Skip to contentSkip to conversation

uprobes/core: handle breakpoint and signal step exception.

PATCH Superseded Srikar Dronamraju 3 participants 18 messages Feb 23, 2012 – Mar 13, 2012

Subsystems (6)
x86 perf events MM core Scheduler +2 more subsystems

v1 of 12 Superseded Latest is v12 →

lore
Share & tools
lore
#0 lore raw

PATCH 1/1 uprobes/core: handle breakpoint and signal step exception.

From: Srikar Dronamraju <[email protected]>

Uprobes uses exception notifiers to get to know if a thread hit a breakpoint or singlestep exception.

When a thread hits a uprobe or is singlestepping post a uprobe hit, the uprobe exception notifier, sets its TIF_UPROBE bit, which will then be checked on its return to userspace path (do_notify_resume() ->uprobe_notify_resume()), where the consumers handlers are run (in task context) based on the defined filters.

Uprobe hits are thread specific and hence we need to maintain information about if a task hit a uprobe, what uprobe was hit, the slot where the original instruction was copied for xol so that it can be singlestepped with appropriate fixups.

In some cases, special care is needed for instructions that are executed out of line (xol). These are architecture specific artefacts, such as handling RIP relative instructions on x86_64.

Since the instruction at which the uprobe was inserted is executed out of line, architecture specific fixups are added so that the thread continues normal execution in the presence of a uprobe.

Postpone the signals until we execute the probed insn. post_xol() path does a recalc_sigpending() before return to user-mode, this ensures the signal can't be lost.

Uprobes relies on DIE_DEBUG notification to notify if a singlestep is complete.

Adds x86 specific uprobe exception notifiers and appropriate hooks needed to determine a uprobe hit and subsequent post processing.

Add requisite x86 fixups for xol for uprobes. Specific cases needing fixups include relative jumps (x86_64), calls, etc.

Where possible, we check and skip singlestepping the breakpointed instructions. For now we skip single byte as well as few multibyte nop instructions. However this can be extended to other instructions too.

Credit to Oleg Nesterov for suggestions/patches related to signal, breakpoint, singlestep handling code.

	---

Changelog

(v10):
- Resolve comments suggested by Ingo Molnar.
- arch specific functions start with arch_uprobe_

(v9):
- Use instruction_pointer_set instead of set_instruction_pointer.
- Changed can_skip_xol to uprobe_skip_sstep as suggested by Ananth

(v7):
- Resolve comments from Dan Carpenter.
- add_utask returns NULL on error.
- Handles architecture agnostic parts of a uprobe breakpoint hit and subsequent xol singlestepping.

(v6)
- added x86 specific hook for aborting xol.

(v5)
- Use srcu_raw instead of synchronize_sched
- Introduce per task srcu_id to store the srcu_id
- Modified comments.
- No more do a i386 specific enable interrupts. (Its not part of another patch posted separately)

---
 arch/x86/include/asm/thread_info.h |    2 
 arch/x86/include/asm/uprobes.h     |   16 ++
 arch/x86/kernel/signal.c           |    6 +
 arch/x86/kernel/uprobes.c          |  275 +++++++++++++++++++++++++++++++++++
 include/linux/sched.h              |    4 +
 include/linux/uprobes.h            |   44 ++++++
 kernel/events/uprobes.c            |  279 ++++++++++++++++++++++++++++++++++++
 kernel/fork.c                      |    7 +
 kernel/signal.c                    |    6 +
 9 files changed, 629 insertions(+), 10 deletions(-)
arch/x86/include/asm/thread_info.h +2 -0
index af1db7e..2f06cdf 100644
@@ -85,6 +85,7 @@ struct thread_info {
 #define TIF_SECCOMP		8	/* secure computing */
 #define TIF_MCE_NOTIFY		10	/* notify userspace of an MCE */
 #define TIF_USER_RETURN_NOTIFY	11	/* notify kernel of userspace return */
+#define TIF_UPROBE		12	/* breakpointed or singlestepping */
 #define TIF_NOTSC		16	/* TSC is not accessible in userland */
 #define TIF_IA32		17	/* IA32 compatibility process */
 #define TIF_FORK		18	/* ret_from_fork */
@@ -109,6 +110,7 @@ struct thread_info {
 #define _TIF_SECCOMP		(1 << TIF_SECCOMP)
 #define _TIF_MCE_NOTIFY		(1 << TIF_MCE_NOTIFY)
 #define _TIF_USER_RETURN_NOTIFY	(1 << TIF_USER_RETURN_NOTIFY)
+#define _TIF_UPROBE		(1 << TIF_UPROBE)
 #define _TIF_NOTSC		(1 << TIF_NOTSC)
 #define _TIF_IA32		(1 << TIF_IA32)
 #define _TIF_FORK		(1 << TIF_FORK)
arch/x86/include/asm/uprobes.h +15 -1
index f7ce310..d331577 100644
@@ -23,6 +23,8 @@
  *	Jim Keniston
  */
 
+#include <linux/notifier.h>
+
 typedef u8 uprobe_opcode_t;
 
 #define MAX_UINSN_BYTES			  16
@@ -39,5 +41,17 @@ struct arch_uprobe {
 #endif
 };
 
-extern int arch_uprobes_analyze_insn(struct mm_struct *mm, struct arch_uprobe *arch_uprobe);
+struct arch_uprobe_task {
+	unsigned long saved_trap_no;
+#ifdef CONFIG_X86_64
+	unsigned long saved_scratch_register;
+#endif
+};
+
+extern int arch_uprobe_analyze_insn(struct mm_struct *mm, struct arch_uprobe *arch_uprobe);
+extern int arch_uprobe_pre_xol(struct arch_uprobe *auprobe, struct pt_regs *regs);
+extern int arch_uprobe_post_xol(struct arch_uprobe *auprobe, struct pt_regs *regs);
+extern bool arch_uprobe_xol_was_trapped(struct task_struct *tsk);
+extern int arch_uprobe_exception_notify(struct notifier_block *self, unsigned long val, void *data);
+extern void arch_uprobe_abort_xol(struct pt_regs *regs, struct arch_uprobe *auprobe);
 #endif	/* _ASM_UPROBES_H */
arch/x86/kernel/signal.c +6 -0
index 5134e17..153fdaf 100644
@@ -18,6 +18,7 @@
 #include <linux/personality.h>
 #include <linux/uaccess.h>
 #include <linux/user-return-notifier.h>
+#include <linux/uprobes.h>
 
 #include <asm/processor.h>
 #include <asm/ucontext.h>
@@ -824,6 +825,11 @@ do_notify_resume(struct pt_regs *regs, void *unused, __u32 thread_info_flags)
 		mce_notify_process();
 #endif /* CONFIG_X86_64 && CONFIG_X86_MCE */
 
+	if (thread_info_flags & _TIF_UPROBE) {
+		clear_thread_flag(TIF_UPROBE);
+		uprobe_notify_resume(regs);
+	}
+
 	/* deal with pending signal delivery */
 	if (thread_info_flags & _TIF_SIGPENDING)
 		do_signal(regs);
arch/x86/kernel/uprobes.c +269 -6

304 lines not on this page · raw message

include/linux/sched.h +4 -0
index 1a63984..f188558 100644
@@ -1605,6 +1605,10 @@ struct task_struct {
 #ifdef CONFIG_HAVE_HW_BREAKPOINT
 	atomic_t ptrace_bp_refcnt;
 #endif
+#ifdef CONFIG_UPROBES
+	struct uprobe_task *utask;
+	int uprobes_srcu_id;
+#endif
 };
 
 /* Future-safe accessor for struct task_struct's cpus_allowed. */
include/linux/uprobes.h +44 -0

67 lines not on this page · raw message

kernel/events/uprobes.c +277 -2

328 lines not on this page · raw message

kernel/fork.c +7 -0
index b77fd55..df711bb 100644
@@ -66,6 +66,7 @@
 #include <linux/user-return-notifier.h>
 #include <linux/oom.h>
 #include <linux/khugepaged.h>
+#include <linux/uprobes.h>
 
 #include <asm/pgtable.h>
 #include <asm/pgalloc.h>
@@ -700,6 +701,8 @@ void mm_release(struct task_struct *tsk, struct mm_struct *mm)
 		exit_pi_state_list(tsk);
 #endif
 
+	uprobe_free_utask(tsk);
+
 	/* Get rid of any cached register state */
 	deactivate_mm(tsk, mm);
 
@@ -1292,6 +1295,10 @@ static struct task_struct *copy_process(unsigned long clone_flags,
 	INIT_LIST_HEAD(&p->pi_state_list);
 	p->pi_state_cache = NULL;
 #endif
+#ifdef CONFIG_UPROBES
+	p->utask = NULL;
+	p->uprobes_srcu_id = -1;
+#endif
 	/*
 	 * sigaltstack should be cleared when sharing the same VM
 	 */
kernel/signal.c +5 -1
index 8511e39..24a9c39 100644
@@ -29,6 +29,7 @@
 #include <linux/pid_namespace.h>
 #include <linux/nsproxy.h>
 #include <linux/user_namespace.h>
+#include <linux/uprobes.h>
 #define CREATE_TRACE_POINTS
 #include <trace/events/signal.h>
 
@@ -2192,6 +2193,9 @@ int get_signal_to_deliver(siginfo_t *info, struct k_sigaction *return_ka,
 	struct signal_struct *signal = current->signal;
 	int signr;
 
+	if (unlikely(uprobe_deny_signal()))
+		return 0;
+
 relock:
 	/*
 	 * We'll jump back here after any time we were stopped in TASK_STOPPED.
@@ -3004,7 +3008,7 @@ int do_sigaction(int sig, struct k_sigaction *act, struct k_sigaction *oact)
 	return 0;
 }
 
-int 
+int
 do_sigaltstack (const stack_t __user *uss, stack_t __user *uoss, unsigned long sp)
 {
 	stack_t oss;
#1 lore raw

a super minor comment below.

Quoted 5 lines · Srikar Dronamraju

From: Srikar Dronamraju <[email protected]>

Uprobes uses exception notifiers to get to know if a thread hit a breakpoint or singlestep exception.

[snip]

Quoted 12 lines
---
 arch/x86/include/asm/thread_info.h |    2 
 arch/x86/include/asm/uprobes.h     |   16 ++
 arch/x86/kernel/signal.c           |    6 +
 arch/x86/kernel/uprobes.c          |  275 +++++++++++++++++++++++++++++++++++
 include/linux/sched.h              |    4 +
 include/linux/uprobes.h            |   44 ++++++
 kernel/events/uprobes.c            |  279 ++++++++++++++++++++++++++++++++++++
 kernel/fork.c                      |    7 +
 kernel/signal.c                    |    6 +
 9 files changed, 629 insertions(+), 10 deletions(-)

[snip]

kernel/events/uprobes.c static struct uprobe *alloc_uprobe(struct inode *inode, loff_t offset) · line 539

Quoted 9 lines
kernel/events/uprobes.c +9 -0
+static void handler_chain(struct uprobe *uprobe, struct pt_regs *regs)
+{
+	struct uprobe_consumer *consumer;
+
+	if (!(uprobe->flags & UPROBES_RUN_HANDLER))
+		return;
+
+	down_read(&uprobe->consumer_rwsem);
+	consumer = uprobe->consumers;

The line above is not needed, looks like a leftover. Perhaps Ingo can cut it before commit.

kernel/events/uprobes.c static struct uprobe *alloc_uprobe(struct inode *inode, loff_t offset) · line 549

Quoted 10 lines
kernel/events/uprobes.c +7 -0
+	for (consumer = uprobe->consumers; consumer; consumer = consumer->next) {
+		if (!consumer->filter || consumer->filter(consumer, current))
+			consumer->handler(consumer, regs);
+	}
+	up_read(&uprobe->consumer_rwsem);
+}
+
 /* Returns the previous consumer */
 static struct uprobe_consumer *
 consumer_add(struct uprobe *uprobe, struct uprobe_consumer *consumer)

[snip]

Anton.

#2 lore raw
Quoted 12 lines

kernel/events/uprobes.c static struct uprobe *alloc_uprobe(struct inode *inode, loff_t offset) · line 539

Quoted 10 lines
kernel/events/uprobes.c +9 -0
+static void handler_chain(struct uprobe *uprobe, struct pt_regs *regs)
+{
+	struct uprobe_consumer *consumer;
+
+	if (!(uprobe->flags & UPROBES_RUN_HANDLER))
+		return;
+
+	down_read(&uprobe->consumer_rwsem);
+	consumer = uprobe->consumers;

The line above is not needed, looks like a leftover. Perhaps Ingo can cut it before commit.

Right Anton.

Ingo,

Please let me know if you want me to send an updated patch with line removed.

Quoted 9 lines
Quoted 9 lines
kernel/events/uprobes.c +7 -0
+	for (consumer = uprobe->consumers; consumer; consumer = consumer->next) {
+		if (!consumer->filter || consumer->filter(consumer, current))
+			consumer->handler(consumer, regs);
+	}
+	up_read(&uprobe->consumer_rwsem);
+}
+
 /* Returns the previous consumer */
Signature · 2 lines
Thanks and Regards
Srikar
#3 lore raw
In reply to Srikar Dronamraju
Quoted 7 lines · Srikar Dronamraju

From: Srikar Dronamraju <[email protected]>

Uprobes uses exception notifiers to get to know if a thread hit a breakpoint or singlestep exception.

When a thread hits a uprobe or is singlestepping post a uprobe hit, the uprobe exception notifier, sets its TIF_UPROBE bit,

[ only two commas needed in this line ]

Quoted 15 lines

which will then be checked on its return to userspace path (do_notify_resume() ->uprobe_notify_resume()), where the consumers handlers are run (in task context) based on the defined filters.

Uprobe hits are thread specific and hence we need to maintain information about if a task hit a uprobe, what uprobe was hit, the slot where the original instruction was copied for xol so that it can be singlestepped with appropriate fixups.

