* [PATCH 1/9] arm64: signal: Make force_signal_inject more robust
2018-02-22 14:00 [PATCH 0/9] arm64: Consolidate use of show_unhandled_signals Will Deacon
@ 2018-02-22 14:00 ` Will Deacon
2018-02-22 14:00 ` [PATCH 2/9] arm64: signal: Force SIGKILL for unknown signals in force_signal_inject Will Deacon
` (7 subsequent siblings)
8 siblings, 0 replies; 14+ messages in thread
From: Will Deacon @ 2018-02-22 14:00 UTC (permalink / raw)
To: linux-arm-kernel
force_signal_inject is a little flakey:
* It only knows about SIGILL and SIGSEGV, so can potentially deliver
other signals based on a partially initialised siginfo_t
* It sets si_addr to point at the PC for SIGSEGV
* It always operates on current, so doesn't need the regs argument
This patch fixes these issues by always assigning the si_addr field to
the address parameter of the function and updates the callers (including
those that indirectly call via arm64_notify_segfault) accordingly.
Signed-off-by: Will Deacon <will.deacon@arm.com>
---
arch/arm64/include/asm/traps.h | 6 ++----
arch/arm64/kernel/fpsimd.c | 5 ++---
arch/arm64/kernel/traps.c | 23 ++++++++++++-----------
3 files changed, 16 insertions(+), 18 deletions(-)
diff --git a/arch/arm64/include/asm/traps.h b/arch/arm64/include/asm/traps.h
index 178e338d2889..1ee63dc38579 100644
--- a/arch/arm64/include/asm/traps.h
+++ b/arch/arm64/include/asm/traps.h
@@ -35,10 +35,8 @@ struct undef_hook {
void register_undef_hook(struct undef_hook *hook);
void unregister_undef_hook(struct undef_hook *hook);
-void force_signal_inject(int signal, int code, struct pt_regs *regs,
- unsigned long address);
-
-void arm64_notify_segfault(struct pt_regs *regs, unsigned long addr);
+void force_signal_inject(int signal, int code, unsigned long address);
+void arm64_notify_segfault(unsigned long addr);
/*
* Move regs->pc to next instruction and do necessary setup before it
diff --git a/arch/arm64/kernel/fpsimd.c b/arch/arm64/kernel/fpsimd.c
index e7226c4c7493..6964ff867d4a 100644
--- a/arch/arm64/kernel/fpsimd.c
+++ b/arch/arm64/kernel/fpsimd.c
@@ -285,8 +285,7 @@ static void task_fpsimd_save(void)
* re-enter user with corrupt state.
* There's no way to recover, so kill it:
*/
- force_signal_inject(
- SIGKILL, 0, current_pt_regs(), 0);
+ force_signal_inject(SIGKILL, 0, 0);
return;
}
@@ -831,7 +830,7 @@ asmlinkage void do_sve_acc(unsigned int esr, struct pt_regs *regs)
{
/* Even if we chose not to use SVE, the hardware could still trap: */
if (unlikely(!system_supports_sve()) || WARN_ON(is_compat_task())) {
- force_signal_inject(SIGILL, ILL_ILLOPC, regs, 0);
+ force_signal_inject(SIGILL, ILL_ILLOPC, regs->pc);
return;
}
diff --git a/arch/arm64/kernel/traps.c b/arch/arm64/kernel/traps.c
index eb2d15147e8d..c478d8e27649 100644
--- a/arch/arm64/kernel/traps.c
+++ b/arch/arm64/kernel/traps.c
@@ -311,12 +311,13 @@ static int call_undef_hook(struct pt_regs *regs)
return fn ? fn(regs, instr) : 1;
}
-void force_signal_inject(int signal, int code, struct pt_regs *regs,
- unsigned long address)
+void force_signal_inject(int signal, int code, unsigned long address)
{
siginfo_t info;
- void __user *pc = (void __user *)instruction_pointer(regs);
const char *desc;
+ struct pt_regs *regs = current_pt_regs();
+
+ clear_siginfo(&info);
switch (signal) {
case SIGILL:
@@ -332,15 +333,15 @@ void force_signal_inject(int signal, int code, struct pt_regs *regs,
if (unhandled_signal(current, signal) &&
show_unhandled_signals_ratelimited()) {
- pr_info("%s[%d]: %s: pc=%p\n",
- current->comm, task_pid_nr(current), desc, pc);
+ pr_info("%s[%d]: %s: pc=%08llx\n",
+ current->comm, task_pid_nr(current), desc, regs->pc);
dump_instr(KERN_INFO, regs);
}
info.si_signo = signal;
info.si_errno = 0;
info.si_code = code;
- info.si_addr = pc;
+ info.si_addr = (void __user *)address;
arm64_notify_die(desc, regs, &info, 0);
}
@@ -348,7 +349,7 @@ void force_signal_inject(int signal, int code, struct pt_regs *regs,
/*
* Set up process info to signal segmentation fault - called on access error.
