Hi List,
looking into the performance of syscall entry/exit after s390 switched
to generic entry showed that there's quite some overhead calling some
of the entry/exit work functions even when there's nothing to do.
This patchset moves the entry and exit function to entry-common.h, so
non inlined code gets only called when there is some work pending.
I wrote a small program that just issues invalid syscalls in a loop.
On an s390 machine, this results in the following numbers:
without this series:
$ ./syscall 1000000000
runtime: 94.886581s / per-syscall 9.488658e-08s
with this series:
$ ./syscall 1000000000
runtime: 84.732391s / per-syscall 8.473239e-08s
so the time required for one syscall dropped from 94.8ns to
84.7ns, which is a drop of about 11%.
Sven Schnelle (3):
entry: move exit to usermode functions to header file
entry: move enter_from_user_mode() to header file
entry: move syscall_enter_from_user_mode() to header file
include/linux/entry-common.h | 94 +++++++++++++++++++++++++++++--
kernel/entry/common.c | 106 ++++-------------------------------
2 files changed, 100 insertions(+), 100 deletions(-)
Changes in v2:
- don't move of exit_to_user_mode_loop() to header file
--
2.40.1
To allow inlining of syscall_enter_from_user_mode(), move it
to entry-common.h.
Signed-off-by: Sven Schnelle <[email protected]>
---
include/linux/entry-common.h | 27 +++++++++++++++++++++++++--
kernel/entry/common.c | 32 +-------------------------------
2 files changed, 26 insertions(+), 33 deletions(-)
diff --git a/include/linux/entry-common.h b/include/linux/entry-common.h
index e8f1e4bba1c1..4f8e467eaf3c 100644
--- a/include/linux/entry-common.h
+++ b/include/linux/entry-common.h
@@ -134,6 +134,9 @@ static __always_inline void enter_from_user_mode(struct pt_regs *regs)
*/
void syscall_enter_from_user_mode_prepare(struct pt_regs *regs);
+long syscall_trace_enter(struct pt_regs *regs, long syscall,
+ unsigned long work);
+
/**
* syscall_enter_from_user_mode_work - Check and handle work before invoking
* a syscall
@@ -157,7 +160,15 @@ void syscall_enter_from_user_mode_prepare(struct pt_regs *regs);
* ptrace_report_syscall_entry(), __secure_computing(), trace_sys_enter()
* 2) Invocation of audit_syscall_entry()
*/
-long syscall_enter_from_user_mode_work(struct pt_regs *regs, long syscall);
+static __always_inline long syscall_enter_from_user_mode_work(struct pt_regs *regs, long syscall)
+{
+ unsigned long work = READ_ONCE(current_thread_info()->syscall_work);
+
+ if (work & SYSCALL_WORK_ENTER)
+ syscall = syscall_trace_enter(regs, syscall, work);
+
+ return syscall;
+}
/**
* syscall_enter_from_user_mode - Establish state and check and handle work
@@ -176,7 +187,19 @@ long syscall_enter_from_user_mode_work(struct pt_regs *regs, long syscall);
* Returns: The original or a modified syscall number. See
* syscall_enter_from_user_mode_work() for further explanation.
*/
-long syscall_enter_from_user_mode(struct pt_regs *regs, long syscall);
+static __always_inline long syscall_enter_from_user_mode(struct pt_regs *regs, long syscall)
+{
+ long ret;
+
+ enter_from_user_mode(regs);
+
+ instrumentation_begin();
+ local_irq_enable();
+ ret = syscall_enter_from_user_mode_work(regs, syscall);
+ instrumentation_end();
+
+ return ret;
+}
/**
* local_irq_enable_exit_to_user - Exit to user variant of local_irq_enable()
diff --git a/kernel/entry/common.c b/kernel/entry/common.c
index cd40cd1b4616..a35aaaa9f8bc 100644
--- a/kernel/entry/common.c
+++ b/kernel/entry/common.c
@@ -25,7 +25,7 @@ static inline void syscall_enter_audit(struct pt_regs *regs, long syscall)
}
}
-static long syscall_trace_enter(struct pt_regs *regs, long syscall,
+long syscall_trace_enter(struct pt_regs *regs, long syscall,
unsigned long work)
{
long ret = 0;
@@ -65,36 +65,6 @@ static long syscall_trace_enter(struct pt_regs *regs, long syscall,
return ret ? : syscall;
}
-static __always_inline long
-__syscall_enter_from_user_work(struct pt_regs *regs, long syscall)
-{
- unsigned long work = READ_ONCE(current_thread_info()->syscall_work);
-
- if (work & SYSCALL_WORK_ENTER)
- syscall = syscall_trace_enter(regs, syscall, work);
-
- return syscall;
-}
-
-long syscall_enter_from_user_mode_work(struct pt_regs *regs, long syscall)
-{
- return __syscall_enter_from_user_work(regs, syscall);
-}
-
-noinstr long syscall_enter_from_user_mode(struct pt_regs *regs, long syscall)
-{
- long ret;
-
- enter_from_user_mode(regs);
-
- instrumentation_begin();
- local_irq_enable();
- ret = __syscall_enter_from_user_work(regs, syscall);
- instrumentation_end();
-
- return ret;
-}
-
noinstr void syscall_enter_from_user_mode_prepare(struct pt_regs *regs)
{
enter_from_user_mode(regs);
--
2.40.1
To allow inlining, move exit_to_user_mode() to
entry-common.h.