In some cases, special care is needed for instructions that are executed out of line (xol). These are architecture specific artefacts, such as handling RIP relative instructions on x86_64.

Quoted 7 lines

Since the instruction at which the uprobe was inserted is executed out of line, architecture specific fixups are added so that the thread continues normal execution in the presence of a uprobe.

Postpone the signals until we execute the probed insn. post_xol() path does a recalc_sigpending() before return to user-mode, this ensures the signal can't be lost.

Sigh - this is where a CPU emulator would be so much more robust
- and we already have one in the KVM code.

Quoted 7 lines

Uprobes relies on DIE_DEBUG notification to notify if a singlestep is complete.

Adds x86 specific uprobe exception notifiers and appropriate hooks needed to determine a uprobe hit and subsequent post processing.

Quoted 7 lines

Add requisite x86 fixups for xol for uprobes. Specific cases needing fixups include relative jumps (x86_64), calls, etc.

Where possible, we check and skip singlestepping the breakpointed instructions. For now we skip single byte as well as few multibyte nop instructions. However this can be extended to other instructions too.

Is this an optimization - allowing a NOP to be inserted for easy probe insertion?

Quoted 77 lines

Credit to Oleg Nesterov for suggestions/patches related to signal, breakpoint, singlestep handling code.

	---

Changelog

(v10):
- Resolve comments suggested by Ingo Molnar.
- arch specific functions start with arch_uprobe_

(v9):
- Use instruction_pointer_set instead of set_instruction_pointer.
- Changed can_skip_xol to uprobe_skip_sstep as suggested by Ananth

(v7):
- Resolve comments from Dan Carpenter.
- add_utask returns NULL on error.
- Handles architecture agnostic parts of a uprobe breakpoint hit and subsequent xol singlestepping.

(v6)
- added x86 specific hook for aborting xol.

(v5)
- Use srcu_raw instead of synchronize_sched
- Introduce per task srcu_id to store the srcu_id
- Modified comments.
- No more do a i386 specific enable interrupts. (Its not part of another patch posted separately)

---
 arch/x86/include/asm/thread_info.h |    2 
 arch/x86/include/asm/uprobes.h     |   16 ++
 arch/x86/kernel/signal.c           |    6 +
 arch/x86/kernel/uprobes.c          |  275 +++++++++++++++++++++++++++++++++++
 include/linux/sched.h              |    4 +
 include/linux/uprobes.h            |   44 ++++++
 kernel/events/uprobes.c            |  279 ++++++++++++++++++++++++++++++++++++
 kernel/fork.c                      |    7 +
 kernel/signal.c                    |    6 +
 9 files changed, 629 insertions(+), 10 deletions(-)
arch/x86/include/asm/thread_info.h +2 -0
index af1db7e..2f06cdf 100644
@@ -85,6 +85,7 @@ struct thread_info {
 #define TIF_SECCOMP		8	/* secure computing */
 #define TIF_MCE_NOTIFY		10	/* notify userspace of an MCE */
 #define TIF_USER_RETURN_NOTIFY	11	/* notify kernel of userspace return */
+#define TIF_UPROBE		12	/* breakpointed or singlestepping */
 #define TIF_NOTSC		16	/* TSC is not accessible in userland */
 #define TIF_IA32		17	/* IA32 compatibility process */
 #define TIF_FORK		18	/* ret_from_fork */
@@ -109,6 +110,7 @@ struct thread_info {
 #define _TIF_SECCOMP		(1 << TIF_SECCOMP)
 #define _TIF_MCE_NOTIFY		(1 << TIF_MCE_NOTIFY)
 #define _TIF_USER_RETURN_NOTIFY	(1 << TIF_USER_RETURN_NOTIFY)
+#define _TIF_UPROBE		(1 << TIF_UPROBE)
 #define _TIF_NOTSC		(1 << TIF_NOTSC)
 #define _TIF_IA32		(1 << TIF_IA32)
 #define _TIF_FORK		(1 << TIF_FORK)
arch/x86/include/asm/uprobes.h +2 -0
index f7ce310..d331577 100644
@@ -23,6 +23,8 @@
  *	Jim Keniston
  */
 
+#include <linux/notifier.h>
+
 typedef u8 uprobe_opcode_t;
 
 #define MAX_UINSN_BYTES			  16

arch/x86/include/asm/uprobes.h struct arch_uprobe { · line 45

Quoted 6 lines
arch/x86/include/asm/uprobes.h +2 -1
@@ -39,5 +41,17 @@ struct arch_uprobe {
 #endif
 };
 
-extern int arch_uprobes_analyze_insn(struct mm_struct *mm, struct arch_uprobe *arch_uprobe);
+struct arch_uprobe_task {
+	unsigned long saved_trap_no;

trap_no in struct thread_struct and now in arch_uprobe_task is a misnomer - I always expect a trap_yes flag next to it ;-)

Please create a separate patch in front of this patch that trivially renames trap_no to something saner like trap_nr, and introduce this field as trap_nr as well.

arch/x86/include/asm/uprobes.h struct arch_uprobe { · line 56

Quoted 11 lines
arch/x86/include/asm/uprobes.h +11 -0
+#ifdef CONFIG_X86_64
+	unsigned long saved_scratch_register;
+#endif
+};
+
+extern int arch_uprobe_analyze_insn(struct mm_struct *mm, struct arch_uprobe *arch_uprobe);
+extern int arch_uprobe_pre_xol(struct arch_uprobe *auprobe, struct pt_regs *regs);
+extern int arch_uprobe_post_xol(struct arch_uprobe *auprobe, struct pt_regs *regs);
+extern bool arch_uprobe_xol_was_trapped(struct task_struct *tsk);
+extern int arch_uprobe_exception_notify(struct notifier_block *self, unsigned long val, void *data);
+extern void arch_uprobe_abort_xol(struct pt_regs *regs, struct arch_uprobe *auprobe);

Small API parameter consistency nit:

We tend to order function parameters by importance: the more important parameter comes first.

In that sense arch_uprobe_pre_xol() and arch_uprobe_abort_xol() have obviously inconsistent ordering: the first one is (aup, regs), the second one is (regs, aup).

Please make it consistent, at first glance the right one appears to be:

aup, mm
aup, regs
aup, regs
t
self, val, data
aup, regs

I'd put the arch_uprobe_analyze_insn() order change into a separate patch, preceding this patch.

Quoted 33 lines
arch/x86/include/asm/uprobes.h +0 -0
 #endif	/* _ASM_UPROBES_H */
arch/x86/kernel/signal.c +6 -0
index 5134e17..153fdaf 100644
@@ -18,6 +18,7 @@
 #include <linux/personality.h>
 #include <linux/uaccess.h>
 #include <linux/user-return-notifier.h>
+#include <linux/uprobes.h>
 
 #include <asm/processor.h>
 #include <asm/ucontext.h>
@@ -824,6 +825,11 @@ do_notify_resume(struct pt_regs *regs, void *unused, __u32 thread_info_flags)
 		mce_notify_process();
 #endif /* CONFIG_X86_64 && CONFIG_X86_MCE */
 
+	if (thread_info_flags & _TIF_UPROBE) {
+		clear_thread_flag(TIF_UPROBE);
+		uprobe_notify_resume(regs);
+	}
+
 	/* deal with pending signal delivery */
 	if (thread_info_flags & _TIF_SIGPENDING)
 		do_signal(regs);
arch/x86/kernel/uprobes.c +0 -0
index 04dfcef..5e60ac2 100644
@@ -24,10 +24,17 @@
 #include <linux/sched.h>
 #include <linux/ptrace.h>
 #include <linux/uprobes.h>

arch/x86/kernel/uprobes.c line 36

Quoted 10 lines
arch/x86/kernel/uprobes.c +6 -0
+#include <linux/uaccess.h>
 
 #include <linux/kdebug.h>
 #include <asm/insn.h>
 
+#ifdef CONFIG_X86_32
+#define is_32bit_app(tsk) 1
+#else
+#define is_32bit_app(tsk) (test_tsk_thread_flag(tsk, TIF_IA32))
+#endif

Small detail, we prefer to use this variant:

#ifdef X

# define foo()
#else
# define bar()
#endif

to give the construct more visual structure.

Also, please put it into asm/compat.h and use it at other places within arch/x86/ as well, there's half a dozen similar patterns of TIP_IA32 tests in arch/x86/. Please make this a separate patch, preceding this patch.

Also, how does this interact with the new x32 ABI code in -tip?

Quoted 20 lines
arch/x86/kernel/uprobes.c +4 -4
+
 /* Post-execution fixups. */
 
 /* No fixup needed */
@@ -221,10 +228,9 @@ static int validate_insn_32bits(struct arch_uprobe *auprobe, struct insn *insn)
 }
 
 /*
- * Figure out which fixups post_xol() will need to perform, and annotate
- * arch_uprobe->fixups accordingly.  To start with,
- * arch_uprobe->fixups is either zero or it reflects rip-related
- * fixups.
+ * Figure out which fixups arch_uprobe_post_xol() will need to perform, and
+ * annotate arch_uprobe->fixups accordingly.  To start with,
+ * arch_uprobe->fixups is either zero or it reflects rip-related fixups.
  */
 static void prepare_fixups(struct arch_uprobe *auprobe, struct insn *insn)
 {
@@ -400,12 +406,12 @@ static int validate_insn_bits(struct mm_struct *mm, struct arch_uprobe *auprobe,
 #endif /* CONFIG_X86_64 */

arch/x86/kernel/uprobes.c static int validate_insn_bits(struct mm_struct *mm, struct arch_uprobe *auprobe, · line 414

Quoted 10 lines
arch/x86/kernel/uprobes.c +2 -2
 
 /**
- * arch_uprobes_analyze_insn - instruction analysis including validity and fixups.
+ * arch_uprobe_analyze_insn - instruction analysis including validity and fixups.
  * @mm: the probed address space.
  * @arch_uprobe: the probepoint information.
  * Return 0 on success or a -ve number on error.
  */
-int arch_uprobes_analyze_insn(struct mm_struct *mm, struct arch_uprobe *auprobe)
+int arch_uprobe_analyze_insn(struct mm_struct *mm, struct arch_uprobe *auprobe)

Btw., there's a few more 'uprobes' naming leftovers in the code, like UPROBES_COPY_INSN.

arch/x86/kernel/uprobes.c int arch_uprobes_analyze_insn(struct mm_struct *mm, struct arch_uprobe *auprobe) · line 430

Quoted 9 lines
arch/x86/kernel/uprobes.c +2 -0
 {
 	int ret;
 	struct insn insn;
@@ -420,3 +426,260 @@ int arch_uprobes_analyze_insn(struct mm_struct *mm, struct arch_uprobe *auprobe)
 
 	return 0;
 }
+
+#define	UPROBE_TRAP_NO		UINT_MAX

this would be TRAP_NR too then.

Also, please put such defines at the top of the .c file, so that it's easily in view.

Quoted 11 lines
arch/x86/kernel/uprobes.c +11 -0
+#ifdef CONFIG_X86_64
+/*
+ * If we're emulating a rip-relative instruction, save the contents
+ * of the scratch register and store the target address in that register.
+ */
+static void
+pre_xol_rip_insn(struct arch_uprobe *auprobe, struct pt_regs *regs, struct arch_uprobe_task *tskinfo)
+{
+	if (auprobe->fixups & UPROBES_FIX_RIP_AX) {
+		tskinfo->saved_scratch_register = regs->ax;
+		regs->ax = current->utask->vaddr;

arch/x86/kernel/uprobes.c int arch_uprobes_analyze_insn(struct mm_struct *mm, struct arch_uprobe *auprobe) · line 451

Quoted 10 lines
arch/x86/kernel/uprobes.c +10 -0
+		regs->ax += auprobe->rip_rela_target_address;
+	} else if (auprobe->fixups & UPROBES_FIX_RIP_CX) {
+		tskinfo->saved_scratch_register = regs->cx;
+		regs->cx = current->utask->vaddr;
+		regs->cx += auprobe->rip_rela_target_address;
+	}
+}
+#else
+static void
+pre_xol_rip_insn(struct arch_uprobe *auprobe, struct pt_regs *regs, struct arch_uprobe_task *tskinfo)

If the standard variable naming for 'struct arch_uprobe' is 'auprobe', then the consistent one for 'struct arch_uprobe_task' would be something like 'autask', not 'tskinfo'.

'autask' would be a nice companion to the 'utask' name as well.

Please propagate the new naming through all other uses of struct arch_uprobe_task as well, and through related local variables and structure fields.

Quoted 10 lines
arch/x86/kernel/uprobes.c +10 -0
+{
+	/* No RIP-relative addressing on 32-bit */
+}
+#endif
+
+/*
+ * arch_uprobe_pre_xol - prepare to execute out of line.
+ * @auprobe: the probepoint information.
+ * @regs: reflects the saved user state of @tsk.
+ */

arch/x86/kernel/uprobes.c int arch_uprobes_analyze_insn(struct mm_struct *mm, struct arch_uprobe *auprobe) · line 471

Quoted 10 lines
arch/x86/kernel/uprobes.c +10 -0
+int arch_uprobe_pre_xol(struct arch_uprobe *auprobe, struct pt_regs *regs)
+{
+	struct arch_uprobe_task *tskinfo;
+
+	tskinfo = &current->utask->tskinfo;
+	tskinfo->saved_trap_no = current->thread.trap_no;
+	current->thread.trap_no = UPROBE_TRAP_NO;
+	regs->ip = current->utask->xol_vaddr;
+	pre_xol_rip_insn(auprobe, regs, tskinfo);
+	return 0;

Please put a newline before the return.

