Skip to contentSkip to conversation

add time_now_after and other macros which compare with jiffies

PATCH In review Dave Young 4 participants 12 messages Mar 7, 2008 – Mar 10, 2008

lore
Share & tools
lore

Dave Young

author
#0 lore raw

PATCH v2 1/1 add time_now_after and other macros which compare with jiffies

Changes from previous version:
1. Add comments
2. Change names easy to understand. For example, now time_now_after(a) means time now after a will return true.

---

---
include/linux/jiffies.h | 16 ++++++++++++++++

1 file changed, 16 insertions(+)
include/linux/jiffies.h +16 -0
diff -upr linux/include/linux/jiffies.h linux.new/include/linux/jiffies.h
@@ -135,6 +135,22 @@ static inline u64 get_jiffies_64(void)
 #define time_before_eq64(a,b)	time_after_eq64(b,a)
 
 /*
+ * These four macros compare jiffies and 'a' for convenience.
+ */
+
+/* time_now_after(a) return true if now (jiffies) is after a */
+#define time_now_after(a) time_after(jiffies, a)
+
+/* time_now_before(a) return true if now (jiffies) is before a */
+#define time_now_before(a) time_before(jiffies, a)
+
+/* time_now_after_eq(a) return true if now (jiffies) is after or equal to a */
+#define time_now_after_eq(a) time_after_eq(jiffies, a)
+
+/* time_now_before_eq(a) return true if now (jiffies) is before or equal to a */
+#define time_now_before_eq(a) time_before_eq(jiffies, a)
+
+/*
  * Have the 32 bit jiffies value wrap 5 minutes after boot
  * so jiffies wrap bugs show up earlier.
  */
#1 lore raw

Hi Dave, Andrew and all

include/linux/jiffies.h static inline u64 get_jiffies_64(void) · line 151

Quoted 11 lines · Dave Young
include/linux/jiffies.h +11 -0
+/* time_now_after(a) return true if now (jiffies) is after a */
+#define time_now_after(a) time_after(jiffies, a)
+
+/* time_now_before(a) return true if now (jiffies) is before a */
+#define time_now_before(a) time_before(jiffies, a)
+
+/* time_now_after_eq(a) return true if now (jiffies) is after or equal to a */
+#define time_now_after_eq(a) time_after_eq(jiffies, a)
+
+/* time_now_before_eq(a) return true if now (jiffies) is before or equal to a */
+#define time_now_before_eq(a) time_before_eq(jiffies, a)

How about even more obvious names like time_is_past(), time_is_future(), ...?

Sorry, I missed v1. Should have proposed that earlier.

Hannes

author
#2 lore raw
Quoted 5 lines · Johannes Weiner

Hi Dave, Andrew and all

Dave Young <[email protected]> writes:

Quoted 14 lines

include/linux/jiffies.h static inline u64 get_jiffies_64(void) · line 151

Quoted 11 lines
include/linux/jiffies.h +11 -0
+/* time_now_after(a) return true if now (jiffies) is after a */
+#define time_now_after(a) time_after(jiffies, a)
+
+/* time_now_before(a) return true if now (jiffies) is before a */
+#define time_now_before(a) time_before(jiffies, a)
+
+/* time_now_after_eq(a) return true if now (jiffies) is after or equal to a */
+#define time_now_after_eq(a) time_after_eq(jiffies, a)
+
+/* time_now_before_eq(a) return true if now (jiffies) is before or equal to a */
+#define time_now_before_eq(a) time_before_eq(jiffies, a)

How about even more obvious names like time_is_past(), time_is_future(), ...?

Thanks for comment.

Then how do we name the _eq version? IMHO, the time_now_* is enough.

Quoted 5 lines

Sorry, I missed v1. Should have proposed that earlier.

Hannes

Regards
dave

#3 lore raw
Quoted 9 lines
Quoted 5 lines

include/linux/jiffies.h static inline u64 get_jiffies_64(void) · line 151

Quoted 2 lines
include/linux/jiffies.h +2 -0
+/* time_now_before_eq(a) return true if now (jiffies) is before or equal to a */
+#define time_now_before_eq(a) time_before_eq(jiffies, a)

How about even more obvious names like time_is_past(), time_is_future(), ...?

Thanks for comment.

Then how do we name the _eq version? IMHO, the time_now_* is enough.

Why do you even need them. I don't see the point of *any* of these extra macros as they simply obfuscate code and hide what is actually going on. The initial macros were added because of the type safety and correct comparison rules being complex. They have a purpose.

Even if you want these you can use !time_future() if you don't want the _eq variants. Generally speaking drivers should be using timers not polled loops, and most of our loops comparing with jiffies want removing.

Alan

author
#4 lore raw
Quoted 8 lines · Alan Cox
Quoted 8 lines
Quoted 5 lines

include/linux/jiffies.h static inline u64 get_jiffies_64(void) · line 151

Quoted 2 lines
include/linux/jiffies.h +2 -0
+/* time_now_before_eq(a) return true if now (jiffies) is before or equal to a */
+#define time_now_before_eq(a) time_before_eq(jiffies, a)

How about even more obvious names like time_is_past(), time_is_future(), ...?

Thanks for comment.