Signed-off-by: Sven Schnelle <[email protected]>
---
include/linux/entry-common.h | 52 +++++++++++++++++++++++++++++++++++-
kernel/entry/common.c | 50 +++++-----------------------------
2 files changed, 58 insertions(+), 44 deletions(-)
diff --git a/include/linux/entry-common.h b/include/linux/entry-common.h
index d95ab85f96ba..b08aceb26e8e 100644
--- a/include/linux/entry-common.h
+++ b/include/linux/entry-common.h
@@ -7,6 +7,10 @@
#include <linux/syscalls.h>
#include <linux/seccomp.h>
#include <linux/sched.h>
+#include <linux/context_tracking.h>
+#include <linux/livepatch.h>
+#include <linux/resume_user_mode.h>
+#include <linux/tick.h>
#include <asm/entry-common.h>
@@ -258,6 +262,42 @@ static __always_inline void arch_exit_to_user_mode(void) { }
*/
void arch_do_signal_or_restart(struct pt_regs *regs);
+/**
+ * exit_to_user_mode_loop - do any pending work before leaving to user space
+ */
+unsigned long exit_to_user_mode_loop(struct pt_regs *regs,
+ unsigned long ti_work);
+
+/**
+ * exit_to_user_mode_prepare - call exit_to_user_mode_loop() if required
+ *
+ * 1) check that interrupts are disabled
+ * 2) call tick_nohz_user_enter_prepare()
+ * 3) call exit_to_user_mode_loop() if any flags from
+ * EXIT_TO_USER_MODE_WORK are set
+ * 4) check that interrupts are still disabled
+ */
+static __always_inline void exit_to_user_mode_prepare(struct pt_regs *regs)
+{
+ unsigned long ti_work;
+
+ lockdep_assert_irqs_disabled();
+
+ /* Flush pending rcuog wakeup before the last need_resched() check */
+ tick_nohz_user_enter_prepare();
+
+ ti_work = read_thread_flags();
+ if (unlikely(ti_work & EXIT_TO_USER_MODE_WORK))
+ ti_work = exit_to_user_mode_loop(regs, ti_work);
+
+ arch_exit_to_user_mode_prepare(regs, ti_work);
+
+ /* Ensure that kernel state is sane for a return to userspace */
+ kmap_assert_nomap();
+ lockdep_assert_irqs_disabled();
+ lockdep_sys_exit();
+}
+
/**
* exit_to_user_mode - Fixup state when exiting to user mode
*
@@ -276,7 +316,17 @@ void arch_do_signal_or_restart(struct pt_regs *regs);
* non-instrumentable.
* The caller has to invoke syscall_exit_to_user_mode_work() before this.
*/
-void exit_to_user_mode(void);
+static __always_inline void exit_to_user_mode(void)
+{
+ instrumentation_begin();
+ trace_hardirqs_on_prepare();
+ lockdep_hardirqs_on_prepare();
+ instrumentation_end();
+
+ user_enter_irqoff();
+ arch_exit_to_user_mode();
+ lockdep_hardirqs_on(CALLER_ADDR0);
+}
/**
* syscall_exit_to_user_mode_work - Handle work before returning to user mode
diff --git a/kernel/entry/common.c b/kernel/entry/common.c
index d7ee4bc3f2ba..113bd3e8e73e 100644
--- a/kernel/entry/common.c
+++ b/kernel/entry/common.c
@@ -123,29 +123,14 @@ noinstr void syscall_enter_from_user_mode_prepare(struct pt_regs *regs)
instrumentation_end();
}
-/* See comment for exit_to_user_mode() in entry-common.h */
-static __always_inline void __exit_to_user_mode(void)
-{
- instrumentation_begin();
- trace_hardirqs_on_prepare();
- lockdep_hardirqs_on_prepare();
- instrumentation_end();
-
- user_enter_irqoff();
- arch_exit_to_user_mode();
- lockdep_hardirqs_on(CALLER_ADDR0);
-}
-
-void noinstr exit_to_user_mode(void)
-{
- __exit_to_user_mode();
-}
-
/* Workaround to allow gradual conversion of architecture code */
void __weak arch_do_signal_or_restart(struct pt_regs *regs) { }
-static unsigned long exit_to_user_mode_loop(struct pt_regs *regs,
- unsigned long ti_work)
+/**
+ * exit_to_user_mode_loop - do any pending work before leaving to user space
+ */
+__always_inline unsigned long exit_to_user_mode_loop(struct pt_regs *regs,
+ unsigned long ti_work)
{
/*
* Before returning to user space ensure that all pending work
@@ -190,27 +175,6 @@ static unsigned long exit_to_user_mode_loop(struct pt_regs *regs,
return ti_work;
}
-static void exit_to_user_mode_prepare(struct pt_regs *regs)
-{
- unsigned long ti_work;
-
- lockdep_assert_irqs_disabled();
-
- /* Flush pending rcuog wakeup before the last need_resched() check */
- tick_nohz_user_enter_prepare();
-
- ti_work = read_thread_flags();
- if (unlikely(ti_work & EXIT_TO_USER_MODE_WORK))
- ti_work = exit_to_user_mode_loop(regs, ti_work);
-
- arch_exit_to_user_mode_prepare(regs, ti_work);
-
- /* Ensure that kernel state is sane for a return to userspace */
- kmap_assert_nomap();
- lockdep_assert_irqs_disabled();
- lockdep_sys_exit();
-}
-
/*
* If SYSCALL_EMU is set, then the only reason to report is when
* SINGLESTEP is set (i.e. PTRACE_SYSEMU_SINGLESTEP). This syscall
@@ -295,7 +259,7 @@ __visible noinstr void syscall_exit_to_user_mode(struct pt_regs *regs)
instrumentation_begin();
__syscall_exit_to_user_mode_work(regs);
instrumentation_end();
- __exit_to_user_mode();
+ exit_to_user_mode();
}
noinstr void irqentry_enter_from_user_mode(struct pt_regs *regs)
@@ -308,7 +272,7 @@ noinstr void irqentry_exit_to_user_mode(struct pt_regs *regs)
instrumentation_begin();
exit_to_user_mode_prepare(regs);
instrumentation_end();
- __exit_to_user_mode();
+ exit_to_user_mode();
}
noinstr irqentry_state_t irqentry_enter(struct pt_regs *regs)
--
2.40.1
To allow inlining of enter_from_user_mode(), move it to
entry-common.h.