Quoted 117 lines
arch/x86/kernel/uprobes.c +117 -0
+}
+
+/*
+ * Called by arch_uprobe_post_xol() to adjust the return address pushed by a
+ * call instruction executed out of line.
+ */
+static int adjust_ret_addr(unsigned long sp, long correction)
+{
+	int rasize, ncopied;
+	long ra = 0;
+
+	if (is_32bit_app(current))
+		rasize = 4;
+	else
+		rasize = 8;
+
+	ncopied = copy_from_user(&ra, (void __user *)sp, rasize);
+	if (unlikely(ncopied))
+		return -EFAULT;
+
+	ra += correction;
+	ncopied = copy_to_user((void __user *)sp, &ra, rasize);
+	if (unlikely(ncopied))
+		return -EFAULT;
+
+	return 0;
+}
+
+#ifdef CONFIG_X86_64
+static bool is_riprel_insn(struct arch_uprobe *auprobe)
+{
+	return ((auprobe->fixups & (UPROBES_FIX_RIP_AX | UPROBES_FIX_RIP_CX)) != 0);
+}
+
+static void handle_riprel_post_xol(struct arch_uprobe *auprobe, struct pt_regs *regs, long *correction)
+{
+	if (is_riprel_insn(auprobe)) {
+		struct arch_uprobe_task *tskinfo;
+
+		tskinfo = &current->utask->tskinfo;
+		if (auprobe->fixups & UPROBES_FIX_RIP_AX)
+			regs->ax = tskinfo->saved_scratch_register;
+		else
+			regs->cx = tskinfo->saved_scratch_register;
+
+		/*
+		 * The original instruction includes a displacement, and so
+		 * is 4 bytes longer than what we've just single-stepped.
+		 * Fall through to handle stuff like "jmpq *...(%rip)" and
+		 * "callq *...(%rip)".
+		 */
+		if (correction)
+			*correction += 4;
+	}
+}
+#else
+static void handle_riprel_post_xol(struct arch_uprobe *auprobe, struct pt_regs *regs, long *correction)
+{
+	/* No RIP-relative addressing on 32-bit */
+}
+#endif
+
+/*
+ * If xol insn itself traps and generates a signal(Say,
+ * SIGILL/SIGSEGV/etc), then detect the case where a singlestepped
+ * instruction jumps back to its own address. It is assumed that anything
+ * like do_page_fault/do_trap/etc sets thread.trap_no != -1.
+ *
+ * arch_uprobe_pre_xol/arch_uprobe_post_xol save/restore thread.trap_no,
+ * arch_uprobe_xol_was_trapped() simply checks that ->trap_no is not equal to
+ * UPROBE_TRAP_NO == -1 set by arch_uprobe_pre_xol().
+ */
+bool arch_uprobe_xol_was_trapped(struct task_struct *tsk)
+{
+	if (tsk->thread.trap_no != UPROBE_TRAP_NO)
+		return true;
+
+	return false;
+}
+
+/*
+ * Called after single-stepping. To avoid the SMP problems that can
+ * occur when we temporarily put back the original opcode to
+ * single-step, we single-stepped a copy of the instruction.
+ *
+ * This function prepares to resume execution after the single-step.
+ * We have to fix things up as follows:
+ *
+ * Typically, the new ip is relative to the copied instruction.  We need
+ * to make it relative to the original instruction (FIX_IP).  Exceptions
+ * are return instructions and absolute or indirect jump or call instructions.
+ *
+ * If the single-stepped instruction was a call, the return address that
+ * is atop the stack is the address following the copied instruction.  We
+ * need to make it the address following the original instruction (FIX_CALL).
+ *
+ * If the original instruction was a rip-relative instruction such as
+ * "movl %edx,0xnnnn(%rip)", we have instead executed an equivalent
+ * instruction using a scratch register -- e.g., "movl %edx,(%rax)".
+ * We need to restore the contents of the scratch register and adjust
+ * the ip, keeping in mind that the instruction we executed is 4 bytes
+ * shorter than the original instruction (since we squeezed out the offset
+ * field).  (FIX_RIP_AX or FIX_RIP_CX)
+ */
+int arch_uprobe_post_xol(struct arch_uprobe *auprobe, struct pt_regs *regs)
+{
+	struct uprobe_task *utask;
+	long correction;
+	int result = 0;
+
+	WARN_ON_ONCE(current->thread.trap_no != UPROBE_TRAP_NO);
+
+	utask = current->utask;
+	current->thread.trap_no = utask->tskinfo.saved_trap_no;
+	correction = (long)(utask->vaddr - utask->xol_vaddr);
+	handle_riprel_post_xol(auprobe, regs, &correction);
+	if (auprobe->fixups & UPROBES_FIX_IP)

arch/x86/kernel/uprobes.c int arch_uprobes_analyze_insn(struct mm_struct *mm, struct arch_uprobe *auprobe) · line 598

Quoted 10 lines
arch/x86/kernel/uprobes.c +10 -0
+		regs->ip += correction;
+
+	if (auprobe->fixups & UPROBES_FIX_CALL)
+		result = adjust_ret_addr(regs->sp, correction);
+
+	return result;
+}
+
+/*
+ * Wrapper routine for handling exceptions.

It's not a wrapper, it's a callback function.

Quoted 6 lines
arch/x86/kernel/uprobes.c +6 -0
+ */
+int arch_uprobe_exception_notify(struct notifier_block *self, unsigned long val, void *data)
+{
+	struct die_args *args = data;
+	struct pt_regs *regs = args->regs;
+	int ret = NOTIFY_DONE;

arch/x86/kernel/uprobes.c int arch_uprobes_analyze_insn(struct mm_struct *mm, struct arch_uprobe *auprobe) · line 614

Quoted 10 lines
arch/x86/kernel/uprobes.c +10 -0
+
+	/* We are only interested in userspace traps */
+	if (regs && !user_mode_vm(regs))
+		return NOTIFY_DONE;
+
+	switch (val) {
+	case DIE_INT3:
+		/* Run your handler here */
+		if (uprobe_bkpt_notifier(regs))
+			ret = NOTIFY_STOP;

This comment looks somewhat out of place.

Also, I have not noticed this in the first patch, but 'bkpt' is not a standard way to refer to breakpoints: we either use 'breakpoint' or 'bp'.

arch/x86/kernel/uprobes.c int arch_uprobes_analyze_insn(struct mm_struct *mm, struct arch_uprobe *auprobe) · line 620

Quoted 6 lines
arch/x86/kernel/uprobes.c +6 -0
+
+		break;
+
+	case DIE_DEBUG:
+		if (uprobe_post_notifier(regs))
+			ret = NOTIFY_STOP;

This wants to be post_bp_notifier, right?

I'd suggest to name them in a symmetric way:

   uprobe_pre_bp_notifier();
   uprobe_post_bp_notifier();

so that the connection and ordering is obvious in the higher levels as well.

arch/x86/kernel/uprobes.c int arch_uprobes_analyze_insn(struct mm_struct *mm, struct arch_uprobe *auprobe) · line 631

Quoted 11 lines
arch/x86/kernel/uprobes.c +11 -0
+
+	default:
+		break;
+	}
+
+	return ret;
+}
+
+/*
+ * xol insn either trapped or thread has a fatal signal, so reset the
+ * instruction pointer to its probed address.

Verb tense mismatch at the beginning of the sentence. It is also very friendly to do round, explanatory sentences like:

'This function gets called when the XOL instruction either
trapped or the thread had a fatal signal, ...'

Please review all other comments for similar patterns as well and fix them.

Quoted 17 lines
arch/x86/kernel/uprobes.c +17 -0
+ */
+void arch_uprobe_abort_xol(struct pt_regs *regs, struct arch_uprobe *auprobe)
+{
+	struct uprobe_task *utask = current->utask;
+
+	current->thread.trap_no = utask->tskinfo.saved_trap_no;
+	handle_riprel_post_xol(auprobe, regs, NULL);
+	instruction_pointer_set(regs, utask->vaddr);
+}
+
+/*
+ * Skip these instructions:
+ *
+ * 0f 19 90 90 90 90 90		nopl   -0x6f6f6f70(%rax)
+ * 0f 1f 00			nopl (%rax)
+ * 0f 1f 40 00			nopl 0x0(%rax)
+ * 0f 1f 44 00 00		nopl 0x0(%rax,%rax,1)

arch/x86/kernel/uprobes.c int arch_uprobes_analyze_insn(struct mm_struct *mm, struct arch_uprobe *auprobe) · line 658

Quoted 10 lines
arch/x86/kernel/uprobes.c +10 -0
+ * 0f 1f 80 00 00 00 00		nopl 0x0(%rax)
+ * 0f 1f 84 00 00 00 00		nopl 0x0(%rax,%rax,1)
+ * 66 0f 1f 44 00 00 00		nopw 0x0(%rax,%rax,1)
+ * 66 0f 1f 84 00 00 00		nopw 0x0(%rax,%rax,1)
+ * 66 87 c0			xchg %eax,%eax
+ * 66 90			nop
+ * 87 c0			xchg %eax,%eax
+ * 90 				nop
+ */
+

unnecessary newline.

arch/x86/kernel/uprobes.c int arch_uprobes_analyze_insn(struct mm_struct *mm, struct arch_uprobe *auprobe) · line 671

Quoted 13 lines
arch/x86/kernel/uprobes.c +13 -0
+bool arch_uprobe_skip_sstep(struct pt_regs *regs, struct arch_uprobe *auprobe)
+{
+	int i;
+
+	for (i = 0; i < MAX_UINSN_BYTES; i++) {
+		if ((auprobe->insn[i] == 0x66))
+			continue;
+
+		if (auprobe->insn[i] == 0x90)
+			return true;
+
+		if (i == (MAX_UINSN_BYTES - 1))
+			break;

Looks like the loop could run from 0 to MAX_UINSN_BYTES-2 and then this break would be superfluous.

arch/x86/kernel/uprobes.c int arch_uprobes_analyze_insn(struct mm_struct *mm, struct arch_uprobe *auprobe) · line 684

Quoted 13 lines
arch/x86/kernel/uprobes.c +13 -0
+
+		if ((auprobe->insn[i] == 0x0f) && (auprobe->insn[i+1] == 0x1f))
+			return true;
+
+		if ((auprobe->insn[i] == 0x0f) && (auprobe->insn[i+1] == 0x19))
+			return true;
+
+		if ((auprobe->insn[i] == 0x87) && (auprobe->insn[i+1] == 0xc0))
+			return true;
+
+		break;
+	}
+	return false;

Looks like we will skip not just the instructions listed above, but a lot more patterns:

0x66* { 0x90 | 0x0f 0x1f | 0x0f 0x19 | 0x87 0xc0 }

It might make sense to mention the pattern in the comment and mention that it's intended to skip the instructions listed, under the currently known x86 ISA.

Quoted 37 lines
arch/x86/kernel/uprobes.c +1 -0
+}
include/linux/sched.h +4 -0
index 1a63984..f188558 100644
@@ -1605,6 +1605,10 @@ struct task_struct {
 #ifdef CONFIG_HAVE_HW_BREAKPOINT
 	atomic_t ptrace_bp_refcnt;
 #endif
+#ifdef CONFIG_UPROBES
+	struct uprobe_task *utask;
+	int uprobes_srcu_id;
+#endif
 };
 
 /* Future-safe accessor for struct task_struct's cpus_allowed. */
include/linux/uprobes.h +6 -0
index f85797e..73c8711 100644
@@ -39,6 +39,8 @@ struct vm_area_struct;
 
 /* Dont run handlers when first register/ last unregister in progress*/
 #define UPROBES_RUN_HANDLER	0x2
+/* Can skip singlestep */
+#define UPROBES_SKIP_SSTEP	0x4
 
 struct uprobe_consumer {
 	int (*handler)(struct uprobe_consumer *self, struct pt_regs *regs);
@@ -52,12 +54,40 @@ struct uprobe_consumer {
 };
 
 #ifdef CONFIG_UPROBES
+enum uprobe_task_state {
+	UTASK_RUNNING,
+	UTASK_BP_HIT,
+	UTASK_SSTEP,

include/linux/uprobes.h struct uprobe_consumer { · line 70

Quoted 10 lines
include/linux/uprobes.h +10 -0
+	UTASK_SSTEP_ACK,
+	UTASK_SSTEP_TRAPPED,
+};
+
+/*
+ * uprobe_task: Metadata of a task while it singlesteps.
+ */
+struct uprobe_task {
+	unsigned long xol_vaddr;
+	unsigned long vaddr;

These two fields are never actually used outside of architecture code.

Unless there's a good reason to keep them outside I'd suggest to move them into struct arch_uprobe_task. This has another
benefit: we can pass struct arch_uprobe_task to the architecture methods, instead of struct uprobe_task. This would allow the moving of the struct uprobe_task into uprobes.c - no code outside uprobes.c needs to know its structure.

include/linux/uprobes.h struct uprobe_consumer { · line 76

Quoted 6 lines
include/linux/uprobes.h +6 -0
+
+	enum uprobe_task_state state;
+	struct arch_uprobe_task tskinfo;
+
+	struct uprobe *active_uprobe;
+};

Also, please check the style of 'struct uprobe' and make 'struct uprobe_task' match that style.

Quoted 22 lines
include/linux/uprobes.h +9 -0
+
 extern int __weak set_bkpt(struct mm_struct *mm, struct arch_uprobe *auprobe, unsigned long vaddr);
 extern int __weak set_orig_insn(struct mm_struct *mm, struct arch_uprobe *auprobe, unsigned long vaddr, bool verify);
 extern bool __weak is_bkpt_insn(uprobe_opcode_t *insn);
 extern int uprobe_register(struct inode *inode, loff_t offset, struct uprobe_consumer *consumer);
 extern void uprobe_unregister(struct inode *inode, loff_t offset, struct uprobe_consumer *consumer);
 extern int uprobe_mmap(struct vm_area_struct *vma);
+extern void uprobe_free_utask(struct task_struct *tsk);
+extern unsigned long __weak get_uprobe_bkpt_addr(struct pt_regs *regs);
+extern int uprobe_post_notifier(struct pt_regs *regs);
+extern int uprobe_bkpt_notifier(struct pt_regs *regs);
+extern void uprobe_notify_resume(struct pt_regs *regs);
+extern bool uprobe_deny_signal(void);
+extern bool __weak arch_uprobe_skip_sstep(struct pt_regs *regs, struct arch_uprobe *auprobe);
 #else /* CONFIG_UPROBES is not defined */
 static inline int
 uprobe_register(struct inode *inode, loff_t offset, struct uprobe_consumer *consumer)
@@ -72,5 +102,19 @@ static inline int uprobe_mmap(struct vm_area_struct *vma)
 {
 	return 0;
 }
+static inline void uprobe_notify_resume(struct pt_regs *regs)

include/linux/uprobes.h static inline int uprobe_mmap(struct vm_area_struct *vma) · line 115

Quoted 10 lines
include/linux/uprobes.h +10 -0
+{
+}
+static inline bool uprobe_deny_signal(void)
+{
+	return false;
+}
+static inline unsigned long get_uprobe_bkpt_addr(struct pt_regs *regs)
+{
+	return 0;
+}

Please use the standard uprobe method naming pattern for get_uprobe_bkpt_addr().