Quoted 7 lines
Quoted 1 line

Then how do we name the _eq version? IMHO, the time_now_* is enough.

Why do you even need them. I don't see the point of *any* of these extra
macros as they simply obfuscate code and hide what is actually going on. The initial macros were added because of the type safety and correct comparison rules being complex. They have a purpose.

Yes, This has a purpose as well. You will find most of the usage of these macros are just compare some value with jiffies after doing some grep.

For these cases this adding will easy use and understand.

Quoted 5 lines

Even if you want these you can use !time_future() if you don't want the _eq variants. Generally speaking drivers should be using timers not polled loops, and most of our loops comparing with jiffies want removing.

IMO time_after is a confusing name actually, If I don't read the comment I will think it as time_future

To some extent I agree with you that the timer will be better.

Quoted 3 lines

Alan

Thanks

#5 lore raw
Quoted 2 lines

Yes, This has a purpose as well. You will find most of the usage of these macros are just compare some value with jiffies after doing some grep.

That is no reason to add more macros that hide how they do it, or make jiffies itself invisible to the grep search or variable usage finding tools (and for tickless that very thing is important)

Quoted 1 line

For these cases this adding will easy use and understand.

It hides information.

Alan

#6 lore raw
Quoted 6 lines · On Sun, 9 Mar 2008 11:08:08 +0000 Alan Cox <[email protected]> wrote:
Quoted 2 lines

Yes, This has a purpose as well. You will find most of the usage of these macros are just compare some value with jiffies after doing some grep.

That is no reason to add more macros that hide how they do it, or make jiffies itself invisible to the grep search or variable usage finding tools (and for tickless that very thing is important)

We could call them time_is_after_jiffies(), time_is_before_jiffies(), etc.

Those names are actually better than time_after_now(), etc.

#7 lore raw
Quoted 3 lines

We could call them time_is_after_jiffies(), time_is_before_jiffies(), etc.

Those names are actually better than time_after_now(), etc.

Agreed - and if you think its worth doing I bow to your expertise on buggy patch submissions..

author
#8 lore raw
In reply to Andrew Morton
Quoted 5 lines · Andrew Morton
Quoted 3 lines · On Sun, 9 Mar 2008 11:08:08 +0000 Alan Cox <[email protected]> wrote:
Quoted 2 lines

Yes, This has a purpose as well. You will find most of the usage of these macros are just compare some value with jiffies after doing some grep.

Quoted 9 lines
Quoted 3 lines

That is no reason to add more macros that hide how they do it, or make jiffies itself invisible to the grep search or variable usage finding tools (and for tickless that very thing is important)

We could call them time_is_after_jiffies(), time_is_before_jiffies(), etc.

Those names are actually better than time_after_now(), etc.

Andrew, Hannes, Alan,
thanks for your comment, I have send the new time_is_* patch as v3.

Hannes, It's more like your is_past/is_future but looks better when along with _eq

#9 lore raw

Hi Dave,

Quoted 2 lines · Dave Young

Hannes, It's more like your is_past/is_future but looks better when along with _eq

I agree. And it's better to have `jiffies' in the names, anyway.

Hannes

#10 lore raw
In reply to Alan Cox
Quoted 6 lines · On Sun, 9 Mar 2008 09:58:02 +0000 Alan Cox <[email protected]> wrote:
Quoted 6 lines
Quoted 5 lines

include/linux/jiffies.h static inline u64 get_jiffies_64(void) · line 151

Quoted 2 lines
include/linux/jiffies.h +2 -0
+/* time_now_before_eq(a) return true if now (jiffies) is before or equal to a */
+#define time_now_before_eq(a) time_before_eq(jiffies, a)

How about even more obvious names like time_is_past(), time_is_future(), ...?

Quoted 6 lines
Quoted 3 lines

Thanks for comment.

Then how do we name the _eq version? IMHO, the time_now_* is enough.

Why do you even need them. I don't see the point of *any* of these extra macros as they simply obfuscate code and hide what is actually going on.

Two reasons:

a) the existing macros are (I believe) a right royal pita. It's very hard to remember which order the args are supposed to be in.

So each time I see a time_foo() when reviewing a patch I have to go off and re-read the implementation then have a big think to check that they got it right (a sure sign of a poor interface, no)?

And I'm not the only one - people get this wrong fairly regularly.

b) around 90% of the usages of time_after() are to compare against jiffies!

The macros which Dave is developing aren't just less-error-prone, easier-to-review transformations - they offer higher-level functionality. Because time_after() is just a basic comparison operation, whereas time_now_before() is an *application* of that operation.

Trust me on this - they will lead to easier-to-review code and less bugs.

Quoted 2 lines

The initial macros were added because of the type safety and correct comparison rules being complex. They have a purpose.

They are hard to use and hard to review.

Quoted 3 lines

Even if you want these you can use !time_future() if you don't want the _eq variants. Generally speaking drivers should be using timers not polled loops, and most of our loops comparing with jiffies want removing.

#11 lore raw
Quoted 2 lines

a) the existing macros are (I believe) a right royal pita. It's very hard to remember which order the args are supposed to be in.

Its operator(first, second) - the C++ infix operators aren't available in C so there isn't much choice.

Really we should be trying to stamp out all these uses of jiffies tests so that we can make tickless work ever better.

Alan