Signed-off-by: Sven Schnelle <[email protected]>
---
include/linux/entry-common.h | 15 ++++++++++++++-
kernel/entry/common.c | 26 +++-----------------------
2 files changed, 17 insertions(+), 24 deletions(-)
diff --git a/include/linux/entry-common.h b/include/linux/entry-common.h
index b08aceb26e8e..e8f1e4bba1c1 100644
--- a/include/linux/entry-common.h
+++ b/include/linux/entry-common.h
@@ -11,6 +11,7 @@
#include <linux/livepatch.h>
#include <linux/resume_user_mode.h>
#include <linux/tick.h>
+#include <linux/kmsan.h>
#include <asm/entry-common.h>
@@ -102,7 +103,19 @@ static __always_inline void arch_enter_from_user_mode(struct pt_regs *regs) {}
* done between establishing state and enabling interrupts. The caller must
* enable interrupts before invoking syscall_enter_from_user_mode_work().
*/
-void enter_from_user_mode(struct pt_regs *regs);
+static __always_inline void enter_from_user_mode(struct pt_regs *regs)
+{
+ arch_enter_from_user_mode(regs);
+ lockdep_hardirqs_off(CALLER_ADDR0);
+
+ CT_WARN_ON(__ct_state() != CONTEXT_USER);
+ user_exit_irqoff();
+
+ instrumentation_begin();
+ kmsan_unpoison_entry_regs(regs);
+ trace_hardirqs_off_finish();
+ instrumentation_end();
+}
/**
* syscall_enter_from_user_mode_prepare - Establish state and enable interrupts
diff --git a/kernel/entry/common.c b/kernel/entry/common.c
index 113bd3e8e73e..cd40cd1b4616 100644
--- a/kernel/entry/common.c
+++ b/kernel/entry/common.c
@@ -15,26 +15,6 @@
#define CREATE_TRACE_POINTS
#include <trace/events/syscalls.h>
-/* See comment for enter_from_user_mode() in entry-common.h */
-static __always_inline void __enter_from_user_mode(struct pt_regs *regs)
-{
- arch_enter_from_user_mode(regs);
- lockdep_hardirqs_off(CALLER_ADDR0);
-
- CT_WARN_ON(__ct_state() != CONTEXT_USER);
- user_exit_irqoff();
-
- instrumentation_begin();
- kmsan_unpoison_entry_regs(regs);
- trace_hardirqs_off_finish();
- instrumentation_end();
-}
-
-void noinstr enter_from_user_mode(struct pt_regs *regs)
-{
- __enter_from_user_mode(regs);
-}
-
static inline void syscall_enter_audit(struct pt_regs *regs, long syscall)
{
if (unlikely(audit_context())) {
@@ -105,7 +85,7 @@ noinstr long syscall_enter_from_user_mode(struct pt_regs *regs, long syscall)
{
long ret;
- __enter_from_user_mode(regs);
+ enter_from_user_mode(regs);
instrumentation_begin();
local_irq_enable();
@@ -117,7 +97,7 @@ noinstr long syscall_enter_from_user_mode(struct pt_regs *regs, long syscall)
noinstr void syscall_enter_from_user_mode_prepare(struct pt_regs *regs)
{
- __enter_from_user_mode(regs);
+ enter_from_user_mode(regs);
instrumentation_begin();
local_irq_enable();
instrumentation_end();
@@ -264,7 +244,7 @@ __visible noinstr void syscall_exit_to_user_mode(struct pt_regs *regs)
noinstr void irqentry_enter_from_user_mode(struct pt_regs *regs)
{
- __enter_from_user_mode(regs);
+ enter_from_user_mode(regs);
}
noinstr void irqentry_exit_to_user_mode(struct pt_regs *regs)
--
2.40.1
Hi Sven,
kernel test robot noticed the following build warnings:
[auto build test WARNING on tip/core/entry]
[also build test WARNING on linus/master v6.7-rc6 next-20231218]
[If your patch is applied to the wrong git tree, kindly drop us a note.
And when submitting patch, we suggest to use '--base' as documented in
https://git-scm.com/docs/git-format-patch#_base_tree_information]
url: https://github.com/intel-lab-lkp/linux/commits/Sven-Schnelle/entry-move-exit-to-usermode-functions-to-header-file/20231218-154733
base: tip/core/entry
patch link: https://lore.kernel.org/r/20231218074520.1998026-2-svens%40linux.ibm.com
patch subject: [PATCH v2 1/3] entry: move exit to usermode functions to header file
config: x86_64-randconfig-161-20231218 (https://download.01.org/0day-ci/archive/20231219/[email protected]/config)
compiler: gcc-12 (Debian 12.2.0-14) 12.2.0
reproduce (this is a W=1 build): (https://download.01.org/0day-ci/archive/20231219/[email protected]/reproduce)
If you fix the issue in a separate patch/commit (i.e. not just a new version of
the same patch/commit), kindly add following tags
| Reported-by: kernel test robot <[email protected]>
| Closes: https://lore.kernel.org/oe-kbuild-all/[email protected]/
All warnings (new ones prefixed by >>):
>> kernel/entry/common.c:134: warning: Function parameter or member 'regs' not described in 'exit_to_user_mode_loop'
>> kernel/entry/common.c:134: warning: Function parameter or member 'ti_work' not described in 'exit_to_user_mode_loop'
vim +134 kernel/entry/common.c
a9f3a74a29af09 Thomas Gleixner 2020-07-22 128
af18f155cb4bda Sven Schnelle 2023-12-18 129 /**
af18f155cb4bda Sven Schnelle 2023-12-18 130 * exit_to_user_mode_loop - do any pending work before leaving to user space
af18f155cb4bda Sven Schnelle 2023-12-18 131 */
af18f155cb4bda Sven Schnelle 2023-12-18 132 __always_inline unsigned long exit_to_user_mode_loop(struct pt_regs *regs,
a9f3a74a29af09 Thomas Gleixner 2020-07-22 133 unsigned long ti_work)
a9f3a74a29af09 Thomas Gleixner 2020-07-22 @134 {
a9f3a74a29af09 Thomas Gleixner 2020-07-22 135 /*
a9f3a74a29af09 Thomas Gleixner 2020-07-22 136 * Before returning to user space ensure that all pending work
a9f3a74a29af09 Thomas Gleixner 2020-07-22 137 * items have been completed.