Quoted 39 lines
include/linux/uprobes.h +3 -0
+static inline void uprobe_free_utask(struct task_struct *tsk)
+{
+}
 #endif /* CONFIG_UPROBES */
 #endif	/* _LINUX_UPROBES_H */
kernel/events/uprobes.c +9 -0
index 5ce32e3..cc955e8 100644
@@ -30,9 +30,12 @@
 #include <linux/rmap.h>		/* anon_vma_prepare */
 #include <linux/mmu_notifier.h>	/* set_pte_at_notify */
 #include <linux/swap.h>		/* try_to_free_swap */
+#include <linux/ptrace.h>	/* user_enable_single_step */
+#include <linux/kdebug.h>	/* notifier mechanism */
 
 #include <linux/uprobes.h>
 
+static struct srcu_struct uprobes_srcu;
 static struct rb_root uprobes_tree = RB_ROOT;
 
 static DEFINE_SPINLOCK(uprobes_treelock);	/* serialize rbtree access */
@@ -486,6 +489,8 @@ static struct uprobe *insert_uprobe(struct uprobe *uprobe)
 	u = __insert_uprobe(uprobe);
 	spin_unlock_irqrestore(&uprobes_treelock, flags);
 
+	/* For now assume that the instruction need not be single-stepped */
+	uprobe->flags |= UPROBES_SKIP_SSTEP;
 	return u;
 }
 
@@ -523,6 +528,22 @@ static struct uprobe *alloc_uprobe(struct inode *inode, loff_t offset)
 	return uprobe;
 }
 
+static void handler_chain(struct uprobe *uprobe, struct pt_regs *regs)
+{
+	struct uprobe_consumer *consumer;
+

kernel/events/uprobes.c static struct uprobe *alloc_uprobe(struct inode *inode, loff_t offset) · line 544

Quoted 10 lines
kernel/events/uprobes.c +10 -0
+	if (!(uprobe->flags & UPROBES_RUN_HANDLER))
+		return;
+
+	down_read(&uprobe->consumer_rwsem);
+	consumer = uprobe->consumers;
+	for (consumer = uprobe->consumers; consumer; consumer = consumer->next) {
+		if (!consumer->filter || consumer->filter(consumer, current))
+			consumer->handler(consumer, regs);
+	}
+	up_read(&uprobe->consumer_rwsem);

Suggestion: as a local variable 'consumer' is mentioned many, many times throughout the code. It might make sense to standardize, for the strict case of local variables (not structure fields), on the following pattern:

	struct uprobe_consumer *uc;

	...

	uc = uprobe->consumers;
	for (uc = uprobe->consumers; uc; uc = uc->next) {
		if (!uc->filter || uc->filter(uc, current))
			uc->handler(uc, regs);
	}

See how much more compact and readable it is? This shorter form also makes it immediately obvious that the first line before the loop is superfulous, it's already done by the loop initializer.

Please propagate this new naming through other places within uprobes.c as well.

Quoted 17 lines
kernel/events/uprobes.c +3 -1
+}
+
 /* Returns the previous consumer */
 static struct uprobe_consumer *
 consumer_add(struct uprobe *uprobe, struct uprobe_consumer *consumer)
@@ -643,7 +664,7 @@ static int install_breakpoint(struct mm_struct *mm, struct uprobe *uprobe,
 		if (is_bkpt_insn((uprobe_opcode_t *)uprobe->arch.insn))
 			return -EEXIST;
 
-		ret = arch_uprobes_analyze_insn(mm, &uprobe->arch);
+		ret = arch_uprobe_analyze_insn(mm, &uprobe->arch);
 		if (ret)
 			return ret;
 
@@ -659,10 +680,21 @@ static void remove_breakpoint(struct mm_struct *mm, struct uprobe *uprobe, loff_
 	set_orig_insn(mm, &uprobe->arch, (unsigned long)vaddr, true);
 }

kernel/events/uprobes.c static void remove_breakpoint(struct mm_struct *mm, struct uprobe *uprobe, loff_ · line 691

Quoted 10 lines
kernel/events/uprobes.c +9 -0
 
+/*
+ * There could be threads that have hit the breakpoint and are entering the
+ * notifier code and trying to acquire the uprobes_treelock. The thread
+ * calling delete_uprobe() that is removing the uprobe from the rb_tree can
+ * race with these threads and might acquire the uprobes_treelock compared
+ * to some of the breakpoint hit threads. In such a case, the breakpoint hit
+ * threads will not find the uprobe. Hence wait till the current breakpoint
+ * hit threads acquire the uprobes_treelock before the uprobe is removed
+ * from the rbtree.

Hm, the last sentence does not parse for me. (even if it's correct English it might make sense to rephrase it to be clearer what is meant.)

Quoted 50 lines
kernel/events/uprobes.c +39 -0
+ */
 static void delete_uprobe(struct uprobe *uprobe)
 {
 	unsigned long flags;
 
+	synchronize_srcu(&uprobes_srcu);
 	spin_lock_irqsave(&uprobes_treelock, flags);
 	rb_erase(&uprobe->rb_node, &uprobes_tree);
 	spin_unlock_irqrestore(&uprobes_treelock, flags);
@@ -1007,6 +1039,248 @@ int uprobe_mmap(struct vm_area_struct *vma)
 	return ret;
 }
 
+/**
+ * get_uprobe_bkpt_addr - compute address of bkpt given post-bkpt regs
+ * @regs: Reflects the saved state of the task after it has hit a breakpoint
+ * instruction.
+ * Return the address of the breakpoint instruction.
+ */
+unsigned long __weak get_uprobe_bkpt_addr(struct pt_regs *regs)
+{
+	return instruction_pointer(regs) - UPROBES_BKPT_INSN_SIZE;
+}
+
+/*
+ * Called with no locks held.
+ * Called in context of a exiting or a exec-ing thread.
+ */
+void uprobe_free_utask(struct task_struct *tsk)
+{
+	struct uprobe_task *utask = tsk->utask;
+
+	if (tsk->uprobes_srcu_id != -1)
+		srcu_read_unlock_raw(&uprobes_srcu, tsk->uprobes_srcu_id);
+
+	if (!utask)
+		return;
+
+	if (utask->active_uprobe)
+		put_uprobe(utask->active_uprobe);
+
+	kfree(utask);
+	tsk->utask = NULL;
+}
+
+/*
+ * Allocate a uprobe_task object for the task.
+ * Called when the thread hits a breakpoint for the first time.
+ *
+ * Returns:

kernel/events/uprobes.c int uprobe_mmap(struct vm_area_struct *vma) · line 1088

Quoted 10 lines
kernel/events/uprobes.c +10 -0
+ * - pointer to new uprobe_task on success
+ * - NULL otherwise
+ */
+static struct uprobe_task *add_utask(void)
+{
+	struct uprobe_task *utask;
+
+	utask = kzalloc(sizeof *utask, GFP_KERNEL);
+	if (unlikely(utask == NULL))
+		return NULL;

!task

kernel/events/uprobes.c int uprobe_mmap(struct vm_area_struct *vma) · line 1092

Quoted 4 lines
kernel/events/uprobes.c +4 -0
+
+	utask->active_uprobe = NULL;
+	current->utask = utask;
+	return utask;

newline.

Quoted 31 lines
kernel/events/uprobes.c +31 -0
+}
+
+/* Prepare to single-step probed instruction out of line. */
+static int pre_ssout(struct uprobe *uprobe, struct pt_regs *regs, unsigned long vaddr)
+{
+	return -EFAULT;
+}
+
+/*
+ * If we are singlestepping, then ensure this thread is not connected to
+ * non-fatal signals until completion of singlestep.  When xol insn itself
+ * triggers the signal,  restart the original insn even if the task is
+ * already SIGKILL'ed (since coredump should report the correct ip).  This
+ * is even more important if the task has a handler for SIGSEGV/etc, The
+ * _same_ instruction should be repeated again after return from the signal
+ * handler, and SSTEP can never finish in this case.
+ */
+bool uprobe_deny_signal(void)
+{
+	struct task_struct *tsk = current;
+	struct uprobe_task *utask = tsk->utask;
+
+	if (likely(!utask || !utask->active_uprobe))
+		return false;
+
+	WARN_ON_ONCE(utask->state != UTASK_SSTEP);
+
+	if (signal_pending(tsk)) {
+		spin_lock_irq(&tsk->sighand->siglock);
+		clear_tsk_thread_flag(tsk, TIF_SIGPENDING);
+		spin_unlock_irq(&tsk->sighand->siglock);

kernel/events/uprobes.c int uprobe_mmap(struct vm_area_struct *vma) · line 1133

Quoted 10 lines
kernel/events/uprobes.c +10 -0
+
+		if (__fatal_signal_pending(tsk) || arch_uprobe_xol_was_trapped(tsk)) {
+			utask->state = UTASK_SSTEP_TRAPPED;
+			set_tsk_thread_flag(tsk, TIF_UPROBE);
+			set_tsk_thread_flag(tsk, TIF_NOTIFY_RESUME);
+		}
+	}
+
+	return true;
+}

Hm, seems racy to me: what happens if we get a signal shortly *after* this code got called and before we take the siglock? It might be safe, but it's not clearly obvious why and the code does not explain it.

Why is this logic not within the 'relock' loop within get_signal_to_deliver(), called with the siglock already taken? That would avoid races and would also remove the siglock taking/dropping above.

Quoted 5 lines
kernel/events/uprobes.c +5 -0
+
+static bool uprobe_skip_sstep(struct pt_regs *regs, struct uprobe *u)
+{
+	if (arch_uprobe_skip_sstep(regs, &u->arch))
+		return true;

kernel/events/uprobes.c int uprobe_mmap(struct vm_area_struct *vma) · line 1148

Quoted 10 lines
kernel/events/uprobes.c +10 -0
+
+	u->flags &= ~UPROBES_SKIP_SSTEP;
+	return false;
+}
+
+/*
+ * On breakpoint hit, breakpoint notifier sets the TIF_UPROBE flag.  (and on
+ * subsequent probe hits on the thread sets the state to UTASK_BP_HIT) and
+ * allows the thread to return from interrupt.  While returning to
+ * userspace, thread noticies the TIF_UPROBE flag and calls

typo.

Quoted 8 lines
kernel/events/uprobes.c +8 -0
+ * uprobe_notify_resume(). uprobe_notify_resume will run the handler and ask
+ * the thread to singlestep.
+ *
+ * On subsequent singlestep exception, singlestep notifier sets the
+ * TIF_UPROBE flag and also sets the state to UTASK_SSTEP_ACK and allows the
+ * thread to return from interrupt. While returning to userspace, thread
+ * notices the TIF_UPROBE and calls uprobe_notify_resume().
+ * uprobe_notify_resume disables singlestep and performs the required

kernel/events/uprobes.c int uprobe_mmap(struct vm_area_struct *vma) · line 1166

Quoted 10 lines
kernel/events/uprobes.c +10 -0
+ * fix-ups.
+ *
+ * All non-fatal signals cannot interrupt thread while the thread singlesteps.
+ */
+void uprobe_notify_resume(struct pt_regs *regs)
+{
+	struct vm_area_struct *vma;
+	struct uprobe_task *utask;
+	struct mm_struct *mm;
+	struct uprobe *u;

'u' is not what we use for a uprobe in other places - please use consistent naming all across the code!

kernel/events/uprobes.c int uprobe_mmap(struct vm_area_struct *vma) · line 1167

Quoted 1 line
kernel/events/uprobes.c +1 -0
+	unsigned long probept;

this wants to be the breakpoint virtual address, right?

If yes then 'bp_vaddr' would be a *much* clearer name for it.

kernel/events/uprobes.c int uprobe_mmap(struct vm_area_struct *vma) · line 1172

Quoted 5 lines
kernel/events/uprobes.c +5 -0
+
+	utask = current->utask;
+	u = NULL;
+	mm = current->mm;
+	if (!utask || utask->state == UTASK_BP_HIT) {

Please put these two starkly different pieces of functionality on the two sides of the branch into two helper inline functions, named apporpriately.

kernel/events/uprobes.c int uprobe_mmap(struct vm_area_struct *vma) · line 1179

Quoted 7 lines
kernel/events/uprobes.c +7 -0
+		probept = get_uprobe_bkpt_addr(regs);
+		down_read(&mm->mmap_sem);
+		vma = find_vma(mm, probept);
+
+		if (vma && vma->vm_start <= probept && valid_vma(vma, false))
+			u = find_uprobe(vma->vm_file->f_mapping->host, probept - vma->vm_start +
+					(vma->vm_pgoff << PAGE_SHIFT));

I'd stick the offset into a loff_t local variable first, plus use an inode variable as well, then it would all be so much easier to read:

			inode = vma->vm_file->f_mapping->host;
			offset = vma->vm_pgoff << PAGE_SHIFT;
			offset += bp_vaddr - vma->vm_start;

			uprobe = find_uprobe(inode, offset);
Quoted 4 lines
kernel/events/uprobes.c +4 -0
+
+		srcu_read_unlock_raw(&uprobes_srcu, current->uprobes_srcu_id);
+		current->uprobes_srcu_id = -1;
+		up_read(&mm->mmap_sem);

kernel/events/uprobes.c int uprobe_mmap(struct vm_area_struct *vma) · line 1193

Quoted 10 lines
kernel/events/uprobes.c +10 -0
+
+		if (!u)
+			/* No matching uprobe; signal SIGTRAP. */
+			goto cleanup_ret;
+
+		if (!utask) {
+			utask = add_utask();
+			/* Cannot Allocate; re-execute the instruction. */
+			if (!utask)
+				goto cleanup_ret;

Weird capitalization.

kernel/events/uprobes.c int uprobe_mmap(struct vm_area_struct *vma) · line 1205

Quoted 12 lines
kernel/events/uprobes.c +12 -0
+		}
+		utask->active_uprobe = u;
+		handler_chain(u, regs);
+		if (u->flags & UPROBES_SKIP_SSTEP && uprobe_skip_sstep(regs, u))
+			goto cleanup_ret;
+
+		utask->state = UTASK_SSTEP;
+		if (!pre_ssout(u, regs, probept))
+			user_enable_single_step(current);
+		else
+			/* Cannot Singlestep; re-execute the instruction. */
+			goto cleanup_ret;

Weird capitalization.

