[PATCH v2] kdb: use ktime_get_mono_fast_ns() instead of ktime_get_ts()

[PATCH v2] kdb: use ktime_get_mono_fast_ns() instead of ktime_get_ts()

[Baolin Wang]

The kdb code will print the monotonic time by ktime_get_ts(), but
the ktime_get_ts() will be protected by a sequence lock, that will
introduce one deadlock risk if the lock was already held in the
context from which we entered the debugger.

Thus we can use the ktime_get_mono_fast_ns() to get the monotonic
time, which is NMI safe access to clock monotonic. Moreover we can
remove the 'struct timespec', which is not y2038 safe.

Signed-off-by: Baolin Wang <baolin.wang@linaro.org>
---
Changes since v1:
 - Use ktime_get_mono_fast_ns() instead of ktime_get_seconds().
---
 kernel/debug/kdb/kdb_main.c |    6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)

diff --git a/kernel/debug/kdb/kdb_main.c b/kernel/debug/kdb/kdb_main.c
index 69e70f4..bf1d0e0 100644
--- a/kernel/debug/kdb/kdb_main.c
+++ b/kernel/debug/kdb/kdb_main.c
@@ -2486,10 +2486,10 @@ static int kdb_kill(int argc, const char **argv)
  */
 static void kdb_sysinfo(struct sysinfo *val)
 {
-	struct timespec uptime;
-	ktime_get_ts(&uptime);
+	u64 uptime = ktime_get_mono_fast_ns();
+
 	memset(val, 0, sizeof(*val));
-	val->uptime = uptime.tv_sec;
+	val->uptime = div_u64(uptime, NSEC_PER_SEC);
 	val->loads[0] = avenrun[0];
 	val->loads[1] = avenrun[1];
 	val->loads[2] = avenrun[2];
-- 
1.7.9.5

Re: [PATCH v2] kdb: use ktime_get_mono_fast_ns() instead of ktime_get_ts()

[Daniel Thompson]

On Mon, Jan 29, 2018 at 10:22:51AM +0800, Baolin Wang wrote:
> The kdb code will print the monotonic time by ktime_get_ts(), but
> the ktime_get_ts() will be protected by a sequence lock, that will
> introduce one deadlock risk if the lock was already held in the
> context from which we entered the debugger.
> 
> Thus we can use the ktime_get_mono_fast_ns() to get the monotonic
> time, which is NMI safe access to clock monotonic. Moreover we can
> remove the 'struct timespec', which is not y2038 safe.
> 
> Signed-off-by: Baolin Wang <baolin.wang@linaro.org>

Reviewed-by: Daniel Thompson <daniel.thompson@linaro.org>


> ---
> Changes since v1:
>  - Use ktime_get_mono_fast_ns() instead of ktime_get_seconds().
> ---
>  kernel/debug/kdb/kdb_main.c |    6 +++---
>  1 file changed, 3 insertions(+), 3 deletions(-)
> 
> diff --git a/kernel/debug/kdb/kdb_main.c b/kernel/debug/kdb/kdb_main.c
> index 69e70f4..bf1d0e0 100644
> --- a/kernel/debug/kdb/kdb_main.c
> +++ b/kernel/debug/kdb/kdb_main.c
> @@ -2486,10 +2486,10 @@ static int kdb_kill(int argc, const char **argv)
>   */
>  static void kdb_sysinfo(struct sysinfo *val)
>  {
> -	struct timespec uptime;
> -	ktime_get_ts(&uptime);
> +	u64 uptime = ktime_get_mono_fast_ns();
> +
>  	memset(val, 0, sizeof(*val));
> -	val->uptime = uptime.tv_sec;
> +	val->uptime = div_u64(uptime, NSEC_PER_SEC);
>  	val->loads[0] = avenrun[0];
>  	val->loads[1] = avenrun[1];
>  	val->loads[2] = avenrun[2];
> -- 
> 1.7.9.5
> 

Re: [PATCH v2] kdb: use ktime_get_mono_fast_ns() instead of ktime_get_ts()

[Arnd Bergmann]

On Mon, Jan 29, 2018 at 3:22 AM, Baolin Wang <baolin.wang@linaro.org> wrote:
> The kdb code will print the monotonic time by ktime_get_ts(), but
> the ktime_get_ts() will be protected by a sequence lock, that will
> introduce one deadlock risk if the lock was already held in the
> context from which we entered the debugger.
>
> Thus we can use the ktime_get_mono_fast_ns() to get the monotonic
> time, which is NMI safe access to clock monotonic. Moreover we can
> remove the 'struct timespec', which is not y2038 safe.
>
> Signed-off-by: Baolin Wang <baolin.wang@linaro.org>

Reviewed-by: Arnd Bergmann <arnd@arndb.de>

Re: [PATCH v2] kdb: use ktime_get_mono_fast_ns() instead of ktime_get_ts()

[Jason Wessel]

On 01/30/2018 07:35 AM, Arnd Bergmann wrote:
> On Mon, Jan 29, 2018 at 3:22 AM, Baolin Wang <baolin.wang@linaro.org> wrote:
>> The kdb code will print the monotonic time by ktime_get_ts(), but
>> the ktime_get_ts() will be protected by a sequence lock, that will
>> introduce one deadlock risk if the lock was already held in the
>> context from which we entered the debugger.
>>
>> Thus we can use the ktime_get_mono_fast_ns() to get the monotonic
>> time, which is NMI safe access to clock monotonic. Moreover we can
>> remove the 'struct timespec', which is not y2038 safe.
>>
>> Signed-off-by: Baolin Wang <baolin.wang@linaro.org>
> 
> Reviewed-by: Arnd Bergmann <arnd@arndb.de>
> 


I'll add this in today, run some tests and it should be in the merge window.

Thanks,
Jason