a9f3a74a29af09 Thomas Gleixner 2020-07-22 138 */
a9f3a74a29af09 Thomas Gleixner 2020-07-22 139 while (ti_work & EXIT_TO_USER_MODE_WORK) {
a9f3a74a29af09 Thomas Gleixner 2020-07-22 140
a9f3a74a29af09 Thomas Gleixner 2020-07-22 141 local_irq_enable_exit_to_user(ti_work);
a9f3a74a29af09 Thomas Gleixner 2020-07-22 142
a9f3a74a29af09 Thomas Gleixner 2020-07-22 143 if (ti_work & _TIF_NEED_RESCHED)
a9f3a74a29af09 Thomas Gleixner 2020-07-22 144 schedule();
a9f3a74a29af09 Thomas Gleixner 2020-07-22 145
a9f3a74a29af09 Thomas Gleixner 2020-07-22 146 if (ti_work & _TIF_UPROBE)
a9f3a74a29af09 Thomas Gleixner 2020-07-22 147 uprobe_notify_resume(regs);
a9f3a74a29af09 Thomas Gleixner 2020-07-22 148
a9f3a74a29af09 Thomas Gleixner 2020-07-22 149 if (ti_work & _TIF_PATCH_PENDING)
a9f3a74a29af09 Thomas Gleixner 2020-07-22 150 klp_update_patch_state(current);
a9f3a74a29af09 Thomas Gleixner 2020-07-22 151
12db8b690010cc Jens Axboe 2020-10-26 152 if (ti_work & (_TIF_SIGPENDING | _TIF_NOTIFY_SIGNAL))
8ba62d37949e24 Eric W. Biederman 2022-02-09 153 arch_do_signal_or_restart(regs);
a9f3a74a29af09 Thomas Gleixner 2020-07-22 154
a68de80f61f6af Sean Christopherson 2021-09-01 155 if (ti_work & _TIF_NOTIFY_RESUME)
03248addadf1a5 Eric W. Biederman 2022-02-09 156 resume_user_mode_work(regs);
a9f3a74a29af09 Thomas Gleixner 2020-07-22 157
a9f3a74a29af09 Thomas Gleixner 2020-07-22 158 /* Architecture specific TIF work */
a9f3a74a29af09 Thomas Gleixner 2020-07-22 159 arch_exit_to_user_mode_work(regs, ti_work);
a9f3a74a29af09 Thomas Gleixner 2020-07-22 160
a9f3a74a29af09 Thomas Gleixner 2020-07-22 161 /*
a9f3a74a29af09 Thomas Gleixner 2020-07-22 162 * Disable interrupts and reevaluate the work flags as they
a9f3a74a29af09 Thomas Gleixner 2020-07-22 163 * might have changed while interrupts and preemption was
a9f3a74a29af09 Thomas Gleixner 2020-07-22 164 * enabled above.
a9f3a74a29af09 Thomas Gleixner 2020-07-22 165 */
a9f3a74a29af09 Thomas Gleixner 2020-07-22 166 local_irq_disable_exit_to_user();
47b8ff194c1fd7 Frederic Weisbecker 2021-02-01 167
47b8ff194c1fd7 Frederic Weisbecker 2021-02-01 168 /* Check if any of the above work has queued a deferred wakeup */
f268c3737ecaef Frederic Weisbecker 2021-05-27 169 tick_nohz_user_enter_prepare();
47b8ff194c1fd7 Frederic Weisbecker 2021-02-01 170
6ce895128b3bff Mark Rutland 2021-11-29 171 ti_work = read_thread_flags();
a9f3a74a29af09 Thomas Gleixner 2020-07-22 172 }
a9f3a74a29af09 Thomas Gleixner 2020-07-22 173
a9f3a74a29af09 Thomas Gleixner 2020-07-22 174 /* Return the latest work state for arch_exit_to_user_mode() */
a9f3a74a29af09 Thomas Gleixner 2020-07-22 175 return ti_work;
a9f3a74a29af09 Thomas Gleixner 2020-07-22 176 }
a9f3a74a29af09 Thomas Gleixner 2020-07-22 177
--
0-DAY CI Kernel Test Service
https://github.com/intel/lkp-tests/wiki
Hi Sven,
kernel test robot noticed the following build warnings:
[auto build test WARNING on tip/core/entry]
[also build test WARNING on linus/master v6.7-rc6 next-20231220]
[If your patch is applied to the wrong git tree, kindly drop us a note.