Quoted 24 lines
kernel/events/uprobes.c +24 -0
+
+	} else {
+		u = utask->active_uprobe;
+		if (utask->state == UTASK_SSTEP_ACK)
+			arch_uprobe_post_xol(&u->arch, regs);
+		else if (utask->state == UTASK_SSTEP_TRAPPED)
+			arch_uprobe_abort_xol(regs, &u->arch);
+		else
+			WARN_ON_ONCE(1);
+
+		put_uprobe(u);
+		utask->active_uprobe = NULL;
+		utask->state = UTASK_RUNNING;
+		user_disable_single_step(current);
+
+		spin_lock_irq(&current->sighand->siglock);
+		recalc_sigpending(); /* see uprobe_deny_signal() */
+		spin_unlock_irq(&current->sighand->siglock);
+	}
+	return;
+
+cleanup_ret:
+	if (utask) {
+		utask->active_uprobe = NULL;

kernel/events/uprobes.c int uprobe_mmap(struct vm_area_struct *vma) · line 1239

Quoted 10 lines
kernel/events/uprobes.c +10 -0
+		utask->state = UTASK_RUNNING;
+	}
+	if (u) {
+		if (!(u->flags & UPROBES_SKIP_SSTEP))
+			instruction_pointer_set(regs, probept);
+
+		put_uprobe(u);
+	} else {
+		send_sig(SIGTRAP, current, 0);
+	}

Please use a standard cleanup sequence.

The code flow is extremely mixed throughout uprobe_notify_resume(), please clean it all up and make it much more obvious.

Quoted 32 lines
kernel/events/uprobes.c +32 -0
+}
+
+/*
+ * uprobe_bkpt_notifier gets called from interrupt context as part of
+ * notifier mechanism. Set TIF_UPROBE flag and indicate breakpoint hit.
+ */
+int uprobe_bkpt_notifier(struct pt_regs *regs)
+{
+	struct uprobe_task *utask;
+
+	if (!current->mm)
+		return 0;
+
+	utask = current->utask;
+	if (utask)
+		utask->state = UTASK_BP_HIT;
+
+	set_thread_flag(TIF_UPROBE);
+	current->uprobes_srcu_id = srcu_read_lock_raw(&uprobes_srcu);
+	return 1;
+}
+
+/*
+ * uprobe_post_notifier gets called in interrupt context as part of notifier
+ * mechanism. Set TIF_UPROBE flag and indicate completion of singlestep.
+ */
+int uprobe_post_notifier(struct pt_regs *regs)
+{
+	struct uprobe_task *utask = current->utask;
+
+	if (!current->mm || !utask || !utask->active_uprobe)
+		/* task is currently not uprobed */

kernel/events/uprobes.c int uprobe_mmap(struct vm_area_struct *vma) · line 1281

Quoted 10 lines
kernel/events/uprobes.c +10 -0
+		return 0;
+
+	utask->state = UTASK_SSTEP_ACK;
+	set_thread_flag(TIF_UPROBE);
+	return 1;
+}
+
+struct notifier_block uprobe_exception_nb = {
+	.notifier_call = arch_uprobe_exception_notify,
+	.priority = INT_MAX - 1,	/* notified after kprobes, kgdb */

This too should align vertically.

Quoted 37 lines
kernel/events/uprobes.c +4 -1
+};
+
 static int __init init_uprobes(void)
 {
 	int i;
@@ -1015,7 +1289,8 @@ static int __init init_uprobes(void)
 		mutex_init(&uprobes_mutex[i]);
 		mutex_init(&uprobes_mmap_mutex[i]);
 	}
-	return 0;
+	init_srcu_struct(&uprobes_srcu);
+	return register_die_notifier(&uprobe_exception_nb);
 }
 
 static void __exit exit_uprobes(void)
kernel/fork.c +3 -0
index b77fd55..df711bb 100644
@@ -66,6 +66,7 @@
 #include <linux/user-return-notifier.h>
 #include <linux/oom.h>
 #include <linux/khugepaged.h>
+#include <linux/uprobes.h>
 
 #include <asm/pgtable.h>
 #include <asm/pgalloc.h>
@@ -700,6 +701,8 @@ void mm_release(struct task_struct *tsk, struct mm_struct *mm)
 		exit_pi_state_list(tsk);
 #endif
 
+	uprobe_free_utask(tsk);
+
 	/* Get rid of any cached register state */
 	deactivate_mm(tsk, mm);
 

kernel/fork.c static struct task_struct *copy_process(unsigned long clone_flags, · line 1301

Quoted 7 lines
kernel/fork.c +4 -0
@@ -1292,6 +1295,10 @@ static struct task_struct *copy_process(unsigned long clone_flags,
 	INIT_LIST_HEAD(&p->pi_state_list);
 	p->pi_state_cache = NULL;
 #endif
+#ifdef CONFIG_UPROBES
+	p->utask = NULL;
+	p->uprobes_srcu_id = -1;
+#endif

Should be a uprobe_copy_process() callback.

Quoted 26 lines
kernel/fork.c +0 -0
 	/*
 	 * sigaltstack should be cleared when sharing the same VM
 	 */
kernel/signal.c +4 -0
index 8511e39..24a9c39 100644
@@ -29,6 +29,7 @@
 #include <linux/pid_namespace.h>
 #include <linux/nsproxy.h>
 #include <linux/user_namespace.h>
+#include <linux/uprobes.h>
 #define CREATE_TRACE_POINTS
 #include <trace/events/signal.h>
 
@@ -2192,6 +2193,9 @@ int get_signal_to_deliver(siginfo_t *info, struct k_sigaction *return_ka,
 	struct signal_struct *signal = current->signal;
 	int signr;
 
+	if (unlikely(uprobe_deny_signal()))
+		return 0;
+
 relock:
 	/*
 	 * We'll jump back here after any time we were stopped in TASK_STOPPED.

kernel/signal.c int do_sigaction(int sig, struct k_sigaction *act, struct k_sigaction *oact) · line 3014

Quoted 8 lines
kernel/signal.c +1 -1
@@ -3004,7 +3008,7 @@ int do_sigaction(int sig, struct k_sigaction *act, struct k_sigaction *oact)
 	return 0;
 }
 
-int 
+int
 do_sigaltstack (const stack_t __user *uss, stack_t __user *uoss, unsigned long sp)
 {
 	stack_t oss;

So ... this patch needs to split up into the preparatory patches
+ main patch and it all needs to be reworked.

Thanks,

Ingo

#4 lore raw
Quoted 9 lines
Quoted 5 lines

Where possible, we check and skip singlestepping the breakpointed instructions. For now we skip single byte as well as few multibyte nop instructions. However this can be extended to other instructions too.

Is this an optimization - allowing a NOP to be inserted for easy probe insertion?

Yes, Its an optimization by which we avoid singlestep exception.

Quoted 13 lines

arch/x86/include/asm/uprobes.h struct arch_uprobe { · line 45

Quoted 9 lines

#define MAX_UINSN_BYTES 16

arch/x86/include/asm/uprobes.h +2 -1
@@ -39,5 +41,17 @@ struct arch_uprobe {
 #endif
 };
 
-extern int arch_uprobes_analyze_insn(struct mm_struct *mm, struct arch_uprobe *arch_uprobe);
+struct arch_uprobe_task {
+	unsigned long saved_trap_no;

trap_no in struct thread_struct and now in arch_uprobe_task is a misnomer - I always expect a trap_yes flag next to it ;-)

Quoted 3 lines

Please create a separate patch in front of this patch that trivially renames trap_no to something saner like trap_nr, and introduce this field as trap_nr as well.

Okay, Will do.

Quoted 32 lines

arch/x86/include/asm/uprobes.h struct arch_uprobe { · line 56

Quoted 12 lines
arch/x86/include/asm/uprobes.h +11 -0
+#ifdef CONFIG_X86_64
+	unsigned long saved_scratch_register;
+#endif
+};
+
+extern int arch_uprobe_analyze_insn(struct mm_struct *mm, struct arch_uprobe *arch_uprobe);
+extern int arch_uprobe_pre_xol(struct arch_uprobe *auprobe, struct pt_regs *regs);
+extern int arch_uprobe_post_xol(struct arch_uprobe *auprobe, struct pt_regs *regs);
+extern bool arch_uprobe_xol_was_trapped(struct task_struct *tsk);
+extern int arch_uprobe_exception_notify(struct notifier_block *self, unsigned long val, void *data);
+extern void arch_uprobe_abort_xol(struct pt_regs *regs, struct arch_uprobe *auprobe);

Small API parameter consistency nit:

We tend to order function parameters by importance: the more important parameter comes first.

In that sense arch_uprobe_pre_xol() and arch_uprobe_abort_xol() have obviously inconsistent ordering: the first one is (aup, regs), the second one is (regs, aup).

Please make it consistent, at first glance the right one appears to be:

aup, mm
aup, regs
aup, regs
t
self, val, data
aup, regs

Quoted 2 lines

I'd put the arch_uprobe_analyze_insn() order change into a separate patch, preceding this patch.

Okay, Will do.

Quoted 18 lines

arch/x86/kernel/uprobes.c line 36

Quoted 9 lines
arch/x86/kernel/uprobes.c +0 -0
 #include <linux/kdebug.h>

#include <asm/insn.h>

arch/x86/kernel/uprobes.c +5 -0
+#ifdef CONFIG_X86_32
+#define is_32bit_app(tsk) 1
+#else
+#define is_32bit_app(tsk) (test_tsk_thread_flag(tsk, TIF_IA32))
+#endif

Small detail, we prefer to use this variant:

#ifdef X

# define foo()
#else
# define bar()
#endif
Quoted 6 lines

to give the construct more visual structure.

Also, please put it into asm/compat.h and use it at other places within arch/x86/ as well, there's half a dozen similar patterns of TIP_IA32 tests in arch/x86/. Please make this a separate patch, preceding this patch.

Okay.

Quoted 3 lines

Also, how does this interact with the new x32 ABI code in -tip?

I havent taken a look at X86_X32_ABI. I will comeback to you on this.

Quoted 7 lines

arch/x86/kernel/uprobes.c static int validate_insn_bits(struct mm_struct *mm, struct arch_uprobe *auprobe, · line 414

Quoted 3 lines

*/

arch/x86/kernel/uprobes.c +1 -1
-int arch_uprobes_analyze_insn(struct mm_struct *mm, struct arch_uprobe *auprobe)
+int arch_uprobe_analyze_insn(struct mm_struct *mm, struct arch_uprobe *auprobe)

Btw., there's a few more 'uprobes' naming leftovers in the code, like UPROBES_COPY_INSN.

Okay,

Quoted 12 lines

arch/x86/kernel/uprobes.c int arch_uprobes_analyze_insn(struct mm_struct *mm, struct arch_uprobe *auprobe) · line 430

Quoted 9 lines

{

arch/x86/kernel/uprobes.c +2 -0
 	int ret;
 	struct insn insn;
@@ -420,3 +426,260 @@ int arch_uprobes_analyze_insn(struct mm_struct *mm, struct arch_uprobe *auprobe)
 
 	return 0;
 }
+
+#define	UPROBE_TRAP_NO		UINT_MAX

this would be TRAP_NR too then.

Quoted 3 lines

Also, please put such defines at the top of the .c file, so that it's easily in view.

Okay,

Quoted 8 lines

arch/x86/kernel/uprobes.c int arch_uprobes_analyze_insn(struct mm_struct *mm, struct arch_uprobe *auprobe) · line 451

Quoted 3 lines
arch/x86/kernel/uprobes.c +3 -0
+#else
+static void
+pre_xol_rip_insn(struct arch_uprobe *auprobe, struct pt_regs *regs, struct arch_uprobe_task *tskinfo)

If the standard variable naming for 'struct arch_uprobe' is 'auprobe', then the consistent one for 'struct arch_uprobe_task' would be something like 'autask', not 'tskinfo'.

Quoted 6 lines

'autask' would be a nice companion to the 'utask' name as well.

Please propagate the new naming through all other uses of struct arch_uprobe_task as well, and through related local variables and structure fields.

Okay.

Quoted 8 lines

arch/x86/kernel/uprobes.c int arch_uprobes_analyze_insn(struct mm_struct *mm, struct arch_uprobe *auprobe) · line 614

Quoted 5 lines
arch/x86/kernel/uprobes.c +5 -0
+	switch (val) {
+	case DIE_INT3:
+		/* Run your handler here */
+		if (uprobe_bkpt_notifier(regs))
+			ret = NOTIFY_STOP;

This comment looks somewhat out of place.

Quoted 4 lines

Also, I have not noticed this in the first patch, but 'bkpt' is not a standard way to refer to breakpoints: we either use 'breakpoint' or 'bp'.

This is again one of those things that I changed from bp to bkpt based on LKML feedback. I am okay to go back to bp.

