vmx)
 
 	vmx->arch_capabilities = kvm_get_arch_capabilities();
 
+	/* disable AC split lock by default */
+	vmx->msr_test_ctl = 0;
+
 	vm_exit_controls_init(vmx, vmx_vmexit_ctrl());
 
 	/* 22.2.1, 20.8.1 */
@@ -4145,6 +4162,7 @@ static void vmx_vcpu_reset(struct kvm_vcpu *vcpu,

bool
init_event)

arch/x86/kvm/vmx/vmx.c +16 -0
 	vmx->rmode.vm86_active = 0;
 	vmx->spec_ctrl = 0;
+	vmx->msr_test_ctl = 0;
 
 	vcpu->arch.microcode_version = 0x100000000ULL;
 	vmx->vcpu.arch.regs[VCPU_REGS_RDX] = get_rdx_init_val();
@@ -6344,6 +6362,21 @@ static void atomic_switch_perf_msrs(struct vcpu_vmx
 *vmx)
 					msrs[i].host, false);
 }
 
+static void atomic_switch_msr_test_ctl(struct vcpu_vmx *vmx)
+{
+	u64 host_msr_test_ctl;
+
+	if (!boot_cpu_has(X86_FEATURE_SPLIT_LOCK_DETECT))
+		return;
+
+	rdmsrl(MSR_TEST_CTL, host_msr_test_ctl);
+	if (host_msr_test_ctl == vmx->msr_test_ctl)
+		clear_atomic_switch_msr(vmx, MSR_TEST_CTL);
+	else
+		add_atomic_switch_msr(vmx, MSR_TEST_CTL, vmx->msr_test_ctl,
+				      host_msr_test_ctl, false);
+}
+
 static void vmx_arm_hv_timer(struct vcpu_vmx *vmx, u32 val)
 {
arch/x86/kvm/vmx/vmx.c +2 -0
 	vmcs_write32(VMX_PREEMPTION_TIMER_VALUE, val);
@@ -6585,6 +6618,8 @@ static void vmx_vcpu_run(struct kvm_vcpu *vcpu)
 
 	atomic_switch_perf_msrs(vmx);
 
+	atomic_switch_msr_test_ctl(vmx);
+
 	vmx_update_hv_timer(vcpu);
 
 	/*
arch/x86/kvm/vmx/vmx.h +1 -0
index cc22379991f3..e8831609c6c3 100644
@@ -191,6 +191,7 @@ struct vcpu_vmx {
 	u64		      msr_guest_kernel_gs_base;
 #endif
 
+	u64		      msr_test_ctl;
 	u64		      core_capability;
 	u64		      arch_capabilities;
 	u64		      spec_ctrl;
#62 lore raw

Re: [PATCH v4 17/17] kvm: vmx: Emulate TEST_CTL MSR

Quoted 22 lines · Paolo Bonzini
Quoted 13 lines · Xiaoyao Li
Quoted 12 lines · Paolo Bonzini
Quoted 7 lines · Xiaoyao Li

Hi, Paolo,

Do you have any comments on this patch?

We are preparing v5 patches for split lock detection, if you have any comments
about this one, please let me know.

No, my only comment is that it should be placed _before_ the other two for bisectability. I think I have already sent that small remark.

Quoted 7 lines
Quoted 3 lines

Thanks,

Paolo

I cannot find the small remark you sent before. Maybe I missed something. But I'am confused why it should be placed _before_ the other two. This patch just use the vmx->core_capability that defined it the previous patch.

Quoted 8 lines

Because otherwise the guest can see core_capability != 0 and will #GP when trying to use split lock detection.

But you are right, this patch must be the last. Instead, kvm-get_core_capability() should always return 0 until the previous patch. Then in this patch you add the rdmsr and boot_cpu_has() tests.

Paolo

Hi, Paolo

Thanks a lot for your explanation. It makes me better understand the bisectability of a patchset. I really appreciate you and I love the community. I can learn a lot here and it makes me better.

I will fix the bisectability issue following your comments.