And when submitting patch, we suggest to use '--base' as documented in
https://git-scm.com/docs/git-format-patch#_base_tree_information]
url: https://github.com/intel-lab-lkp/linux/commits/Sven-Schnelle/entry-move-exit-to-usermode-functions-to-header-file/20231218-154733
base: tip/core/entry
patch link: https://lore.kernel.org/r/20231218074520.1998026-2-svens%40linux.ibm.com
patch subject: [PATCH v2 1/3] entry: move exit to usermode functions to header file
config: x86_64-rhel-8.3-kunit (https://download.01.org/0day-ci/archive/20231221/[email protected]/config)
compiler: gcc-12 (Debian 12.2.0-14) 12.2.0
reproduce (this is a W=1 build): (https://download.01.org/0day-ci/archive/20231221/[email protected]/reproduce)
If you fix the issue in a separate patch/commit (i.e. not just a new version of
the same patch/commit), kindly add following tags
| Reported-by: kernel test robot <[email protected]>
| Closes: https://lore.kernel.org/oe-kbuild-all/[email protected]/
All warnings (new ones prefixed by >>):
>> kernel/entry/common.c:134: warning: Function parameter or struct member 'regs' not described in 'exit_to_user_mode_loop'
>> kernel/entry/common.c:134: warning: Function parameter or struct member 'ti_work' not described in 'exit_to_user_mode_loop'
vim +134 kernel/entry/common.c
a9f3a74a29af09 Thomas Gleixner 2020-07-22 128
af18f155cb4bda Sven Schnelle 2023-12-18 129 /**
af18f155cb4bda Sven Schnelle 2023-12-18 130 * exit_to_user_mode_loop - do any pending work before leaving to user space
af18f155cb4bda Sven Schnelle 2023-12-18 131 */
af18f155cb4bda Sven Schnelle 2023-12-18 132 __always_inline unsigned long exit_to_user_mode_loop(struct pt_regs *regs,
a9f3a74a29af09 Thomas Gleixner 2020-07-22 133 unsigned long ti_work)
a9f3a74a29af09 Thomas Gleixner 2020-07-22 @134 {
a9f3a74a29af09 Thomas Gleixner 2020-07-22 135 /*
a9f3a74a29af09 Thomas Gleixner 2020-07-22 136 * Before returning to user space ensure that all pending work
a9f3a74a29af09 Thomas Gleixner 2020-07-22 137 * items have been completed.
a9f3a74a29af09 Thomas Gleixner 2020-07-22 138 */
a9f3a74a29af09 Thomas Gleixner 2020-07-22 139 while (ti_work & EXIT_TO_USER_MODE_WORK) {
a9f3a74a29af09 Thomas Gleixner 2020-07-22 140
a9f3a74a29af09 Thomas Gleixner 2020-07-22 141 local_irq_enable_exit_to_user(ti_work);
a9f3a74a29af09 Thomas Gleixner 2020-07-22 142
a9f3a74a29af09 Thomas Gleixner 2020-07-22 143 if (ti_work & _TIF_NEED_RESCHED)
a9f3a74a29af09 Thomas Gleixner 2020-07-22 144 schedule();
a9f3a74a29af09 Thomas Gleixner 2020-07-22 145
a9f3a74a29af09 Thomas Gleixner 2020-07-22 146 if (ti_work & _TIF_UPROBE)
a9f3a74a29af09 Thomas Gleixner 2020-07-22 147 uprobe_notify_resume(regs);
a9f3a74a29af09 Thomas Gleixner 2020-07-22 148
a9f3a74a29af09 Thomas Gleixner 2020-07-22 149 if (ti_work & _TIF_PATCH_PENDING)
a9f3a74a29af09 Thomas Gleixner 2020-07-22 150 klp_update_patch_state(current);
a9f3a74a29af09 Thomas Gleixner 2020-07-22 151
12db8b690010cc Jens Axboe 2020-10-26 152 if (ti_work & (_TIF_SIGPENDING | _TIF_NOTIFY_SIGNAL))
8ba62d37949e24 Eric W. Biederman 2022-02-09 153 arch_do_signal_or_restart(regs);
a9f3a74a29af09 Thomas Gleixner 2020-07-22 154
a68de80f61f6af Sean Christopherson 2021-09-01 155 if (ti_work & _TIF_NOTIFY_RESUME)
03248addadf1a5 Eric W. Biederman 2022-02-09 156 resume_user_mode_work(regs);
a9f3a74a29af09 Thomas Gleixner 2020-07-22 157
a9f3a74a29af09 Thomas Gleixner 2020-07-22 158 /* Architecture specific TIF work */
a9f3a74a29af09 Thomas Gleixner 2020-07-22 159 arch_exit_to_user_mode_work(regs, ti_work);
a9f3a74a29af09 Thomas Gleixner 2020-07-22 160
a9f3a74a29af09 Thomas Gleixner 2020-07-22 161 /*
a9f3a74a29af09 Thomas Gleixner 2020-07-22 162 * Disable interrupts and reevaluate the work flags as they
a9f3a74a29af09 Thomas Gleixner 2020-07-22 163 * might have changed while interrupts and preemption was
a9f3a74a29af09 Thomas Gleixner 2020-07-22 164 * enabled above.