*/
-void arm64_notify_segfault(struct pt_regs *regs, unsigned long addr)
+void arm64_notify_segfault(unsigned long addr)
{
int code;
@@ -359,7 +360,7 @@ void arm64_notify_segfault(struct pt_regs *regs, unsigned long addr)
code = SEGV_ACCERR;
up_read(¤t->mm->mmap_sem);
- force_signal_inject(SIGSEGV, code, regs, addr);
+ force_signal_inject(SIGSEGV, code, addr);
}
asmlinkage void __exception do_undefinstr(struct pt_regs *regs)
@@ -371,7 +372,7 @@ asmlinkage void __exception do_undefinstr(struct pt_regs *regs)
if (call_undef_hook(regs) == 0)
return;
- force_signal_inject(SIGILL, ILL_ILLOPC, regs, 0);
+ force_signal_inject(SIGILL, ILL_ILLOPC, regs->pc);
}
int cpu_enable_cache_maint_trap(void *__unused)
@@ -426,12 +427,12 @@ static void user_cache_maint_handler(unsigned int esr, struct pt_regs *regs)
__user_cache_maint("ic ivau", address, ret);
break;
default:
- force_signal_inject(SIGILL, ILL_ILLOPC, regs, 0);
+ force_signal_inject(SIGILL, ILL_ILLOPC, regs->pc);
return;
}
if (ret)
- arm64_notify_segfault(regs, address);
+ arm64_notify_segfault(address);
else
arm64_skip_faulting_instruction(regs, AARCH64_INSN_SIZE);
}
--
2.1.4
^ permalink raw reply related [flat|nested] 14+ messages in thread* [PATCH 2/9] arm64: signal: Force SIGKILL for unknown signals in force_signal_inject
2018-02-22 14:00 [PATCH 0/9] arm64: Consolidate use of show_unhandled_signals Will Deacon
2018-02-22 14:00 ` [PATCH 1/9] arm64: signal: Make force_signal_inject more robust Will Deacon
@ 2018-02-22 14:00 ` Will Deacon
2018-02-22 14:00 ` [PATCH 3/9] arm64: Introduce arm64_force_sig_info and hook up in arm64_notify_die Will Deacon
` (6 subsequent siblings)
8 siblings, 0 replies; 14+ messages in thread
From: Will Deacon @ 2018-02-22 14:00 UTC (permalink / raw)
To: linux-arm-kernel
For signals other than SIGKILL or those with siginfo_layout(signal, code)
== SIL_FAULT then force_signal_inject does not initialise the siginfo_t
properly. Since the signal number is determined solely by the caller,
simply WARN on unknown signals and force to SIGKILL.
Reported-by: Dave Martin <Dave.Martin@arm.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
---
arch/arm64/kernel/traps.c | 6 ++++++
1 file changed, 6 insertions(+)
diff --git a/arch/arm64/kernel/traps.c b/arch/arm64/kernel/traps.c
index c478d8e27649..3f52c07b4bf4 100644
--- a/arch/arm64/kernel/traps.c
+++ b/arch/arm64/kernel/traps.c
@@ -331,6 +331,12 @@ void force_signal_inject(int signal, int code, unsigned long address)
break;
}
+ /* Force signals we don't understand to SIGKILL */
+ if (WARN_ON(signal != SIGKILL ||
+ siginfo_layout(signal, code) != SIL_FAULT)) {
+ signal = SIGKILL;
+ }
+
if (unhandled_signal(current, signal) &&
show_unhandled_signals_ratelimited()) {
pr_info("%s[%d]: %s: pc=%08llx\n",
--
2.1.4
^ permalink raw reply related [flat|nested] 14+ messages in thread* [PATCH 3/9] arm64: Introduce arm64_force_sig_info and hook up in arm64_notify_die
2018-02-22 14:00 [PATCH 0/9] arm64: Consolidate use of show_unhandled_signals Will Deacon
2018-02-22 14:00 ` [PATCH 1/9] arm64: signal: Make force_signal_inject more robust Will Deacon
2018-02-22 14:00 ` [PATCH 2/9] arm64: signal: Force SIGKILL for unknown signals in force_signal_inject Will Deacon
@ 2018-02-22 14:00 ` Will Deacon
2018-02-22 15:39 ` Mark Rutland
2018-02-22 14:00 ` [PATCH 4/9] arm64: signal: Don't print anything directly in force_signal_inject Will Deacon
` (5 subsequent siblings)
8 siblings, 1 reply; 14+ messages in thread
From: Will Deacon @ 2018-02-22 14:00 UTC (permalink / raw)
To: linux-arm-kernel
In preparation for consolidating our handling of printing unhandled
signals, introduce a wrapper around force_sig_info which can act as
the canonical place for dealing with show_unhandled_signals.
Initially, we just hook this up to arm64_notify_die.
Signed-off-by: Will Deacon <will.deacon@arm.com>
---
arch/arm64/include/asm/traps.h | 2 ++
arch/arm64/kernel/traps.c | 28 +++++++++++++++++++++++++++-
2 files changed, 29 insertions(+), 1 deletion(-)
diff --git a/arch/arm64/include/asm/traps.h b/arch/arm64/include/asm/traps.h
index 1ee63dc38579..c320f3bf6c57 100644
--- a/arch/arm64/include/asm/traps.h
+++ b/arch/arm64/include/asm/traps.h
@@ -37,6 +37,8 @@ void register_undef_hook(struct undef_hook *hook);
void unregister_undef_hook(struct undef_hook *hook);
void force_signal_inject(int signal, int code, unsigned long address);
void arm64_notify_segfault(unsigned long addr);
+void arm64_force_sig_info(struct siginfo *info, const char *str,
+ struct task_struct *tsk);
/*
* Move regs->pc to next instruction and do necessary setup before it
diff --git a/arch/arm64/kernel/traps.c b/arch/arm64/kernel/traps.c
index 3f52c07b4bf4..b46b5be5bd6d 100644
--- a/arch/arm64/kernel/traps.c
+++ b/arch/arm64/kernel/traps.c
@@ -223,13 +223,39 @@ void die(const char *str, struct pt_regs *regs, int err)
do_exit(SIGSEGV);
}
+void arm64_force_sig_info(struct siginfo *info, const char *str,
+ struct task_struct *tsk)
+{
+ unsigned int esr = tsk->thread.fault_code;