Quoted 9 lines

arch/x86/kernel/uprobes.c int arch_uprobes_analyze_insn(struct mm_struct *mm, struct arch_uprobe *auprobe) · line 620

Quoted 6 lines
arch/x86/kernel/uprobes.c +6 -0
+
+		break;
+
+	case DIE_DEBUG:
+		if (uprobe_post_notifier(regs))
+			ret = NOTIFY_STOP;

This wants to be post_bp_notifier, right?

Quoted 7 lines

I'd suggest to name them in a symmetric way:

   uprobe_pre_bp_notifier();
   uprobe_post_bp_notifier();

so that the connection and ordering is obvious in the higher levels as well.

Okay.

Quoted 16 lines

arch/x86/kernel/uprobes.c int arch_uprobes_analyze_insn(struct mm_struct *mm, struct arch_uprobe *auprobe) · line 631

Quoted 12 lines
arch/x86/kernel/uprobes.c +11 -0
+
+	default:
+		break;
+	}
+
+	return ret;
+}
+
+/*
+ * xol insn either trapped or thread has a fatal signal, so reset the
+ * instruction pointer to its probed address.

Verb tense mismatch at the beginning of the sentence. It is also very friendly to do round, explanatory sentences like:

Quoted 6 lines

'This function gets called when the XOL instruction either
trapped or the thread had a fatal signal, ...'

Please review all other comments for similar patterns as well and fix them.

Okay.

Quoted 8 lines
Quoted 8 lines
arch/x86/kernel/uprobes.c +8 -0
+bool arch_uprobe_skip_sstep(struct pt_regs *regs, struct arch_uprobe *auprobe)
+{
+	int i;
+
+	for (i = 0; i < MAX_UINSN_BYTES; i++) {
+		if ((auprobe->insn[i] == 0x66))
+			continue;
+
Quoted 9 lines

arch/x86/kernel/uprobes.c int arch_uprobes_analyze_insn(struct mm_struct *mm, struct arch_uprobe *auprobe) · line 671

Quoted 5 lines
arch/x86/kernel/uprobes.c +5 -0
+		if (auprobe->insn[i] == 0x90)
+			return true;
+
+		if (i == (MAX_UINSN_BYTES - 1))
+			break;

Looks like the loop could run from 0 to MAX_UINSN_BYTES-2 and then this break would be superfluous.

Even if we were to run from 0 to MAX_UINSN_BYTES - 2, we would have to add extra code to handle 0x66* 0x90 (where 0x90 is stored at index i == MAX_UINSN_BYTES - 1. So I would like to keep this code as is.

Quoted 17 lines

arch/x86/kernel/uprobes.c int arch_uprobes_analyze_insn(struct mm_struct *mm, struct arch_uprobe *auprobe) · line 684

Quoted 13 lines
arch/x86/kernel/uprobes.c +13 -0
+
+		if ((auprobe->insn[i] == 0x0f) && (auprobe->insn[i+1] == 0x1f))
+			return true;
+
+		if ((auprobe->insn[i] == 0x0f) && (auprobe->insn[i+1] == 0x19))
+			return true;
+
+		if ((auprobe->insn[i] == 0x87) && (auprobe->insn[i+1] == 0xc0))
+			return true;
+
+		break;
+	}
+	return false;

Looks like we will skip not just the instructions listed above, but a lot more patterns:

Quoted 5 lines

0x66* { 0x90 | 0x0f 0x1f | 0x0f 0x19 | 0x87 0xc0 }

It might make sense to mention the pattern in the comment and mention that it's intended to skip the instructions listed, under the currently known x86 ISA.

Okay, Will do.

Quoted 12 lines

include/linux/uprobes.h struct uprobe_consumer { · line 70

Quoted 8 lines
include/linux/uprobes.h +7 -0
+
+/*
+ * uprobe_task: Metadata of a task while it singlesteps.
+ */
+struct uprobe_task {
+	unsigned long xol_vaddr;
+	unsigned long vaddr;

These two fields are never actually used outside of architecture code.

Quoted 7 lines

Unless there's a good reason to keep them outside I'd suggest to move them into struct arch_uprobe_task. This has another
benefit: we can pass struct arch_uprobe_task to the architecture methods, instead of struct uprobe_task. This would allow the moving of the struct uprobe_task into uprobes.c - no code outside uprobes.c needs to know its structure.

The Xol layer(which is the next patch) uses them in arch agnostic way. Also vaddr/xol_vaddr are populated/used in arch agnostic way. We could still move them to arch_uprobe_task but we will then have to ensure that every other arch defines them the way uprobes understands.

Quoted 10 lines

include/linux/uprobes.h struct uprobe_consumer { · line 76

Quoted 6 lines
include/linux/uprobes.h +6 -0
+
+	enum uprobe_task_state state;
+	struct arch_uprobe_task tskinfo;
+
+	struct uprobe *active_uprobe;
+};

Also, please check the style of 'struct uprobe' and make 'struct uprobe_task' match that style.

Okay.

Quoted 9 lines

include/linux/uprobes.h static inline int uprobe_mmap(struct vm_area_struct *vma) · line 115

Quoted 5 lines
include/linux/uprobes.h +5 -0
+}
+static inline unsigned long get_uprobe_bkpt_addr(struct pt_regs *regs)
+{
+	return 0;
+}

Please use the standard uprobe method naming pattern for get_uprobe_bkpt_addr().

do you mean uprobe_get_bp_addr ?

Quoted 32 lines
Quoted 6 lines
kernel/events/uprobes.c +4 -0
+static void handler_chain(struct uprobe *uprobe, struct pt_regs *regs)
+{
+	struct uprobe_consumer *consumer;
+

kernel/events/uprobes.c static struct uprobe *alloc_uprobe(struct inode *inode, loff_t offset) · line 544

Quoted 10 lines
kernel/events/uprobes.c +10 -0
+	if (!(uprobe->flags & UPROBES_RUN_HANDLER))
+		return;
+
+	down_read(&uprobe->consumer_rwsem);
+	consumer = uprobe->consumers;
+	for (consumer = uprobe->consumers; consumer; consumer = consumer->next) {
+		if (!consumer->filter || consumer->filter(consumer, current))
+			consumer->handler(consumer, regs);
+	}
+	up_read(&uprobe->consumer_rwsem);

Suggestion: as a local variable 'consumer' is mentioned many, many times throughout the code. It might make sense to standardize, for the strict case of local variables (not structure fields), on the following pattern:

	struct uprobe_consumer *uc;

	...

	uc = uprobe->consumers;
	for (uc = uprobe->consumers; uc; uc = uc->next) {
		if (!uc->filter || uc->filter(uc, current))
			uc->handler(uc, regs);
	}
Quoted 7 lines

See how much more compact and readable it is? This shorter form also makes it immediately obvious that the first line before the loop is superfulous, it's already done by the loop initializer.

Please propagate this new naming through other places within uprobes.c as well.

Okay.

Quoted 17 lines

kernel/events/uprobes.c static void remove_breakpoint(struct mm_struct *mm, struct uprobe *uprobe, loff_ · line 691

Quoted 12 lines
kernel/events/uprobes.c +11 -0
+}
+
 
+/*
+ * There could be threads that have hit the breakpoint and are entering the
+ * notifier code and trying to acquire the uprobes_treelock. The thread
+ * calling delete_uprobe() that is removing the uprobe from the rb_tree can
+ * race with these threads and might acquire the uprobes_treelock compared
+ * to some of the breakpoint hit threads. In such a case, the breakpoint hit
+ * threads will not find the uprobe. Hence wait till the current breakpoint
+ * hit threads acquire the uprobes_treelock before the uprobe is removed
+ * from the rbtree.

Hm, the last sentence does not parse for me. (even if it's correct English it might make sense to rephrase it to be clearer what is meant.)

Would this be okay with you.

The current unregistering thread waits till all other threads that have hit a breakpoint to acquire the uprobes_treelock before the uprobe is removed from the rbtree.

Quoted 40 lines
Quoted 23 lines
kernel/events/uprobes.c +23 -0
+/*
+ * If we are singlestepping, then ensure this thread is not connected to
+ * non-fatal signals until completion of singlestep.  When xol insn itself
+ * triggers the signal,  restart the original insn even if the task is
+ * already SIGKILL'ed (since coredump should report the correct ip).  This
+ * is even more important if the task has a handler for SIGSEGV/etc, The
+ * _same_ instruction should be repeated again after return from the signal
+ * handler, and SSTEP can never finish in this case.
+ */
+bool uprobe_deny_signal(void)
+{
+	struct task_struct *tsk = current;
+	struct uprobe_task *utask = tsk->utask;
+
+	if (likely(!utask || !utask->active_uprobe))
+		return false;
+
+	WARN_ON_ONCE(utask->state != UTASK_SSTEP);
+
+	if (signal_pending(tsk)) {
+		spin_lock_irq(&tsk->sighand->siglock);
+		clear_tsk_thread_flag(tsk, TIF_SIGPENDING);
+		spin_unlock_irq(&tsk->sighand->siglock);

kernel/events/uprobes.c int uprobe_mmap(struct vm_area_struct *vma) · line 1133

Quoted 10 lines
kernel/events/uprobes.c +10 -0
+
+		if (__fatal_signal_pending(tsk) || arch_uprobe_xol_was_trapped(tsk)) {
+			utask->state = UTASK_SSTEP_TRAPPED;
+			set_tsk_thread_flag(tsk, TIF_UPROBE);
+			set_tsk_thread_flag(tsk, TIF_NOTIFY_RESUME);
+		}
+	}
+
+	return true;
+}

Hm, seems racy to me: what happens if we get a signal shortly *after* this code got called and before we take the siglock? It might be safe, but it's not clearly obvious why and the code does not explain it.

Quoted 5 lines

Why is this logic not within the 'relock' loop within get_signal_to_deliver(), called with the siglock already taken? That would avoid races and would also remove the siglock taking/dropping above.

At the time the thread is at get_signal_to_deliver, the thread has either hit a uprobe and still not singlestepped or it isnt handling any uprobes. If the thread has hit a uprobe but not yet singlestepped, we want the thread to not handle signals, Hence it returns from the signal handling function. If a new signal were to be delivered, after we return from get_signal_to_deliver, then we repeat the same, i.e return after checking that uprobes is active.

If the thread was not in the middle of a uprobe hit then we go through the regular signal handling.

Since there is no way this thread can hit a uprobe once a thread has entered get_signal_to_deliver(kernel code), I dont see a reason to move it under relock:

Quoted 40 lines

kernel/events/uprobes.c int uprobe_mmap(struct vm_area_struct *vma) · line 1166

Quoted 8 lines

+ * All non-fatal signals cannot interrupt thread while the thread singlesteps.
+ */

kernel/events/uprobes.c +6 -0
+void uprobe_notify_resume(struct pt_regs *regs)
+{
+	struct vm_area_struct *vma;
+	struct uprobe_task *utask;
+	struct mm_struct *mm;
+	struct uprobe *u;

'u' is not what we use for a uprobe in other places - please use consistent naming all across the code!

kernel/events/uprobes.c int uprobe_mmap(struct vm_area_struct *vma) · line 1167

Quoted 1 line
kernel/events/uprobes.c +1 -0
+	unsigned long probept;

this wants to be the breakpoint virtual address, right?

If yes then 'bp_vaddr' would be a *much* clearer name for it.

kernel/events/uprobes.c int uprobe_mmap(struct vm_area_struct *vma) · line 1172

Quoted 5 lines
kernel/events/uprobes.c +5 -0
+
+	utask = current->utask;
+	u = NULL;
+	mm = current->mm;
+	if (!utask || utask->state == UTASK_BP_HIT) {

Please put these two starkly different pieces of functionality on the two sides of the branch into two helper inline functions, named apporpriately.

kernel/events/uprobes.c int uprobe_mmap(struct vm_area_struct *vma) · line 1179

Quoted 7 lines
kernel/events/uprobes.c +7 -0
+		probept = get_uprobe_bkpt_addr(regs);
+		down_read(&mm->mmap_sem);
+		vma = find_vma(mm, probept);
+
+		if (vma && vma->vm_start <= probept && valid_vma(vma, false))
+			u = find_uprobe(vma->vm_file->f_mapping->host, probept - vma->vm_start +
+					(vma->vm_pgoff << PAGE_SHIFT));

I'd stick the offset into a loff_t local variable first, plus use an inode variable as well, then it would all be so much easier to read:

Quoted 6 lines
			inode = vma->vm_file->f_mapping->host;
			offset = vma->vm_pgoff << PAGE_SHIFT;
			offset += bp_vaddr - vma->vm_start;

			uprobe = find_uprobe(inode, offset);

Okay,

Quoted 69 lines
Quoted 4 lines
kernel/events/uprobes.c +4 -0
+
+		srcu_read_unlock_raw(&uprobes_srcu, current->uprobes_srcu_id);
+		current->uprobes_srcu_id = -1;
+		up_read(&mm->mmap_sem);

kernel/events/uprobes.c int uprobe_mmap(struct vm_area_struct *vma) · line 1193

Quoted 10 lines
kernel/events/uprobes.c +10 -0
+
+		if (!u)
+			/* No matching uprobe; signal SIGTRAP. */
+			goto cleanup_ret;
+
+		if (!utask) {
+			utask = add_utask();
+			/* Cannot Allocate; re-execute the instruction. */
+			if (!utask)
+				goto cleanup_ret;

Weird capitalization.

kernel/events/uprobes.c int uprobe_mmap(struct vm_area_struct *vma) · line 1205

Quoted 12 lines
kernel/events/uprobes.c +12 -0
+		}
+		utask->active_uprobe = u;
+		handler_chain(u, regs);
+		if (u->flags & UPROBES_SKIP_SSTEP && uprobe_skip_sstep(regs, u))
+			goto cleanup_ret;
+
+		utask->state = UTASK_SSTEP;
+		if (!pre_ssout(u, regs, probept))
+			user_enable_single_step(current);
+		else
+			/* Cannot Singlestep; re-execute the instruction. */
+			goto cleanup_ret;

Weird capitalization.