a9f3a74a29af09 Thomas Gleixner 2020-07-22 165 */
a9f3a74a29af09 Thomas Gleixner 2020-07-22 166 local_irq_disable_exit_to_user();
47b8ff194c1fd7 Frederic Weisbecker 2021-02-01 167
47b8ff194c1fd7 Frederic Weisbecker 2021-02-01 168 /* Check if any of the above work has queued a deferred wakeup */
f268c3737ecaef Frederic Weisbecker 2021-05-27 169 tick_nohz_user_enter_prepare();
47b8ff194c1fd7 Frederic Weisbecker 2021-02-01 170
6ce895128b3bff Mark Rutland 2021-11-29 171 ti_work = read_thread_flags();
a9f3a74a29af09 Thomas Gleixner 2020-07-22 172 }
a9f3a74a29af09 Thomas Gleixner 2020-07-22 173
a9f3a74a29af09 Thomas Gleixner 2020-07-22 174 /* Return the latest work state for arch_exit_to_user_mode() */
a9f3a74a29af09 Thomas Gleixner 2020-07-22 175 return ti_work;
a9f3a74a29af09 Thomas Gleixner 2020-07-22 176 }
a9f3a74a29af09 Thomas Gleixner 2020-07-22 177
--
0-DAY CI Kernel Test Service
https://github.com/intel/lkp-tests/wiki
The following commit has been merged into the core/entry branch of tip:
Commit-ID: 221a164035fd8b554a44bd7c4bf8e7715a497561
Gitweb: https://git.kernel.org/tip/221a164035fd8b554a44bd7c4bf8e7715a497561
Author: Sven Schnelle <[email protected]>
AuthorDate: Mon, 18 Dec 2023 08:45:20 +01:00
Committer: Thomas Gleixner <[email protected]>
CommitterDate: Thu, 21 Dec 2023 23:12:18 +01:00
entry: Move syscall_enter_from_user_mode() to header file
To allow inlining of syscall_enter_from_user_mode(), move it
to entry-common.h.
Signed-off-by: Sven Schnelle <[email protected]>
Signed-off-by: Thomas Gleixner <[email protected]>
Link: https://lore.kernel.org/r/[email protected]
---
include/linux/entry-common.h | 27 +++++++++++++++++++++++++--
kernel/entry/common.c | 32 +-------------------------------
2 files changed, 26 insertions(+), 33 deletions(-)
diff --git a/include/linux/entry-common.h b/include/linux/entry-common.h
index c420539..b0fb775 100644
--- a/include/linux/entry-common.h
+++ b/include/linux/entry-common.h
@@ -134,6 +134,9 @@ static __always_inline void enter_from_user_mode(struct pt_regs *regs)
*/
void syscall_enter_from_user_mode_prepare(struct pt_regs *regs);
+long syscall_trace_enter(struct pt_regs *regs, long syscall,
+ unsigned long work);
+
/**
* syscall_enter_from_user_mode_work - Check and handle work before invoking
* a syscall
@@ -157,7 +160,15 @@ void syscall_enter_from_user_mode_prepare(struct pt_regs *regs);
* ptrace_report_syscall_entry(), __secure_computing(), trace_sys_enter()
* 2) Invocation of audit_syscall_entry()
*/
-long syscall_enter_from_user_mode_work(struct pt_regs *regs, long syscall);
+static __always_inline long syscall_enter_from_user_mode_work(struct pt_regs *regs, long syscall)
+{
+ unsigned long work = READ_ONCE(current_thread_info()->syscall_work);
+
+ if (work & SYSCALL_WORK_ENTER)
+ syscall = syscall_trace_enter(regs, syscall, work);
+
+ return syscall;
+}
/**
* syscall_enter_from_user_mode - Establish state and check and handle work
@@ -176,7 +187,19 @@ long syscall_enter_from_user_mode_work(struct pt_regs *regs, long syscall);
* Returns: The original or a modified syscall number. See
* syscall_enter_from_user_mode_work() for further explanation.
*/
-long syscall_enter_from_user_mode(struct pt_regs *regs, long syscall);
+static __always_inline long syscall_enter_from_user_mode(struct pt_regs *regs, long syscall)
+{
+ long ret;
+
+ enter_from_user_mode(regs);
+
+ instrumentation_begin();
+ local_irq_enable();
+ ret = syscall_enter_from_user_mode_work(regs, syscall);
+ instrumentation_end();
+
+ return ret;
+}
/**
* local_irq_enable_exit_to_user - Exit to user variant of local_irq_enable()
diff --git a/kernel/entry/common.c b/kernel/entry/common.c
index 0616f23..88cb3c8 100644
--- a/kernel/entry/common.c
+++ b/kernel/entry/common.c
@@ -25,7 +25,7 @@ static inline void syscall_enter_audit(struct pt_regs *regs, long syscall)
}
}
-static long syscall_trace_enter(struct pt_regs *regs, long syscall,
+long syscall_trace_enter(struct pt_regs *regs, long syscall,
unsigned long work)
{
long ret = 0;
@@ -65,36 +65,6 @@ static long syscall_trace_enter(struct pt_regs *regs, long syscall,
return ret ? : syscall;
}
-static __always_inline long
-__syscall_enter_from_user_work(struct pt_regs *regs, long syscall)
-{
- unsigned long work = READ_ONCE(current_thread_info()->syscall_work);
-
- if (work & SYSCALL_WORK_ENTER)
- syscall = syscall_trace_enter(regs, syscall, work);
-
- return syscall;
-}
-
-long syscall_enter_from_user_mode_work(struct pt_regs *regs, long syscall)
-{
- return __syscall_enter_from_user_work(regs, syscall);
-}
-
-noinstr long syscall_enter_from_user_mode(struct pt_regs *regs, long syscall)
-{
- long ret;
-
- enter_from_user_mode(regs);
-
- instrumentation_begin();
- local_irq_enable();
- ret = __syscall_enter_from_user_work(regs, syscall);
- instrumentation_end();
-
- return ret;
-}
-
noinstr void syscall_enter_from_user_mode_prepare(struct pt_regs *regs)
{
enter_from_user_mode(regs);
The following commit has been merged into the core/entry branch of tip:
Commit-ID: d68019471995ba47e56a9da355df13a1cdb5bf7e
Gitweb: https://git.kernel.org/tip/d68019471995ba47e56a9da355df13a1cdb5bf7e
Author: Sven Schnelle <[email protected]>
AuthorDate: Mon, 18 Dec 2023 08:45:18 +01:00
Committer: Thomas Gleixner <[email protected]>
CommitterDate: Thu, 21 Dec 2023 23:12:18 +01:00
entry: Move exit to usermode functions to header file
To allow inlining, move exit_to_user_mode() to
entry-common.h.