+ struct pt_regs *regs = task_pt_regs(tsk);
+
+ if (!unhandled_signal(tsk, info->si_signo))
+ goto send_sig;
+
+ if (!show_unhandled_signals_ratelimited())
+ goto send_sig;
+
+ pr_info("%s[%d]: unhandled exception: ", tsk->comm, task_pid_nr(tsk));
+ if (esr)
+ pr_cont("%s, ESR 0x%03x, ", esr_get_class_string(esr), esr);
+
+ pr_cont("%s", str);
+ print_vma_addr(KERN_CONT " in ", regs->pc);
+ pr_cont("\n");
+ __show_regs(regs);
+
+send_sig:
+ force_sig_info(info->si_signo, info, tsk);
+}
+
void arm64_notify_die(const char *str, struct pt_regs *regs,
struct siginfo *info, int err)
{
if (user_mode(regs)) {
+ WARN_ON(regs != current_pt_regs());
current->thread.fault_address = 0;
current->thread.fault_code = err;
- force_sig_info(info->si_signo, info, current);
+ arm64_force_sig_info(info, str, current);
} else {
die(str, regs, err);
}
--
2.1.4
^ permalink raw reply related [flat|nested] 14+ messages in thread* [PATCH 3/9] arm64: Introduce arm64_force_sig_info and hook up in arm64_notify_die
2018-02-22 14:00 ` [PATCH 3/9] arm64: Introduce arm64_force_sig_info and hook up in arm64_notify_die Will Deacon
@ 2018-02-22 15:39 ` Mark Rutland
0 siblings, 0 replies; 14+ messages in thread
From: Mark Rutland @ 2018-02-22 15:39 UTC (permalink / raw)
To: linux-arm-kernel
On Thu, Feb 22, 2018 at 02:00:51PM +0000, Will Deacon wrote:
> +void arm64_force_sig_info(struct siginfo *info, const char *str,
> + struct task_struct *tsk)
> +{
> + unsigned int esr = tsk->thread.fault_code;
> + struct pt_regs *regs = task_pt_regs(tsk);
> +
> + if (!unhandled_signal(tsk, info->si_signo))
> + goto send_sig;
> +
> + if (!show_unhandled_signals_ratelimited())
> + goto send_sig;
> +
> + pr_info("%s[%d]: unhandled exception: ", tsk->comm, task_pid_nr(tsk));
> + if (esr)
> + pr_cont("%s, ESR 0x%03x, ", esr_get_class_string(esr), esr);
Can we make that 0x%08x please? ESR_ELx.EC occupies bits 31-26...
I guess this was copied from __do_user_fault, and it's odd there, as
wherever that's called the EC is non-zero.
Thanks,
Mark.
^ permalink raw reply [flat|nested] 14+ messages in thread
* [PATCH 4/9] arm64: signal: Don't print anything directly in force_signal_inject
2018-02-22 14:00 [PATCH 0/9] arm64: Consolidate use of show_unhandled_signals Will Deacon
` (2 preceding siblings ...)
2018-02-22 14:00 ` [PATCH 3/9] arm64: Introduce arm64_force_sig_info and hook up in arm64_notify_die Will Deacon
@ 2018-02-22 14:00 ` Will Deacon
2018-02-22 14:00 ` [PATCH 5/9] arm64: Pass user fault info to arm64_notify_die instead of printing it Will Deacon
` (4 subsequent siblings)
8 siblings, 0 replies; 14+ messages in thread
From: Will Deacon @ 2018-02-22 14:00 UTC (permalink / raw)
To: linux-arm-kernel
arm64_notify_die deals with printing out information regarding unhandled
signals, so there's no need to roll our own code here.
Signed-off-by: Will Deacon <will.deacon@arm.com>
---
arch/arm64/kernel/traps.c | 7 -------
1 file changed, 7 deletions(-)
diff --git a/arch/arm64/kernel/traps.c b/arch/arm64/kernel/traps.c
index b46b5be5bd6d..637ae3702950 100644
--- a/arch/arm64/kernel/traps.c
+++ b/arch/arm64/kernel/traps.c
@@ -363,13 +363,6 @@ void force_signal_inject(int signal, int code, unsigned long address)
signal = SIGKILL;
}
- if (unhandled_signal(current, signal) &&
- show_unhandled_signals_ratelimited()) {
- pr_info("%s[%d]: %s: pc=%08llx\n",
- current->comm, task_pid_nr(current), desc, regs->pc);
- dump_instr(KERN_INFO, regs);
- }
-
info.si_signo = signal;
info.si_errno = 0;
info.si_code = code;
--
2.1.4
^ permalink raw reply related [flat|nested] 14+ messages in thread* [PATCH 5/9] arm64: Pass user fault info to arm64_notify_die instead of printing it
2018-02-22 14:00 [PATCH 0/9] arm64: Consolidate use of show_unhandled_signals Will Deacon
` (3 preceding siblings ...)
2018-02-22 14:00 ` [PATCH 4/9] arm64: signal: Don't print anything directly in force_signal_inject Will Deacon
@ 2018-02-22 14:00 ` Will Deacon
2018-02-22 14:00 ` [PATCH 6/9] arm64: mm: Rework unhandled user pagefaults to call arm64_force_sig_info Will Deacon
` (3 subsequent siblings)
8 siblings, 0 replies; 14+ messages in thread
From: Will Deacon @ 2018-02-22 14:00 UTC (permalink / raw)
To: linux-arm-kernel
There's no need for callers of arm64_notify_die to print information
about user faults. Instead, they can pass a string to arm64_notify_die
which will be printed subject to show_unhandled_signals.
Signed-off-by: Will Deacon <will.deacon@arm.com>
---
arch/arm64/mm/fault.c | 30 ++++++++----------------------
1 file changed, 8 insertions(+), 22 deletions(-)
diff --git a/arch/arm64/mm/fault.c b/arch/arm64/mm/fault.c
index bff11553eb05..fd5928afd9cd 100644
--- a/arch/arm64/mm/fault.c
+++ b/arch/arm64/mm/fault.c
@@ -582,8 +582,6 @@ static int do_sea(unsigned long addr, unsigned int esr, struct pt_regs *regs)
const struct fault_info *inf;
inf = esr_to_fault_info(esr);
- pr_err("Synchronous External Abort: %s (0x%08x) at 0x%016lx\n",
- inf->name, esr, addr);
/*
* Synchronous aborts may interrupt code which had interrupts masked.