Quoted 24 lines
kernel/events/uprobes.c +24 -0
+
+	} else {
+		u = utask->active_uprobe;
+		if (utask->state == UTASK_SSTEP_ACK)
+			arch_uprobe_post_xol(&u->arch, regs);
+		else if (utask->state == UTASK_SSTEP_TRAPPED)
+			arch_uprobe_abort_xol(regs, &u->arch);
+		else
+			WARN_ON_ONCE(1);
+
+		put_uprobe(u);
+		utask->active_uprobe = NULL;
+		utask->state = UTASK_RUNNING;
+		user_disable_single_step(current);
+
+		spin_lock_irq(&current->sighand->siglock);
+		recalc_sigpending(); /* see uprobe_deny_signal() */
+		spin_unlock_irq(&current->sighand->siglock);
+	}
+	return;
+
+cleanup_ret:
+	if (utask) {
+		utask->active_uprobe = NULL;

kernel/events/uprobes.c int uprobe_mmap(struct vm_area_struct *vma) · line 1239

Quoted 10 lines
kernel/events/uprobes.c +10 -0
+		utask->state = UTASK_RUNNING;
+	}
+	if (u) {
+		if (!(u->flags & UPROBES_SKIP_SSTEP))
+			instruction_pointer_set(regs, probept);
+
+		put_uprobe(u);
+	} else {
+		send_sig(SIGTRAP, current, 0);
+	}

Please use a standard cleanup sequence.

Quoted 3 lines

The code flow is extremely mixed throughout uprobe_notify_resume(), please clean it all up and make it much more obvious.

Okay.

Quoted 37 lines
Quoted 37 lines
kernel/events/uprobes.c +36 -0
+}
+
+/*
+ * uprobe_bkpt_notifier gets called from interrupt context as part of
+ * notifier mechanism. Set TIF_UPROBE flag and indicate breakpoint hit.
+ */
+int uprobe_bkpt_notifier(struct pt_regs *regs)
+{
+	struct uprobe_task *utask;
+
+	if (!current->mm)
+		return 0;
+
+	utask = current->utask;
+	if (utask)
+		utask->state = UTASK_BP_HIT;
+
+	set_thread_flag(TIF_UPROBE);
+	current->uprobes_srcu_id = srcu_read_lock_raw(&uprobes_srcu);
+	return 1;
+}
+
+/*
+ * uprobe_post_notifier gets called in interrupt context as part of notifier
+ * mechanism. Set TIF_UPROBE flag and indicate completion of singlestep.
+ */
+int uprobe_post_notifier(struct pt_regs *regs)
+{
+	struct uprobe_task *utask = current->utask;
+
+	if (!current->mm || !utask || !utask->active_uprobe)
+		/* task is currently not uprobed */
+		return 0;
+
+	utask->state = UTASK_SSTEP_ACK;
+	set_thread_flag(TIF_UPROBE);
Quoted 9 lines

kernel/events/uprobes.c int uprobe_mmap(struct vm_area_struct *vma) · line 1281

Quoted 6 lines
kernel/events/uprobes.c +6 -0
+	return 1;
+}
+
+struct notifier_block uprobe_exception_nb = {
+	.notifier_call = arch_uprobe_exception_notify,
+	.priority = INT_MAX - 1,	/* notified after kprobes, kgdb */

This too should align vertically.

Okay,

Quoted 11 lines

kernel/fork.c static struct task_struct *copy_process(unsigned long clone_flags, · line 1301

Quoted 8 lines
kernel/fork.c +4 -0
@@ -1292,6 +1295,10 @@ static struct task_struct *copy_process(unsigned long clone_flags,
 	INIT_LIST_HEAD(&p->pi_state_list);
 	p->pi_state_cache = NULL;
 #endif
+#ifdef CONFIG_UPROBES
+	p->utask = NULL;
+	p->uprobes_srcu_id = -1;
+#endif

Should be a uprobe_copy_process() callback.

Okay.

Quoted 3 lines

So ... this patch needs to split up into the preparatory patches
+ main patch and it all needs to be reworked.

Okay.

Signature · 2 lines
Thanks and Regards
Srikar
#5 lore raw
Quoted 10 lines · Srikar Dronamraju
Quoted 7 lines
Quoted 4 lines

Where possible, we check and skip singlestepping the breakpointed instructions. For now we skip single byte as well as few multibyte nop instructions. However this can be extended to other instructions too.

Is this an optimization - allowing a NOP to be inserted for easy probe insertion?

Yes, Its an optimization by which we avoid singlestep exception.

That would be nice to comment in the code - nowhere in the 'skip' logic is this fact mentioned, and it's really useful information to pretty much anyone reading the code.

It's also a nice optimization, there's no need to obfuscate its existence.

Quoted 7 lines
Quoted 7 lines

arch/x86/kernel/uprobes.c int arch_uprobes_analyze_insn(struct mm_struct *mm, struct arch_uprobe *auprobe) · line 614

Quoted 4 lines
arch/x86/kernel/uprobes.c +4 -0
+	case DIE_INT3:
+		/* Run your handler here */
+		if (uprobe_bkpt_notifier(regs))
+			ret = NOTIFY_STOP;

This comment looks somewhat out of place.

Quoted 6 lines
Quoted 3 lines

Also, I have not noticed this in the first patch, but 'bkpt' is not a standard way to refer to breakpoints: we either use 'breakpoint' or 'bp'.

This is again one of those things that I changed from bp to bkpt based on LKML feedback. I am okay to go back to bp.

:-/ I can understand it somewhat, 'bp' also means other things.

'hwbp' is a common name - you could use 'swbp' which would pair with that nicely?

Quoted 18 lines
Quoted 8 lines
Quoted 8 lines
arch/x86/kernel/uprobes.c +8 -0
+bool arch_uprobe_skip_sstep(struct pt_regs *regs, struct arch_uprobe *auprobe)
+{
+	int i;
+
+	for (i = 0; i < MAX_UINSN_BYTES; i++) {
+		if ((auprobe->insn[i] == 0x66))
+			continue;
+
Quoted 9 lines

arch/x86/kernel/uprobes.c int arch_uprobes_analyze_insn(struct mm_struct *mm, struct arch_uprobe *auprobe) · line 671

Quoted 5 lines
arch/x86/kernel/uprobes.c +5 -0
+		if (auprobe->insn[i] == 0x90)
+			return true;
+
+		if (i == (MAX_UINSN_BYTES - 1))
+			break;

Looks like the loop could run from 0 to MAX_UINSN_BYTES-2 and then this break would be superfluous.

Quoted 4 lines

Even if we were to run from 0 to MAX_UINSN_BYTES - 2, we would have to add extra code to handle 0x66* 0x90 (where 0x90 is stored at index i == MAX_UINSN_BYTES - 1. So I would like to keep this code as is.

Ok.

Quoted 18 lines
Quoted 10 lines

include/linux/uprobes.h struct uprobe_consumer { · line 70

Quoted 6 lines
include/linux/uprobes.h +6 -0
+/*
+ * uprobe_task: Metadata of a task while it singlesteps.
+ */
+struct uprobe_task {
+	unsigned long xol_vaddr;
+	unsigned long vaddr;

These two fields are never actually used outside of architecture code.

Quoted 7 lines

Unless there's a good reason to keep them outside I'd suggest to move them into struct arch_uprobe_task. This has another benefit: we can pass struct arch_uprobe_task to the architecture methods, instead of struct uprobe_task. This would allow the moving of the struct uprobe_task into uprobes.c - no code outside uprobes.c needs to know its structure.

Quoted 5 lines

The Xol layer(which is the next patch) uses them in arch agnostic way. Also vaddr/xol_vaddr are populated/used in arch agnostic way. We could still move them to arch_uprobe_task but we will then have to ensure that every other arch defines them the way uprobes understands.

Correct - and that still isolates the arch code from the core uprobes code.

We could also introduce 'struct generic_arch_uprobe_task' and embedd that inside arch_uprobe via a short field name, to make it easy to access: ->gen.field or so.

You can also leave it as-is for now, I'll reconsider how things look like with the patch following these bits and then make a new suggestion if I see a better way.

Quoted 9 lines
Quoted 7 lines

include/linux/uprobes.h static inline int uprobe_mmap(struct vm_area_struct *vma) · line 115

Quoted 4 lines
include/linux/uprobes.h +4 -0
+static inline unsigned long get_uprobe_bkpt_addr(struct pt_regs *regs)
+{
+	return 0;
+}

Please use the standard uprobe method naming pattern for get_uprobe_bkpt_addr().

do you mean uprobe_get_bp_addr ?

Yeah, that sounds good.

Quoted 15 lines
Quoted 10 lines
Quoted 9 lines

+/*
+ * There could be threads that have hit the breakpoint and are entering the
+ * notifier code and trying to acquire the uprobes_treelock. The thread
+ * calling delete_uprobe() that is removing the uprobe from the rb_tree can
+ * race with these threads and might acquire the uprobes_treelock compared
+ * to some of the breakpoint hit threads. In such a case, the breakpoint hit
+ * threads will not find the uprobe. Hence wait till the current breakpoint
+ * hit threads acquire the uprobes_treelock before the uprobe is removed
+ * from the rbtree.

Quoted 4 lines

Hm, the last sentence does not parse for me. (even if it's correct English it might make sense to rephrase it to be clearer what is meant.)

Quoted 5 lines

Would this be okay with you.

The current unregistering thread waits till all other threads that have hit a breakpoint to acquire the uprobes_treelock before the uprobe is removed from the rbtree.

s/is removed/are removed

?

If yes then indeed this reads better.

Quoted 8 lines

[...]

If the thread was not in the middle of a uprobe hit then we go through the regular signal handling.

Since there is no way this thread can hit a uprobe once a thread has entered get_signal_to_deliver(kernel code), I dont see a reason to move it under relock:

Ok, fair enough.

Thanks,

Ingo

#6 lore raw
Quoted 14 lines · Ingo Molnar
Quoted 10 lines · Srikar Dronamraju
Quoted 7 lines
Quoted 4 lines

Where possible, we check and skip singlestepping the breakpointed instructions. For now we skip single byte as well as few multibyte nop instructions. However this can be extended to other instructions too.

Is this an optimization - allowing a NOP to be inserted for easy probe insertion?

Yes, Its an optimization by which we avoid singlestep exception.

Quoted 6 lines

That would be nice to comment in the code - nowhere in the 'skip' logic is this fact mentioned, and it's really useful information to pretty much anyone reading the code.

It's also a nice optimization, there's no need to obfuscate its existence.

okay, Will add.

Quoted 12 lines
Quoted 12 lines
Quoted 11 lines

arch/x86/kernel/uprobes.c int arch_uprobes_analyze_insn(struct mm_struct *mm, struct arch_uprobe *auprobe) · line 614

Quoted 5 lines
arch/x86/kernel/uprobes.c +4 -0
+	case DIE_INT3:
+		/* Run your handler here */
+		if (uprobe_bkpt_notifier(regs))
+			ret = NOTIFY_STOP;

This comment looks somewhat out of place.

Also, I have not noticed this in the first patch, but 'bkpt' is not a standard way to refer to breakpoints: we either use 'breakpoint' or 'bp'.

Quoted 8 lines
Quoted 2 lines

This is again one of those things that I changed from bp to bkpt based on LKML feedback. I am okay to go back to bp.

:-/ I can understand it somewhat, 'bp' also means other things.

'hwbp' is a common name - you could use 'swbp' which would pair with that nicely?

Okay.
However most of these functions call are called from within uprobes.c and have a uprobe prefix. So there is enough context for people to link bp to breakpoint.

Quoted 4 lines

Correct - and that still isolates the arch code from the core uprobes code.

Quoted 8 lines

We could also introduce 'struct generic_arch_uprobe_task' and embedd that inside arch_uprobe via a short field name, to make it easy to access: ->gen.field or so.

You can also leave it as-is for now, I'll reconsider how things look like with the patch following these bits and then make a new suggestion if I see a better way.

Will leave this as-is for now and wait for your suggestions.

Quoted 18 lines
Quoted 18 lines
Quoted 11 lines
Quoted 10 lines

+/*
+ * There could be threads that have hit the breakpoint and are entering the
+ * notifier code and trying to acquire the uprobes_treelock. The thread
+ * calling delete_uprobe() that is removing the uprobe from the rb_tree can
+ * race with these threads and might acquire the uprobes_treelock compared
+ * to some of the breakpoint hit threads. In such a case, the breakpoint hit
+ * threads will not find the uprobe. Hence wait till the current breakpoint
+ * hit threads acquire the uprobes_treelock before the uprobe is removed
+ * from the rbtree.

Quoted 4 lines

Hm, the last sentence does not parse for me. (even if it's correct English it might make sense to rephrase it to be clearer what is meant.)

Would this be okay with you.

Quoted 8 lines
Quoted 3 lines

The current unregistering thread waits till all other threads that have hit a breakpoint to acquire the uprobes_treelock before the uprobe is removed from the rbtree.

s/is removed/are removed

?

At a time, we are unregistering just one probe,(atleast for now.) Wondering if "before uprobes are remove from rbtree." sounds as if more than one uprobe is being removed at one instance.

Quoted 4 lines

If yes then indeed this reads better.

Quoted 2 lines

[...]

Quoted 9 lines
Quoted 6 lines

If the thread was not in the middle of a uprobe hit then we go through the regular signal handling.

Since there is no way this thread can hit a uprobe once a thread has entered get_signal_to_deliver(kernel code), I dont see a reason to move it under relock:

Ok, fair enough.

Okay,

Signature · 2 lines
Thanks and Regards
Srikar
#7 lore raw
Quoted 7 lines · Srikar Dronamraju
Quoted 7 lines
Quoted 4 lines

The current unregistering thread waits till all other threads that have hit a breakpoint to acquire the uprobes_treelock before the uprobe is removed from the rbtree.

s/is removed/are removed

Quoted 7 lines
Quoted 2 lines

?