Quoted 82 lines
Quoted 80 lines
Quoted 80 lines
Quoted 78 lines
Quoted 78 lines
(quoted diff) +1 -0
+		if (!(vmx->core_capability &

CORE_CAP_SPLIT_LOCK_DETECT))

arch/x86/kvm/vmx/vmx.c +9 -0
+			return 1;
+
+		if (data & ~TEST_CTL_ENABLE_SPLIT_LOCK_DETECT)
+			return 1;
+		vmx->msr_test_ctl = data;
+		break;
 	case MSR_EFER:
 		ret = kvm_set_msr_common(vcpu, msr_info);
 		break;
@@ -4108,6 +4122,9 @@ static void vmx_vcpu_setup(struct vcpu_vmx *vmx)
 
 	vmx->arch_capabilities = kvm_get_arch_capabilities();
 
+	/* disable AC split lock by default */
+	vmx->msr_test_ctl = 0;
+
 	vm_exit_controls_init(vmx, vmx_vmexit_ctrl());
 
 	/* 22.2.1, 20.8.1 */
@@ -4145,6 +4162,7 @@ static void vmx_vcpu_reset(struct kvm_vcpu
 *vcpu,

bool
init_event)

arch/x86/kvm/vmx/vmx.c +1 -0
 	vmx->rmode.vm86_active = 0;
 	vmx->spec_ctrl = 0;
+	vmx->msr_test_ctl = 0;
 
 	vcpu->arch.microcode_version = 0x100000000ULL;
 	vmx->vcpu.arch.regs[VCPU_REGS_RDX] = get_rdx_init_val();
@@ -6344,6 +6362,21 @@ static void atomic_switch_perf_msrs(struct

vcpu_vmx
*vmx)

arch/x86/kvm/vmx/vmx.c +12 -0
 					msrs[i].host, false);
 }
 
+static void atomic_switch_msr_test_ctl(struct vcpu_vmx *vmx)
+{
+	u64 host_msr_test_ctl;
+
+	if (!boot_cpu_has(X86_FEATURE_SPLIT_LOCK_DETECT))
+		return;
+
+	rdmsrl(MSR_TEST_CTL, host_msr_test_ctl);
+	if (host_msr_test_ctl == vmx->msr_test_ctl)
+		clear_atomic_switch_msr(vmx, MSR_TEST_CTL);
+	else
+		add_atomic_switch_msr(vmx, MSR_TEST_CTL, vmx-
Quoted 1 line

msr_test_ctl,

arch/x86/kvm/vmx/vmx.c +5 -0
+				      host_msr_test_ctl, false);
+}
+
 static void vmx_arm_hv_timer(struct vcpu_vmx *vmx, u32 val)
 {
 	vmcs_write32(VMX_PREEMPTION_TIMER_VALUE, val);
@@ -6585,6 +6618,8 @@ static void vmx_vcpu_run(struct kvm_vcpu *vcpu)
 
 	atomic_switch_perf_msrs(vmx);
 
+	atomic_switch_msr_test_ctl(vmx);
+
 	vmx_update_hv_timer(vcpu);
 
 	/*
arch/x86/kvm/vmx/vmx.h +1 -0
index cc22379991f3..e8831609c6c3 100644
@@ -191,6 +191,7 @@ struct vcpu_vmx {
 	u64		      msr_guest_kernel_gs_base;
 #endif
 
+	u64		      msr_test_ctl;
 	u64		      core_capability;
 	u64		      arch_capabilities;
 	u64		      spec_ctrl;
#63 lore raw
In reply to Fenghua Yu
Quoted 4 lines

This capability may also find usage in cloud. A user process with split lock running in one guest can block other cores from accessing shared memory during its split locked memory access. That may cause overall system performance degradation.

"shared memory" ? As in memory shared between two guests?

Quoted 3 lines

Split lock may open a security hole where malicious user code may slow down overall system by executing instructions with split lock.

How much slowdown are we talking?

author
#64 lore raw
Quoted 6 lines · Konrad Rzeszutek Wilk
Quoted 4 lines

This capability may also find usage in cloud. A user process with split lock running in one guest can block other cores from accessing shared memory during its split locked memory access. That may cause overall system performance degradation.

"shared memory" ? As in memory shared between two guests?

Maybe I should remove the "shared" here.

Thanks.

-Fenghua