@@ -607,7 +605,7 @@ static int do_sea(unsigned long addr, unsigned int esr, struct pt_regs *regs)
info.si_addr = NULL;
else
info.si_addr = (void __user *)addr;
- arm64_notify_die("", regs, &info, esr);
+ arm64_notify_die(inf->name, regs, &info, esr);
return 0;
}
@@ -698,19 +696,17 @@ asmlinkage void __exception do_mem_abort(unsigned long addr, unsigned int esr,
if (!inf->fn(addr, esr, regs))
return;
- pr_alert("Unhandled fault: %s at 0x%016lx\n",
- inf->name, addr);
-
- mem_abort_decode(esr);
-
- if (!user_mode(regs))
+ if (!user_mode(regs)) {
+ pr_alert("Unhandled fault at 0x%016lx\n", addr);
+ mem_abort_decode(esr);
show_pte(addr);
+ }
info.si_signo = inf->sig;
info.si_errno = 0;
info.si_code = inf->code;
info.si_addr = (void __user *)addr;
- arm64_notify_die("", regs, &info, esr);
+ arm64_notify_die(inf->name, regs, &info, esr);
}
asmlinkage void __exception do_el0_irq_bp_hardening(void)
@@ -741,7 +737,6 @@ asmlinkage void __exception do_sp_pc_abort(unsigned long addr,
struct pt_regs *regs)
{
struct siginfo info;
- struct task_struct *tsk = current;
if (user_mode(regs)) {
if (instruction_pointer(regs) > TASK_SIZE)
@@ -749,17 +744,11 @@ asmlinkage void __exception do_sp_pc_abort(unsigned long addr,
local_irq_enable();
}
- if (show_unhandled_signals && unhandled_signal(tsk, SIGBUS))
- pr_info_ratelimited("%s[%d]: %s exception: pc=%p sp=%p\n",
- tsk->comm, task_pid_nr(tsk),
- esr_get_class_string(esr), (void *)regs->pc,
- (void *)regs->sp);
-
info.si_signo = SIGBUS;
info.si_errno = 0;
info.si_code = BUS_ADRALN;
info.si_addr = (void __user *)addr;
- arm64_notify_die("Oops - SP/PC alignment exception", regs, &info, esr);
+ arm64_notify_die("SP/PC alignment exception", regs, &info, esr);
}
int __init early_brk64(unsigned long addr, unsigned int esr,
@@ -814,14 +803,11 @@ asmlinkage int __exception do_debug_exception(unsigned long addr,
if (!inf->fn(addr, esr, regs)) {
rv = 1;
} else {
- pr_alert("Unhandled debug exception: %s (0x%08x) at 0x%016lx\n",
- inf->name, esr, addr);
-
info.si_signo = inf->sig;
info.si_errno = 0;
info.si_code = inf->code;
info.si_addr = (void __user *)addr;
- arm64_notify_die("", regs, &info, 0);
+ arm64_notify_die(inf->name, regs, &info, esr);
rv = 0;
}
--
2.1.4
^ permalink raw reply related [flat|nested] 14+ messages in thread* [PATCH 6/9] arm64: mm: Rework unhandled user pagefaults to call arm64_force_sig_info
2018-02-22 14:00 [PATCH 0/9] arm64: Consolidate use of show_unhandled_signals Will Deacon
` (4 preceding siblings ...)
2018-02-22 14:00 ` [PATCH 5/9] arm64: Pass user fault info to arm64_notify_die instead of printing it Will Deacon
@ 2018-02-22 14:00 ` Will Deacon
2018-02-22 14:00 ` [PATCH 7/9] arm64: signal: Call arm64_notify_segfault when failing to deliver signal Will Deacon
` (2 subsequent siblings)
8 siblings, 0 replies; 14+ messages in thread
From: Will Deacon @ 2018-02-22 14:00 UTC (permalink / raw)
To: linux-arm-kernel
Reporting unhandled user pagefaults via arm64_force_sig_info means
that __do_user_fault can be drastically simplified, since it no longer
has to worry about printing the fault information and can consequently
just take the siginfo as a parameter.
Signed-off-by: Will Deacon <will.deacon@arm.com>
---
arch/arm64/mm/fault.c | 87 +++++++++++++++++++++------------------------------
1 file changed, 36 insertions(+), 51 deletions(-)
diff --git a/arch/arm64/mm/fault.c b/arch/arm64/mm/fault.c
index fd5928afd9cd..49dfb08a6c4d 100644
--- a/arch/arm64/mm/fault.c
+++ b/arch/arm64/mm/fault.c
@@ -43,6 +43,7 @@
#include <asm/system_misc.h>
#include <asm/pgtable.h>
#include <asm/tlbflush.h>
+#include <asm/traps.h>
#include <acpi/ghes.h>
@@ -289,58 +290,31 @@ static void __do_kernel_fault(unsigned long addr, unsigned int esr,
do_exit(SIGKILL);
}
-static void __do_user_fault(struct task_struct *tsk, unsigned long addr,
- unsigned int esr, unsigned int sig, int code,
- struct pt_regs *regs, int fault)
+static void __do_user_fault(struct siginfo *info, unsigned int esr)
{
- struct siginfo si;
- const struct fault_info *inf;
- unsigned int lsb = 0;
-
- if (unhandled_signal(tsk, sig) && show_unhandled_signals_ratelimited()) {
- inf = esr_to_fault_info(esr);
- pr_info("%s[%d]: unhandled %s (%d) at 0x%08lx, esr 0x%03x",
- tsk->comm, task_pid_nr(tsk), inf->name, sig,
- addr, esr);
- print_vma_addr(KERN_CONT ", in ", regs->pc);
- pr_cont("\n");
- __show_regs(regs);
- }
-
- tsk->thread.fault_address = addr;
- tsk->thread.fault_code = esr;
- si.si_signo = sig;
- si.si_errno = 0;
- si.si_code = code;
- si.si_addr = (void __user *)addr;
- /*
- * Either small page or large page may be poisoned.
- * In other words, VM_FAULT_HWPOISON_LARGE and
- * VM_FAULT_HWPOISON are mutually exclusive.
