Subject: Re: Revised futex(2) man page for review

Hi David,

On 03/31/2015 04:45 PM, Davidlohr Bueso wrote:
> On Sat, 2015-03-28 at 12:47 +0100, Peter Zijlstra wrote:
>
>> The condition is represented by the futex word, which is an address in
>> memory supplied to the futex() system call, and the value at this mem‐
>> ory location. (While the virtual addresses for the same memory in sep‐
>> arate processes may not be equal, the kernel maps them internally so
>> that the same memory mapped in different locations will correspond for
>> futex() calls.)
>>
>> When executing a futex operation that requests to block a thread, the
>> kernel will only block if the futex word has the value that the calling
>
> Given the use of "word", you should probably state right away that
> futexes are only 32bit.

So, I made the opening sentence here:

The condition is represented by the futex word, which is an
address in memory supplied to the futex() system call, and the
32-bit value at this memory location.

Okay?

Cheers,

Michael

--
Michael Kerrisk
Linux man-pages maintainer; http://www.kernel.org/doc/man-pages/
Linux/UNIX System Programming Training: http://man7.org/training/


2015-07-28 03:16:39

by Davidlohr Bueso

[permalink] [raw]
Subject: Re: Revised futex(2) man page for review

On Mon, 2015-07-27 at 13:10 +0200, Michael Kerrisk (man-pages) wrote:
> Hi David,
>
> On 03/31/2015 04:45 PM, Davidlohr Bueso wrote:
> > On Sat, 2015-03-28 at 12:47 +0100, Peter Zijlstra wrote:
> >
> >> The condition is represented by the futex word, which is an address in
> >> memory supplied to the futex() system call, and the value at this mem‐
> >> ory location. (While the virtual addresses for the same memory in sep‐
> >> arate processes may not be equal, the kernel maps them internally so
> >> that the same memory mapped in different locations will correspond for
> >> futex() calls.)
> >>
> >> When executing a futex operation that requests to block a thread, the
> >> kernel will only block if the futex word has the value that the calling
> >
> > Given the use of "word", you should probably state right away that
> > futexes are only 32bit.
>
> So, I made the opening sentence here:
>
> The condition is represented by the futex word, which is an
> address in memory supplied to the futex() system call, and the
> 32-bit value at this memory location.
>
> Okay?

I think we can still improve :)

I've re-read the whole first paragraphs, and have a few comments that
touch upon this specific wording. Lets see. You have:

> The futex() system call provides a method for waiting until a certain
> condition becomes true. It is typically used as a blocking construct
> in the context of shared-memory synchronization: The program implements
> the majority of the synchronization in user space, and uses one of
> operations of the system call when it is likely that it has to block
> for a longer time until the condition becomes true. The program uses
> another operation of the system call to wake anyone waiting for a par‐
> ticular condition.

I've rephrased the next paragraph. How about adding this to follow?

A futex is in essence a 32-bit user-space address. All futex operations and
conditions are governed by this variable, from now on referred to as 'futex
word'. As such, a futex is identified by the address in shared memory, which
may or may not be shared between different processes. For virtual memory, the
kernel will internally handle and resolve the later. This futex word, along
with the value at its the memory location, are supplied to the futex() system
call.

Feel free to reword however you think is better.

Thanks,
Davidlohr

Subject: Re: Revised futex(2) man page for review

Hi David,

On 07/28/2015 05:16 AM, Davidlohr Bueso wrote:
> On Mon, 2015-07-27 at 13:10 +0200, Michael Kerrisk (man-pages) wrote:
>> Hi David,
>>
>> On 03/31/2015 04:45 PM, Davidlohr Bueso wrote:
>>> On Sat, 2015-03-28 at 12:47 +0100, Peter Zijlstra wrote:
>>>
>>>> The condition is represented by the futex word, which is an address in
>>>> memory supplied to the futex() system call, and the value at this mem‐
>>>> ory location. (While the virtual addresses for the same memory in sep‐
>>>> arate processes may not be equal, the kernel maps them internally so
>>>> that the same memory mapped in different locations will correspond for
>>>> futex() calls.)
>>>>
>>>> When executing a futex operation that requests to block a thread, the
>>>> kernel will only block if the futex word has the value that the calling
>>>
>>> Given the use of "word", you should probably state right away that
>>> futexes are only 32bit.
>>
>> So, I made the opening sentence here:
>>
>> The condition is represented by the futex word, which is an
>> address in memory supplied to the futex() system call, and the
>> 32-bit value at this memory location.
>>
>> Okay?
>
> I think we can still improve :)
>
> I've re-read the whole first paragraphs, and have a few comments that
> touch upon this specific wording. Lets see. You have:
>
>> The futex() system call provides a method for waiting until a certain
>> condition becomes true. It is typically used as a blocking construct
>> in the context of shared-memory synchronization: The program implements
>> the majority of the synchronization in user space, and uses one of
>> operations of the system call when it is likely that it has to block
>> for a longer time until the condition becomes true. The program uses
>> another operation of the system call to wake anyone waiting for a par‐
>> ticular condition.
>
> I've rephrased the next paragraph. How about adding this to follow?
>
> A futex is in essence a 32-bit user-space address. All futex operations and
> conditions are governed by this variable, from now on referred to as 'futex
> word'. As such, a futex is identified by the address in shared memory, which
> may or may not be shared between different processes. For virtual memory, the
> kernel will internally handle and resolve the later. This futex word, along
> with the value at its the memory location, are supplied to the futex() system
> call.
>
> Feel free to reword however you think is better.


I agree with you that that second paragraph is a bit broken. But, like Heinrich,
I'm confused by this term "32-bit ... address".

I've rewritten the paragraph as:

A futex is a 32-bit value—referred to below as a futex word—whose
address is supplied to the futex() system call. (Futexes are
32-bits in size on all platforms, including 64-bit systems.) All
futex operations are governed by this value. In order to share a
futex between processes, the futex is placed in a region of
shared memory, created using (for example) mmap(2) or shmat(2).
(Thus the futex word may have different virtual addresses in dif‐
ferent processes, but these addresses all refer to the same loca‐
tion in physical memory.)

Maybe you still have some further improvements for the paragraph?

Cheers,

Michael

--
Michael Kerrisk
Linux man-pages maintainer; http://www.kernel.org/doc/man-pages/
Linux/UNIX System Programming Training: http://man7.org/training/

2015-07-28 17:53:17

by Davidlohr Bueso

[permalink] [raw]
Subject: Re: Revised futex(2) man page for review

On Tue, 2015-07-28 at 09:44 +0200, Michael Kerrisk (man-pages) wrote:
> Maybe you still have some further improvements for the paragraph?

Nah, this is fine enough. Looks good.

Thanks,
Davidlohr

Subject: Re: Revised futex(2) man page for review

On 07/28/2015 07:52 PM, Davidlohr Bueso wrote:
> On Tue, 2015-07-28 at 09:44 +0200, Michael Kerrisk (man-pages) wrote:
>> Maybe you still have some further improvements for the paragraph?
>
> Nah, this is fine enough. Looks good.

Okay. Thanks. I added a Reviewed-by: for you.

Cheers,

Michael


--
Michael Kerrisk
Linux man-pages maintainer; http://www.kernel.org/doc/man-pages/
Linux/UNIX System Programming Training: http://man7.org/training/