Signed-off-by: Sven Schnelle <[email protected]>
Signed-off-by: Thomas Gleixner <[email protected]>
Link: https://lore.kernel.org/r/[email protected]
---
include/linux/entry-common.h | 53 ++++++++++++++++++++++++++++++++++-
kernel/entry/common.c | 52 +++++-----------------------------
2 files changed, 61 insertions(+), 44 deletions(-)
diff --git a/include/linux/entry-common.h b/include/linux/entry-common.h
index d95ab85..6a6e98f 100644
--- a/include/linux/entry-common.h
+++ b/include/linux/entry-common.h
@@ -7,6 +7,10 @@
#include <linux/syscalls.h>
#include <linux/seccomp.h>
#include <linux/sched.h>
+#include <linux/context_tracking.h>
+#include <linux/livepatch.h>
+#include <linux/resume_user_mode.h>
+#include <linux/tick.h>
#include <asm/entry-common.h>
@@ -259,6 +263,43 @@ static __always_inline void arch_exit_to_user_mode(void) { }
void arch_do_signal_or_restart(struct pt_regs *regs);
/**
+ * exit_to_user_mode_loop - do any pending work before leaving to user space
+ */
+unsigned long exit_to_user_mode_loop(struct pt_regs *regs,
+ unsigned long ti_work);
+
+/**
+ * exit_to_user_mode_prepare - call exit_to_user_mode_loop() if required
+ * @regs: Pointer to pt_regs on entry stack
+ *
+ * 1) check that interrupts are disabled
+ * 2) call tick_nohz_user_enter_prepare()
+ * 3) call exit_to_user_mode_loop() if any flags from
+ * EXIT_TO_USER_MODE_WORK are set
+ * 4) check that interrupts are still disabled
+ */
+static __always_inline void exit_to_user_mode_prepare(struct pt_regs *regs)
+{
+ unsigned long ti_work;
+
+ lockdep_assert_irqs_disabled();
+
+ /* Flush pending rcuog wakeup before the last need_resched() check */
+ tick_nohz_user_enter_prepare();
+
+ ti_work = read_thread_flags();
+ if (unlikely(ti_work & EXIT_TO_USER_MODE_WORK))
+ ti_work = exit_to_user_mode_loop(regs, ti_work);
+
+ arch_exit_to_user_mode_prepare(regs, ti_work);
+
+ /* Ensure that kernel state is sane for a return to userspace */
+ kmap_assert_nomap();
+ lockdep_assert_irqs_disabled();
+ lockdep_sys_exit();
+}
+
+/**
* exit_to_user_mode - Fixup state when exiting to user mode
*
* Syscall/interrupt exit enables interrupts, but the kernel state is
@@ -276,7 +317,17 @@ void arch_do_signal_or_restart(struct pt_regs *regs);
* non-instrumentable.
* The caller has to invoke syscall_exit_to_user_mode_work() before this.
*/
-void exit_to_user_mode(void);
+static __always_inline void exit_to_user_mode(void)
+{
+ instrumentation_begin();
+ trace_hardirqs_on_prepare();
+ lockdep_hardirqs_on_prepare();
+ instrumentation_end();
+
+ user_enter_irqoff();
+ arch_exit_to_user_mode();
+ lockdep_hardirqs_on(CALLER_ADDR0);
+}
/**
* syscall_exit_to_user_mode_work - Handle work before returning to user mode
diff --git a/kernel/entry/common.c b/kernel/entry/common.c
index d7ee4bc..7f8f8c1 100644
--- a/kernel/entry/common.c
+++ b/kernel/entry/common.c
@@ -123,29 +123,16 @@ noinstr void syscall_enter_from_user_mode_prepare(struct pt_regs *regs)
instrumentation_end();
}
-/* See comment for exit_to_user_mode() in entry-common.h */
-static __always_inline void __exit_to_user_mode(void)
-{
- instrumentation_begin();
- trace_hardirqs_on_prepare();
- lockdep_hardirqs_on_prepare();
- instrumentation_end();
-
- user_enter_irqoff();
- arch_exit_to_user_mode();
- lockdep_hardirqs_on(CALLER_ADDR0);
-}
-
-void noinstr exit_to_user_mode(void)
-{
- __exit_to_user_mode();
-}
-
/* Workaround to allow gradual conversion of architecture code */
void __weak arch_do_signal_or_restart(struct pt_regs *regs) { }
-static unsigned long exit_to_user_mode_loop(struct pt_regs *regs,
- unsigned long ti_work)
+/**
+ * exit_to_user_mode_loop - do any pending work before leaving to user space
+ * @regs: Pointer to pt_regs on entry stack
+ * @ti_work: TIF work flags as read by the caller
+ */
+__always_inline unsigned long exit_to_user_mode_loop(struct pt_regs *regs,
+ unsigned long ti_work)
{
/*
* Before returning to user space ensure that all pending work
@@ -190,27 +177,6 @@ static unsigned long exit_to_user_mode_loop(struct pt_regs *regs,
return ti_work;
}
-static void exit_to_user_mode_prepare(struct pt_regs *regs)
-{
- unsigned long ti_work;
-
- lockdep_assert_irqs_disabled();
-
- /* Flush pending rcuog wakeup before the last need_resched() check */
- tick_nohz_user_enter_prepare();
-
- ti_work = read_thread_flags();
- if (unlikely(ti_work & EXIT_TO_USER_MODE_WORK))
- ti_work = exit_to_user_mode_loop(regs, ti_work);