- */
- if (fault & VM_FAULT_HWPOISON_LARGE)
- lsb = hstate_index_to_shift(VM_FAULT_GET_HINDEX(fault));
- else if (fault & VM_FAULT_HWPOISON)
- lsb = PAGE_SHIFT;
- si.si_addr_lsb = lsb;
-
- force_sig_info(sig, &si, tsk);
+ current->thread.fault_address = (unsigned long)info->si_addr;
+ current->thread.fault_code = esr;
+ arm64_force_sig_info(info, esr_to_fault_info(esr)->name, current);
}
static void do_bad_area(unsigned long addr, unsigned int esr, struct pt_regs *regs)
{
- struct task_struct *tsk = current;
- const struct fault_info *inf;
-
/*
* If we are in kernel mode at this point, we have no context to
* handle this fault with.
*/
if (user_mode(regs)) {
- inf = esr_to_fault_info(esr);
- __do_user_fault(tsk, addr, esr, inf->sig, inf->code, regs, 0);
- } else
+ const struct fault_info *inf = esr_to_fault_info(esr);
+ struct siginfo si = {
+ .si_signo = inf->sig,
+ .si_code = inf->code,
+ .si_addr = (void __user *)addr,
+ };
+
+ __do_user_fault(&si, esr);
+ } else {
__do_kernel_fault(addr, esr, regs);
+ }
}
#define VM_FAULT_BADMAP 0x010000
@@ -393,7 +367,8 @@ static int __kprobes do_page_fault(unsigned long addr, unsigned int esr,
{
struct task_struct *tsk;
struct mm_struct *mm;
- int fault, sig, code, major = 0;
+ struct siginfo si;
+ int fault, major = 0;
unsigned long vm_flags = VM_READ | VM_WRITE;
unsigned int mm_flags = FAULT_FLAG_ALLOW_RETRY | FAULT_FLAG_KILLABLE;
@@ -525,27 +500,37 @@ static int __kprobes do_page_fault(unsigned long addr, unsigned int esr,
return 0;
}
+ clear_siginfo(&si);
+ si.si_addr = (void __user *)addr;
+
if (fault & VM_FAULT_SIGBUS) {
/*
* We had some memory, but were unable to successfully fix up
* this page fault.
*/
- sig = SIGBUS;
- code = BUS_ADRERR;
- } else if (fault & (VM_FAULT_HWPOISON | VM_FAULT_HWPOISON_LARGE)) {
- sig = SIGBUS;
- code = BUS_MCEERR_AR;
+ si.si_signo = SIGBUS;
+ si.si_code = BUS_ADRERR;
+ } else if (fault & VM_FAULT_HWPOISON_LARGE) {
+ unsigned int hindex = VM_FAULT_GET_HINDEX(fault);
+
+ si.si_signo = SIGBUS;
+ si.si_code = BUS_MCEERR_AR;
+ si.si_addr_lsb = hstate_index_to_shift(hindex);
+ } else if (fault & VM_FAULT_HWPOISON) {
+ si.si_signo = SIGBUS;
+ si.si_code = BUS_MCEERR_AR;
+ si.si_addr_lsb = PAGE_SHIFT;
} else {
/*
* Something tried to access memory that isn't in our memory
* map.
*/
- sig = SIGSEGV;
- code = fault == VM_FAULT_BADACCESS ?
- SEGV_ACCERR : SEGV_MAPERR;
+ si.si_signo = SIGSEGV;
+ si.si_code = fault == VM_FAULT_BADACCESS ?
+ SEGV_ACCERR : SEGV_MAPERR;
}
- __do_user_fault(tsk, addr, esr, sig, code, regs, fault);
+ __do_user_fault(&si, esr);
return 0;
no_context:
--
2.1.4
^ permalink raw reply related [flat|nested] 14+ messages in thread* [PATCH 7/9] arm64: signal: Call arm64_notify_segfault when failing to deliver signal
2018-02-22 14:00 [PATCH 0/9] arm64: Consolidate use of show_unhandled_signals Will Deacon
` (5 preceding siblings ...)
2018-02-22 14:00 ` [PATCH 6/9] arm64: mm: Rework unhandled user pagefaults to call arm64_force_sig_info Will Deacon
@ 2018-02-22 14:00 ` Will Deacon
2018-02-22 14:00 ` [PATCH 8/9] arm64: Move show_unhandled_signals_ratelimited into traps.c Will Deacon
2018-02-22 14:00 ` [PATCH 9/9] arm64: Use arm64_force_sig_info instead of force_sig_info Will Deacon
8 siblings, 0 replies; 14+ messages in thread
From: Will Deacon @ 2018-02-22 14:00 UTC (permalink / raw)
To: linux-arm-kernel
If we fail to deliver a signal due to taking an unhandled fault on the
stackframe, we can all arm64_notify_segfault to deliver a SEGV can deal
with printing any unhandled signal messages for us, rather than roll our
own printing code.
A side-effect of this change is that we now deliver the frame address
in si_addr along with an si_code of SEGV_{ACC,MAP}ERR, rather than an
si_addr of 0 and an si_code of SI_KERNEL as before.