At a time, we are unregistering just one probe,(atleast for now.) Wondering if "before uprobes are remove from rbtree." sounds as if more than one uprobe is being removed at one instance.

The sentence still does not parse:

" X waits until Ys that have done Z before A is removed from
the rbtree"

Did you want to say:

" The current unregistering thread waits till all other
threads have hit a breakpoint, to acquire the uprobes_treelock before the uprobe is removed from the rbtree. "

(I have removed 'that' and added a comma.)

Thanks,

Ingo

#8 lore raw
In reply to Ingo Molnar
Quoted 18 lines

Review

Quoted 9 lines

#include <linux/kdebug.h>
#include <asm/insn.h>

(quoted diff) +5 -0
+#ifdef CONFIG_X86_32
+#define is_32bit_app(tsk) 1
+#else
+#define is_32bit_app(tsk) (test_tsk_thread_flag(tsk, TIF_IA32))
+#endif

Small detail, we prefer to use this variant:

#ifdef X

# define foo()
#else
# define bar()
#endif
Quoted 6 lines

to give the construct more visual structure.

Also, please put it into asm/compat.h and use it at other places within arch/x86/ as well, there's half a dozen similar patterns of TIP_IA32 tests in arch/x86/. Please make this a separate patch, preceding this patch.

I am facing a problem doing the above.
I did some changes in arch/x86/kernel/signal.c and included compat.h and compilation fails on tip tree that the commit d1a797f388

In file included from ../arch/x86/kernel/signal.c:42:
../arch/x86/include/asm/compat.h: In function ‘arch_compat_alloc_user_space’:
../arch/x86/include/asm/compat.h:238: error: ‘old_rsp’ undeclared (first use in this function) ../arch/x86/include/asm/compat.h:238: error: (Each undeclared identifier is reported only once ../arch/x86/include/asm/compat.h:238: error: for each function it appears in.) ../arch/x86/include/asm/compat.h:238: warning: type defaults to ‘int’ in declaration of ‘pfo_ret__’

I see old_rsp declared only under CONFIG_X86_64 in asm/processor.h

The below helps resolve the problem.
---------

arch/x86/include/asm/compat.h +2 -0
index 9c9b6dc..147e790 100644
@@ -233,9 +233,11 @@ static inline void __user *arch_compat_alloc_user_space(long len)
 
 	if (test_thread_flag(TIF_IA32)) {
 		sp = task_pt_regs(current)->sp;
+#ifdef CONFIG_X86_64
 	} else {
 		/* -128 for the x32 ABI redzone */
 		sp = __this_cpu_read(old_rsp) - 128;
+#endif
 	}
 
 	return (void __user *)round_down(sp - len, 16);

-----------

But I wanted to check if there is any reason why we cant include asm/compat.h in arch/x86/kernel/signal.c.

Signature · 2 lines
Thanks and Regards
Srikar
#9 lore raw

arch/x86/include/asm/compat.h static inline void __user *arch_compat_alloc_user_space(long len) · line 243

Quoted 12 lines · Srikar Dronamraju
arch/x86/include/asm/compat.h +2 -0
@@ -233,9 +233,11 @@ static inline void __user *arch_compat_alloc_user_space(long len)
 
 	if (test_thread_flag(TIF_IA32)) {
 		sp = task_pt_regs(current)->sp;
+#ifdef CONFIG_X86_64
 	} else {
 		/* -128 for the x32 ABI redzone */
 		sp = __this_cpu_read(old_rsp) - 128;
+#endif
 	}
 
 	return (void __user *)round_down(sp - len, 16);

So 'sp' is undefined if that TIF check fails?

Also, on a 32-bit kernel the TIF check probably fails all the time, because we don't set TIF_IA32 (and don't know that flag).

It would probably be better to make the whole helper inline #ifdef 64-bit, it does not look very useful on 32-bit.

Thanks,

Ingo

#10 lore raw
Quoted 18 lines · Ingo Molnar

arch/x86/include/asm/compat.h static inline void __user *arch_compat_alloc_user_space(long len) · line 243

Quoted 12 lines · Srikar Dronamraju
arch/x86/include/asm/compat.h +2 -0
@@ -233,9 +233,11 @@ static inline void __user *arch_compat_alloc_user_space(long len)
 
 	if (test_thread_flag(TIF_IA32)) {
 		sp = task_pt_regs(current)->sp;
+#ifdef CONFIG_X86_64
 	} else {
 		/* -128 for the x32 ABI redzone */
 		sp = __this_cpu_read(old_rsp) - 128;
+#endif
 	}
 
 	return (void __user *)round_down(sp - len, 16);

So 'sp' is undefined if that TIF check fails?

Quoted 7 lines

Also, on a 32-bit kernel the TIF check probably fails all the time, because we don't set TIF_IA32 (and don't know that flag).

It would probably be better to make the whole helper inline #ifdef 64-bit, it does not look very useful on 32-bit.

arch_compat_alloc_user_space gets called from compat_alloc_user_space which is arch agnostic and exported too.

So I will change this to

void __user *arch_compat_alloc_user_space(long len) {
if (is_ia32_compat_task(current))
sp = task_pt_regs(current)->sp;
#ifdef CONFIG_X86_64
else
/* -128 for the x32 ABI redzone */ sp = __this_cpu_read(old_rsp) - 128;
#endif

return (void __user *)round_down(sp - len, 16);
}

where is_ia32_compat_task() is the new macro that you suggested we put in compat.h which would return true if the task is 32 bit emulated on x86_64 or running on i386 machine.

Hence we can avoid the case where sp is not set.

Signature · 2 lines
Thanks and Regards
Srikar
#11 lore raw
Quoted 43 lines · Srikar Dronamraju
Quoted 18 lines · Ingo Molnar

arch/x86/include/asm/compat.h static inline void __user *arch_compat_alloc_user_space(long len) · line 243

Quoted 12 lines · Srikar Dronamraju
arch/x86/include/asm/compat.h +2 -0
@@ -233,9 +233,11 @@ static inline void __user *arch_compat_alloc_user_space(long len)
 
 	if (test_thread_flag(TIF_IA32)) {
 		sp = task_pt_regs(current)->sp;
+#ifdef CONFIG_X86_64
 	} else {
 		/* -128 for the x32 ABI redzone */
 		sp = __this_cpu_read(old_rsp) - 128;
+#endif
 	}
 
 	return (void __user *)round_down(sp - len, 16);

So 'sp' is undefined if that TIF check fails?

Quoted 7 lines

Also, on a 32-bit kernel the TIF check probably fails all the time, because we don't set TIF_IA32 (and don't know that flag).

It would probably be better to make the whole helper inline #ifdef 64-bit, it does not look very useful on 32-bit.

arch_compat_alloc_user_space gets called from compat_alloc_user_space which is arch agnostic and exported too.

So I will change this to

void __user *arch_compat_alloc_user_space(long len) {
if (is_ia32_compat_task(current))
sp = task_pt_regs(current)->sp;
#ifdef CONFIG_X86_64
else
/* -128 for the x32 ABI redzone */ sp = __this_cpu_read(old_rsp) - 128;
#endif

Quoted 8 lines

return (void __user *)round_down(sp - len, 16);
}

where is_ia32_compat_task() is the new macro that you suggested we put in compat.h which would return true if the task is 32 bit emulated on x86_64 or running on i386 machine.

Hence we can avoid the case where sp is not set.

Ok - looks good at first glance.

Thanks,

Ingo

#12 lore raw
Quoted 49 lines · Ingo Molnar
Quoted 46 lines · Srikar Dronamraju
Quoted 18 lines · Ingo Molnar

arch/x86/include/asm/compat.h static inline void __user *arch_compat_alloc_user_space(long len) · line 243

Quoted 12 lines · Srikar Dronamraju
arch/x86/include/asm/compat.h +2 -0
@@ -233,9 +233,11 @@ static inline void __user *arch_compat_alloc_user_space(long len)
 
 	if (test_thread_flag(TIF_IA32)) {
 		sp = task_pt_regs(current)->sp;
+#ifdef CONFIG_X86_64
 	} else {
 		/* -128 for the x32 ABI redzone */
 		sp = __this_cpu_read(old_rsp) - 128;
+#endif
 	}
 
 	return (void __user *)round_down(sp - len, 16);

So 'sp' is undefined if that TIF check fails?

Quoted 7 lines

Also, on a 32-bit kernel the TIF check probably fails all the time, because we don't set TIF_IA32 (and don't know that flag).

It would probably be better to make the whole helper inline #ifdef 64-bit, it does not look very useful on 32-bit.

arch_compat_alloc_user_space gets called from compat_alloc_user_space which is arch agnostic and exported too.

So I will change this to

void __user *arch_compat_alloc_user_space(long len) {
if (is_ia32_compat_task(current))
sp = task_pt_regs(current)->sp;
#ifdef CONFIG_X86_64
else
/* -128 for the x32 ABI redzone */ sp = __this_cpu_read(old_rsp) - 128;
#endif

return (void __user *)round_down(sp - len, 16);
}

Quoted 7 lines
Quoted 5 lines

where is_ia32_compat_task() is the new macro that you suggested we put in compat.h which would return true if the task is 32 bit emulated on x86_64 or running on i386 machine.

Hence we can avoid the case where sp is not set.

Ok - looks good at first glance.

It does not look good on a second glance though, once I checked your latest patches.

arch_compat_alloc_user_space() is arch agnostic on *CONFIG_COMPAT=y* kernels.

It's generally not available on 32-bit builds - CONFIG_COMPAT is a facility to provide 32-bit syscall compatibility on 64-bit kernels. Such a facility is not needed on 32-bit kernels.

So providing this:

Quoted 4 lines

void __user *arch_compat_alloc_user_space(long len)

{
	if (is_ia32_compat_task(current))
              sp = task_pt_regs(current)->sp;

on 32-bit systems makes little sense.

So ... instead of adding is_compat_task() to compat.h it would be better to add it to another x86 header (processor.h might be good but I have not checked very hard) and maybe name it is_32bit_task() or so, to make sure there's no confusion with CONFIG_COMPAT=y methods.

I.e. you could drop this patch altogether:

x86/trivial: Fix 'old_rsp' undefined build failure when including asm/compat.h

And rework the is_ia32_compat_task() patch to use another header and to use the is_32bit_task() name. Also, you should double check whether the x32 execution model needs special consideration as well:

#define TIF_IA32 17 /* IA32 compatibility process */
#define TIF_X32 30 /* 32-bit native x86-64 binary */

otherwise uprobe will not work with x32 tasks properly.

Thanks,

Ingo

#13 lore raw
Quoted 9 lines · Ingo Molnar

I.e. you could drop this patch altogether:

x86/trivial: Fix 'old_rsp' undefined build failure when including asm/compat.h

And rework the is_ia32_compat_task() patch to use another header and to use the is_32bit_task() name. Also, you should double check whether the x32 execution model needs special consideration as well:

Quoted 4 lines

#define TIF_IA32 17 /* IA32 compatibility process */
#define TIF_X32 30 /* 32-bit native x86-64 binary */

otherwise uprobe will not work with x32 tasks properly.

Given that x32 tasks are using native 64-bit syscalls, what you need is the is_ia32_task() check. We have that in -tip already, in compat.h. If uprobes uses that definition then uprobes will be able to probe x32 tasks as well.

To make it available to uprobes please move is_ia32_task() to processor.h and make it return 1 on i386 kernels. This should be a very simple patch - and that's all that is needed.

Thanks,

Ingo

#14 lore raw
Quoted 3 lines · Ingo Molnar

To make it available to uprobes please move is_ia32_task() to processor.h and make it return 1 on i386 kernels. This should be a very simple patch - and that's all that is needed.

Side note: in general I don't mind making compat.h a bit more generic (at all), but is_ia32_task() is really a generic x86 specific check that we can move to processor.h cleanly.

Some other effort should make compat.h bitness agnostic - uprobes doesn't actually make use of the compat types and facilities there, and it resulted in a nonsensical change like making the alloc function generic ...

Thanks,

Ingo

#15 lore raw
Quoted 7 lines · Ingo Molnar
Quoted 3 lines · Ingo Molnar

To make it available to uprobes please move is_ia32_task() to processor.h and make it return 1 on i386 kernels. This should be a very simple patch - and that's all that is needed.

Quoted 9 lines

Side note: in general I don't mind making compat.h a bit more generic (at all), but is_ia32_task() is really a generic x86 specific check that we can move to processor.h cleanly.

Some other effort should make compat.h bitness agnostic - uprobes doesn't actually make use of the compat types and facilities there, and it resulted in a nonsensical change like making the alloc function generic ...

Okay, Will move is_ia32_task() to processor.h and use it.

Signature · 2 lines
Thanks and Regards
Srikar
#16 lore raw
Quoted 13 lines · Srikar Dronamraju
Quoted 11 lines · Ingo Molnar
Quoted 3 lines · Ingo Molnar

To make it available to uprobes please move is_ia32_task() to processor.h and make it return 1 on i386 kernels. This should be a very simple patch - and that's all that is needed.

Side note: in general I don't mind making compat.h a bit more generic (at all), but is_ia32_task() is really a generic x86 specific check that we can move to processor.h cleanly.

Quoted 7 lines
Quoted 5 lines

Some other effort should make compat.h bitness agnostic - uprobes doesn't actually make use of the compat types and facilities there, and it resulted in a nonsensical change like making the alloc function generic ...

Okay, Will move is_ia32_task() to processor.h and use it.

thread_info.h might be a better place, because the TIF_ flags are defined there.

Thanks,

Ingo

Acked-by Jim Keniston Reviewed-by Jim Keniston on v1 1/1
#17 lore raw
In reply to Srikar Dronamraju
Quoted 2 lines · Srikar Dronamraju

That's not a valid signoff sequence - in the future please make this something like:

Or:

If he acks a particular patch. (Multiple authors attribution is done via the copyright notices.)

Thanks,

Ingo