-
- arch_exit_to_user_mode_prepare(regs, ti_work);
-
- /* Ensure that kernel state is sane for a return to userspace */
- kmap_assert_nomap();
- lockdep_assert_irqs_disabled();
- lockdep_sys_exit();
-}
-
/*
* If SYSCALL_EMU is set, then the only reason to report is when
* SINGLESTEP is set (i.e. PTRACE_SYSEMU_SINGLESTEP). This syscall
@@ -295,7 +261,7 @@ __visible noinstr void syscall_exit_to_user_mode(struct pt_regs *regs)
instrumentation_begin();
__syscall_exit_to_user_mode_work(regs);
instrumentation_end();
- __exit_to_user_mode();
+ exit_to_user_mode();
}
noinstr void irqentry_enter_from_user_mode(struct pt_regs *regs)
@@ -308,7 +274,7 @@ noinstr void irqentry_exit_to_user_mode(struct pt_regs *regs)
instrumentation_begin();
exit_to_user_mode_prepare(regs);
instrumentation_end();
- __exit_to_user_mode();
+ exit_to_user_mode();
}
noinstr irqentry_state_t irqentry_enter(struct pt_regs *regs)
The following commit has been merged into the core/entry branch of tip:
Commit-ID: caf4062e35b21cd7d3d35ac2f58f9765d02d32a0
Gitweb: https://git.kernel.org/tip/caf4062e35b21cd7d3d35ac2f58f9765d02d32a0
Author: Sven Schnelle <[email protected]>
AuthorDate: Mon, 18 Dec 2023 08:45:19 +01:00
Committer: Thomas Gleixner <[email protected]>
CommitterDate: Thu, 21 Dec 2023 23:12:18 +01:00
entry: Move enter_from_user_mode() to header file
To allow inlining of enter_from_user_mode(), move it to
entry-common.h.
Signed-off-by: Sven Schnelle <[email protected]>
Signed-off-by: Thomas Gleixner <[email protected]>
Link: https://lore.kernel.org/r/[email protected]
---
include/linux/entry-common.h | 15 ++++++++++++++-
kernel/entry/common.c | 26 +++-----------------------
2 files changed, 17 insertions(+), 24 deletions(-)
diff --git a/include/linux/entry-common.h b/include/linux/entry-common.h
index 6a6e98f..c420539 100644
--- a/include/linux/entry-common.h
+++ b/include/linux/entry-common.h
@@ -11,6 +11,7 @@
#include <linux/livepatch.h>
#include <linux/resume_user_mode.h>
#include <linux/tick.h>
+#include <linux/kmsan.h>
#include <asm/entry-common.h>
@@ -102,7 +103,19 @@ static __always_inline void arch_enter_from_user_mode(struct pt_regs *regs) {}
* done between establishing state and enabling interrupts. The caller must
* enable interrupts before invoking syscall_enter_from_user_mode_work().
*/
-void enter_from_user_mode(struct pt_regs *regs);
+static __always_inline void enter_from_user_mode(struct pt_regs *regs)
+{
+ arch_enter_from_user_mode(regs);
+ lockdep_hardirqs_off(CALLER_ADDR0);
+
+ CT_WARN_ON(__ct_state() != CONTEXT_USER);
+ user_exit_irqoff();
+
+ instrumentation_begin();
+ kmsan_unpoison_entry_regs(regs);
+ trace_hardirqs_off_finish();
+ instrumentation_end();
+}
/**
* syscall_enter_from_user_mode_prepare - Establish state and enable interrupts
diff --git a/kernel/entry/common.c b/kernel/entry/common.c
index 7f8f8c1..0616f23 100644
--- a/kernel/entry/common.c
+++ b/kernel/entry/common.c
@@ -15,26 +15,6 @@
#define CREATE_TRACE_POINTS
#include <trace/events/syscalls.h>
-/* See comment for enter_from_user_mode() in entry-common.h */
-static __always_inline void __enter_from_user_mode(struct pt_regs *regs)
-{
- arch_enter_from_user_mode(regs);
- lockdep_hardirqs_off(CALLER_ADDR0);
-
- CT_WARN_ON(__ct_state() != CONTEXT_USER);
- user_exit_irqoff();
-
- instrumentation_begin();
- kmsan_unpoison_entry_regs(regs);
- trace_hardirqs_off_finish();
- instrumentation_end();
-}
-
-void noinstr enter_from_user_mode(struct pt_regs *regs)
-{
- __enter_from_user_mode(regs);
-}
-
static inline void syscall_enter_audit(struct pt_regs *regs, long syscall)
{
if (unlikely(audit_context())) {
@@ -105,7 +85,7 @@ noinstr long syscall_enter_from_user_mode(struct pt_regs *regs, long syscall)
{
long ret;
- __enter_from_user_mode(regs);
+ enter_from_user_mode(regs);
instrumentation_begin();
local_irq_enable();
@@ -117,7 +97,7 @@ noinstr long syscall_enter_from_user_mode(struct pt_regs *regs, long syscall)
noinstr void syscall_enter_from_user_mode_prepare(struct pt_regs *regs)
{
- __enter_from_user_mode(regs);
+ enter_from_user_mode(regs);
instrumentation_begin();
local_irq_enable();
instrumentation_end();
@@ -266,7 +246,7 @@ __visible noinstr void syscall_exit_to_user_mode(struct pt_regs *regs)
noinstr void irqentry_enter_from_user_mode(struct pt_regs *regs)
{
- __enter_from_user_mode(regs);
+ enter_from_user_mode(regs);
}
noinstr void irqentry_exit_to_user_mode(struct pt_regs *regs)