Signed-off-by: Will Deacon <will.deacon@arm.com>
---
arch/arm64/kernel/signal.c | 7 ++-----
arch/arm64/kernel/signal32.c | 13 +++----------
2 files changed, 5 insertions(+), 15 deletions(-)
diff --git a/arch/arm64/kernel/signal.c b/arch/arm64/kernel/signal.c
index f60c052e8d1c..e5c656d0e316 100644
--- a/arch/arm64/kernel/signal.c
+++ b/arch/arm64/kernel/signal.c
@@ -40,6 +40,7 @@
#include <asm/fpsimd.h>
#include <asm/ptrace.h>
#include <asm/signal32.h>
+#include <asm/traps.h>
#include <asm/vdso.h>
/*
@@ -565,11 +566,7 @@ asmlinkage long sys_rt_sigreturn(struct pt_regs *regs)
return regs->regs[0];
badframe:
- if (show_unhandled_signals)
- pr_info_ratelimited("%s[%d]: bad frame in %s: pc=%08llx sp=%08llx\n",
- current->comm, task_pid_nr(current), __func__,
- regs->pc, regs->sp);
- force_sig(SIGSEGV, current);
+ arm64_notify_segfault(regs->sp);
return 0;
}
diff --git a/arch/arm64/kernel/signal32.c b/arch/arm64/kernel/signal32.c
index 79feb861929b..68f5e07b592b 100644
--- a/arch/arm64/kernel/signal32.c
+++ b/arch/arm64/kernel/signal32.c
@@ -26,6 +26,7 @@
#include <asm/esr.h>
#include <asm/fpsimd.h>
#include <asm/signal32.h>
+#include <asm/traps.h>
#include <linux/uaccess.h>
#include <asm/unistd.h>
@@ -307,11 +308,7 @@ asmlinkage int compat_sys_sigreturn(struct pt_regs *regs)
return regs->regs[0];
badframe:
- if (show_unhandled_signals)
- pr_info_ratelimited("%s[%d]: bad frame in %s: pc=%08llx sp=%08llx\n",
- current->comm, task_pid_nr(current), __func__,
- regs->pc, regs->compat_sp);
- force_sig(SIGSEGV, current);
+ arm64_notify_segfault(regs->compat_sp);
return 0;
}
@@ -344,11 +341,7 @@ asmlinkage int compat_sys_rt_sigreturn(struct pt_regs *regs)
return regs->regs[0];
badframe:
- if (show_unhandled_signals)
- pr_info_ratelimited("%s[%d]: bad frame in %s: pc=%08llx sp=%08llx\n",
- current->comm, task_pid_nr(current), __func__,
- regs->pc, regs->compat_sp);
- force_sig(SIGSEGV, current);
+ arm64_notify_segfault(regs->compat_sp);
return 0;
}
--
2.1.4
^ permalink raw reply related [flat|nested] 14+ messages in thread* [PATCH 8/9] arm64: Move show_unhandled_signals_ratelimited into traps.c
2018-02-22 14:00 [PATCH 0/9] arm64: Consolidate use of show_unhandled_signals Will Deacon
` (6 preceding siblings ...)
2018-02-22 14:00 ` [PATCH 7/9] arm64: signal: Call arm64_notify_segfault when failing to deliver signal Will Deacon
@ 2018-02-22 14:00 ` Will Deacon
2018-02-22 14:00 ` [PATCH 9/9] arm64: Use arm64_force_sig_info instead of force_sig_info Will Deacon
8 siblings, 0 replies; 14+ messages in thread
From: Will Deacon @ 2018-02-22 14:00 UTC (permalink / raw)
To: linux-arm-kernel
show_unhandled_signals_ratelimited is only called in traps.c, so move it
out of its macro in the dreaded system_misc.h and into a static function
in traps.c
Signed-off-by: Will Deacon <will.deacon@arm.com>
---
arch/arm64/include/asm/system_misc.h | 11 -----------
arch/arm64/kernel/traps.c | 7 +++++++
2 files changed, 7 insertions(+), 11 deletions(-)
diff --git a/arch/arm64/include/asm/system_misc.h b/arch/arm64/include/asm/system_misc.h
index 07aa8e3c5630..28893a0b141d 100644
--- a/arch/arm64/include/asm/system_misc.h
+++ b/arch/arm64/include/asm/system_misc.h
@@ -45,17 +45,6 @@ extern void __show_regs(struct pt_regs *);
extern void (*arm_pm_restart)(enum reboot_mode reboot_mode, const char *cmd);
-#define show_unhandled_signals_ratelimited() \
-({ \
- static DEFINE_RATELIMIT_STATE(_rs, \
- DEFAULT_RATELIMIT_INTERVAL, \
- DEFAULT_RATELIMIT_BURST); \
- bool __show_ratelimited = false; \
- if (show_unhandled_signals && __ratelimit(&_rs)) \
- __show_ratelimited = true; \
- __show_ratelimited; \
-})
-
int handle_guest_sea(phys_addr_t addr, unsigned int esr);
#endif /* __ASSEMBLY__ */
diff --git a/arch/arm64/kernel/traps.c b/arch/arm64/kernel/traps.c
index 637ae3702950..4231488a5a9f 100644
--- a/arch/arm64/kernel/traps.c
+++ b/arch/arm64/kernel/traps.c
@@ -223,6 +223,13 @@ void die(const char *str, struct pt_regs *regs, int err)
do_exit(SIGSEGV);
}
+static bool show_unhandled_signals_ratelimited(void)
+{
+ static DEFINE_RATELIMIT_STATE(rs, DEFAULT_RATELIMIT_INTERVAL,
+ DEFAULT_RATELIMIT_BURST);
+ return show_unhandled_signals && __ratelimit(&rs);
+}
+
void arm64_force_sig_info(struct siginfo *info, const char *str,
struct task_struct *tsk)
{
--
2.1.4
^ permalink raw reply related [flat|nested] 14+ messages in thread* [PATCH 9/9] arm64: Use arm64_force_sig_info instead of force_sig_info
2018-02-22 14:00 [PATCH 0/9] arm64: Consolidate use of show_unhandled_signals Will Deacon
` (7 preceding siblings ...)
2018-02-22 14:00 ` [PATCH 8/9] arm64: Move show_unhandled_signals_ratelimited into traps.c Will Deacon
@ 2018-02-22 14:00 ` Will Deacon
2018-02-22 15:48 ` Mark Rutland
8 siblings, 1 reply; 14+ messages in thread
From: Will Deacon @ 2018-02-22 14:00 UTC (permalink / raw)
To: linux-arm-kernel
Using arm64_force_sig_info means that printing messages about unhandled
signals is dealt with for us, so use that in preference to force_sig_info
and remove any homebrew printing code.
Signed-off-by: Will Deacon <will.deacon@arm.com>
---
arch/arm64/kernel/debug-monitors.c | 3 ++-
arch/arm64/kernel/ptrace.c | 2 +-
arch/arm64/kernel/traps.c | 9 ++-------
3 files changed, 5 insertions(+), 9 deletions(-)
diff --git a/arch/arm64/kernel/debug-monitors.c b/arch/arm64/kernel/debug-monitors.c
index 53781f5687c5..06ca574495af 100644
--- a/arch/arm64/kernel/debug-monitors.c
+++ b/arch/arm64/kernel/debug-monitors.c
@@ -33,6 +33,7 @@
#include <asm/daifflags.h>
#include <asm/debug-monitors.h>
#include <asm/system_misc.h>
+#include <asm/traps.h>
/* Determine debug architecture. */
u8 debug_monitors_arch(void)
@@ -223,7 +224,7 @@ static void send_user_sigtrap(int si_code)
if (interrupts_enabled(regs))
local_irq_enable();
- force_sig_info(SIGTRAP, &info, current);
+ arm64_force_sig_info(&info, "User debug trap", current);
}
static int single_step_handler(unsigned long addr, unsigned int esr,
diff --git a/arch/arm64/kernel/ptrace.c b/arch/arm64/kernel/ptrace.c
index 6618036ae6d4..dd359ee02e2b 100644
--- a/arch/arm64/kernel/ptrace.c
+++ b/arch/arm64/kernel/ptrace.c
@@ -209,7 +209,7 @@ static void ptrace_hbptriggered(struct perf_event *bp,
force_sig_ptrace_errno_trap(si_errno, (void __user *)bkpt->trigger);
}
#endif
- force_sig_info(SIGTRAP, &info, current);
+ arm64_force_sig_info(&info, "Hardware breakpoint trap (ptrace)", current);
}
/*
diff --git a/arch/arm64/kernel/traps.c b/arch/arm64/kernel/traps.c
index 4231488a5a9f..cb5ad77524f3 100644
--- a/arch/arm64/kernel/traps.c
+++ b/arch/arm64/kernel/traps.c
@@ -633,11 +633,6 @@ asmlinkage void bad_el0_sync(struct pt_regs *regs, int reason, unsigned int esr)
{
siginfo_t info;
void __user *pc = (void __user *)instruction_pointer(regs);
- console_verbose();
-
- pr_crit("Bad EL0 synchronous exception detected on CPU%d, code 0x%08x -- %s\n",
- smp_processor_id(), esr, esr_get_class_string(esr));
- __show_regs(regs);
info.si_signo = SIGILL;
info.si_errno = 0;
@@ -645,9 +640,9 @@ asmlinkage void bad_el0_sync(struct pt_regs *regs, int reason, unsigned int esr)
info.si_addr = pc;
current->thread.fault_address = 0;
- current->thread.fault_code = 0;
+ current->thread.fault_code = esr;
- force_sig_info(info.si_signo, &info, current);
+ arm64_force_sig_info(&info, "Bad EL0 synchronous exception", current);
}
#ifdef CONFIG_VMAP_STACK
--
2.1.4
^ permalink raw reply related [flat|nested] 14+ messages in thread* [PATCH 9/9] arm64: Use arm64_force_sig_info instead of force_sig_info
2018-02-22 14:00 ` [PATCH 9/9] arm64: Use arm64_force_sig_info instead of force_sig_info Will Deacon
@ 2018-02-22 15:48 ` Mark Rutland
2018-02-22 16:58 ` Will Deacon
0 siblings, 1 reply; 14+ messages in thread
From: Mark Rutland @ 2018-02-22 15:48 UTC (permalink / raw)
To: linux-arm-kernel
On Thu, Feb 22, 2018 at 02:00:57PM +0000, Will Deacon wrote:
> Using arm64_force_sig_info means that printing messages about unhandled
> signals is dealt with for us, so use that in preference to force_sig_info
> and remove any homebrew printing code.
>
> Signed-off-by: Will Deacon <will.deacon@arm.com>
> ---
> arch/arm64/kernel/debug-monitors.c | 3 ++-
> arch/arm64/kernel/ptrace.c | 2 +-
> arch/arm64/kernel/traps.c | 9 ++-------
> 3 files changed, 5 insertions(+), 9 deletions(-)
>
> diff --git a/arch/arm64/kernel/debug-monitors.c b/arch/arm64/kernel/debug-monitors.c
> index 53781f5687c5..06ca574495af 100644
> --- a/arch/arm64/kernel/debug-monitors.c
> +++ b/arch/arm64/kernel/debug-monitors.c
> @@ -33,6 +33,7 @@
> #include <asm/daifflags.h>
> #include <asm/debug-monitors.h>
> #include <asm/system_misc.h>
> +#include <asm/traps.h>
>
> /* Determine debug architecture. */
> u8 debug_monitors_arch(void)
> @@ -223,7 +224,7 @@ static void send_user_sigtrap(int si_code)
> if (interrupts_enabled(regs))
> local_irq_enable();
>
> - force_sig_info(SIGTRAP, &info, current);
> + arm64_force_sig_info(&info, "User debug trap", current);
> }
>
> static int single_step_handler(unsigned long addr, unsigned int esr,
> diff --git a/arch/arm64/kernel/ptrace.c b/arch/arm64/kernel/ptrace.c
> index 6618036ae6d4..dd359ee02e2b 100644
> --- a/arch/arm64/kernel/ptrace.c
> +++ b/arch/arm64/kernel/ptrace.c
> @@ -209,7 +209,7 @@ static void ptrace_hbptriggered(struct perf_event *bp,
> force_sig_ptrace_errno_trap(si_errno, (void __user *)bkpt->trigger);
> }
> #endif
> - force_sig_info(SIGTRAP, &info, current);
> + arm64_force_sig_info(&info, "Hardware breakpoint trap (ptrace)", current);
> }
>
> /*
> diff --git a/arch/arm64/kernel/traps.c b/arch/arm64/kernel/traps.c
> index 4231488a5a9f..cb5ad77524f3 100644
> --- a/arch/arm64/kernel/traps.c
> +++ b/arch/arm64/kernel/traps.c
> @@ -633,11 +633,6 @@ asmlinkage void bad_el0_sync(struct pt_regs *regs, int reason, unsigned int esr)
> {
> siginfo_t info;
> void __user *pc = (void __user *)instruction_pointer(regs);
> - console_verbose();
> -
> - pr_crit("Bad EL0 synchronous exception detected on CPU%d, code 0x%08x -- %s\n",
> - smp_processor_id(), esr, esr_get_class_string(esr));
> - __show_regs(regs);
Here we dump the CPU so that a developer can figure out which
microarchitecture the exception was taken on (e.g. to allow them to
decode an IMP DEF SError). We added that in commit:
8051f4d16ef1d037 ("arm64: report CPU number in bad_mode")
However, in arm64_force_sig_info() we don't dump the CPU, and I guess
that may be preemptible in cases other than bad_el0_sync().
Are we happy to lose this?
Thanks,
Mark.
^ permalink raw reply [flat|nested] 14+ messages in thread* [PATCH 9/9] arm64: Use arm64_force_sig_info instead of force_sig_info
2018-02-22 15:48 ` Mark Rutland
@ 2018-02-22 16:58 ` Will Deacon
2018-02-22 16:59 ` Mark Rutland
0 siblings, 1 reply; 14+ messages in thread
From: Will Deacon @ 2018-02-22 16:58 UTC (permalink / raw)
To: linux-arm-kernel
On Thu, Feb 22, 2018 at 03:48:21PM +0000, Mark Rutland wrote:
> On Thu, Feb 22, 2018 at 02:00:57PM +0000, Will Deacon wrote:
> > Using arm64_force_sig_info means that printing messages about unhandled
> > signals is dealt with for us, so use that in preference to force_sig_info
> > and remove any homebrew printing code.
> >
> > Signed-off-by: Will Deacon <will.deacon@arm.com>
> > ---
> > arch/arm64/kernel/debug-monitors.c | 3 ++-
> > arch/arm64/kernel/ptrace.c | 2 +-
> > arch/arm64/kernel/traps.c | 9 ++-------
> > 3 files changed, 5 insertions(+), 9 deletions(-)
[...]
> > diff --git a/arch/arm64/kernel/traps.c b/arch/arm64/kernel/traps.c
> > index 4231488a5a9f..cb5ad77524f3 100644
> > --- a/arch/arm64/kernel/traps.c
> > +++ b/arch/arm64/kernel/traps.c
> > @@ -633,11 +633,6 @@ asmlinkage void bad_el0_sync(struct pt_regs *regs, int reason, unsigned int esr)
> > {
> > siginfo_t info;
> > void __user *pc = (void __user *)instruction_pointer(regs);
> > - console_verbose();
> > -
> > - pr_crit("Bad EL0 synchronous exception detected on CPU%d, code 0x%08x -- %s\n",
> > - smp_processor_id(), esr, esr_get_class_string(esr));
> > - __show_regs(regs);
>
> Here we dump the CPU so that a developer can figure out which
> microarchitecture the exception was taken on (e.g. to allow them to
> decode an IMP DEF SError). We added that in commit:
>
> 8051f4d16ef1d037 ("arm64: report CPU number in bad_mode")
>
> However, in arm64_force_sig_info() we don't dump the CPU, and I guess
> that may be preemptible in cases other than bad_el0_sync().
__show_regs prints the CPU number already. If we're preemptible, then it
could be inaccurate.
Will
^ permalink raw reply [flat|nested] 14+ messages in thread* [PATCH 9/9] arm64: Use arm64_force_sig_info instead of force_sig_info
2018-02-22 16:58 ` Will Deacon
@ 2018-02-22 16:59 ` Mark Rutland
0 siblings, 0 replies; 14+ messages in thread
From: Mark Rutland @ 2018-02-22 16:59 UTC (permalink / raw)
To: linux-arm-kernel
On Thu, Feb 22, 2018 at 04:58:05PM +0000, Will Deacon wrote:
> On Thu, Feb 22, 2018 at 03:48:21PM +0000, Mark Rutland wrote:
> > On Thu, Feb 22, 2018 at 02:00:57PM +0000, Will Deacon wrote:
> > > Using arm64_force_sig_info means that printing messages about unhandled
> > > signals is dealt with for us, so use that in preference to force_sig_info
> > > and remove any homebrew printing code.
> > >
> > > Signed-off-by: Will Deacon <will.deacon@arm.com>
> > > ---
> > > arch/arm64/kernel/debug-monitors.c | 3 ++-
> > > arch/arm64/kernel/ptrace.c | 2 +-
> > > arch/arm64/kernel/traps.c | 9 ++-------
> > > 3 files changed, 5 insertions(+), 9 deletions(-)
>
> [...]
>
> > > diff --git a/arch/arm64/kernel/traps.c b/arch/arm64/kernel/traps.c
> > > index 4231488a5a9f..cb5ad77524f3 100644
> > > --- a/arch/arm64/kernel/traps.c
> > > +++ b/arch/arm64/kernel/traps.c
> > > @@ -633,11 +633,6 @@ asmlinkage void bad_el0_sync(struct pt_regs *regs, int reason, unsigned int esr)
> > > {
> > > siginfo_t info;
> > > void __user *pc = (void __user *)instruction_pointer(regs);
> > > - console_verbose();
> > > -
> > > - pr_crit("Bad EL0 synchronous exception detected on CPU%d, code 0x%08x -- %s\n",
> > > - smp_processor_id(), esr, esr_get_class_string(esr));
> > > - __show_regs(regs);
> >
> > Here we dump the CPU so that a developer can figure out which
> > microarchitecture the exception was taken on (e.g. to allow them to
> > decode an IMP DEF SError). We added that in commit:
> >
> > 8051f4d16ef1d037 ("arm64: report CPU number in bad_mode")
> >
> > However, in arm64_force_sig_info() we don't dump the CPU, and I guess
> > that may be preemptible in cases other than bad_el0_sync().
>
> __show_regs prints the CPU number already. If we're preemptible, then it
> could be inaccurate.
Ah, that'll do. :)
Mark.
^ permalink raw reply [flat|nested] 14+ messages in thread