Re: Status of pwritev2/preadv2 in glibc?

classic Classic list List threaded Threaded
9 messages Options
Reply | Threaded
Open this post in threaded view
|

Re: Status of pwritev2/preadv2 in glibc?

Christoph Hellwig
It would probably help to add the glibc list as the glibc developers
might know better..

On Sat, Apr 15, 2017 at 04:24:29PM +0000, Stephen  Bates wrote:
> Hi
>
> Does anyone know the status of support for the pwritev2/preadv2 system calls in glibc? I am doing some more IO polling testing and realized that the support is not in 2.25 and does not seem to be staged for later release?
>
> Right now I am using the FIO_HAVE_PWRITEV2 support that fio provides but it’d be nice to be able to write/adapt other user-space programs to test this capability…
>
> Stephen
>
>
---end quoted text---
Reply | Threaded
Open this post in threaded view
|

Re: Status of pwritev2/preadv2 in glibc?

Florian Weimer-5
On 04/21/2017 12:56 PM, Christoph Hellwig wrote:
> It would probably help to add the glibc list as the glibc developers
> might know better..

So far, no one has submitted a patch.

I hope the off_t parameter is passed exactly the same way as for pwritev
and its 64-bit variant, for all architectures.

Thanks,
Florian

Reply | Threaded
Open this post in threaded view
|

Re: Status of pwritev2/preadv2 in glibc?

Adhemerval Zanella-2


On 21/04/2017 08:19, Florian Weimer wrote:
> On 04/21/2017 12:56 PM, Christoph Hellwig wrote:
>> It would probably help to add the glibc list as the glibc developers
>> might know better..
>
> So far, no one has submitted a patch.
>
> I hope the off_t parameter is passed exactly the same way as for pwritev and its 64-bit variant, for all architectures.
>

It seems to have the same kernel abi indeed, I will work on this.  I think
it falls on the criteria for kernel syscalls wrappers [1].

[1] https://sourceware.org/glibc/wiki/Consensus?highlight=%28syscall%29#WIP:_Kernel_syscalls_wrappers
Reply | Threaded
Open this post in threaded view
|

Re: Status of pwritev2/preadv2 in glibc?

Florian Weimer-5
On 04/21/2017 04:18 PM, Adhemerval Zanella wrote:
> It seems to have the same kernel abi indeed, I will work on this.  I think
> it falls on the criteria for kernel syscalls wrappers [1].

We added the non-standard pwritev interface as a GNU API.  It may make
sense to do the same to pwritev2, for consistency.

Thanks for working on this,
Florian
Reply | Threaded
Open this post in threaded view
|

Re: Status of pwritev2/preadv2 in glibc?

Stephen  Bates
In reply to this post by Florian Weimer-5

> So far, no one has submitted a patch.

OK, unless I hear that someone else is working on one I will take a look at this.

> I hope the off_t parameter is passed exactly the same way as for pwritev
> and its 64-bit variant, for all architectures.

Duly noted.

For the kernel peeps I think this makes a case for revisiting the “big hammer” control for IO polling since it is going to be a while before applications can utilize the preadv2/pwritev2 approach… I will take a look at that too.

Stephen


Reply | Threaded
Open this post in threaded view
|

Re: Status of pwritev2/preadv2 in glibc?

Adhemerval Zanella-2


On 24/04/2017 14:55, Stephen  Bates wrote:
>
>> So far, no one has submitted a patch.
>
> OK, unless I hear that someone else is working on one I will take a look at this.

I am working on it btw [1]

>
>> I hope the off_t parameter is passed exactly the same way as for pwritev
>> and its 64-bit variant, for all architectures.
>
> Duly noted.
>
> For the kernel peeps I think this makes a case for revisiting the “big hammer” control for IO polling since it is going to be a while before applications can utilize the preadv2/pwritev2 approach… I will take a look at that too.
>
> Stephen

[1] https://sourceware.org/ml/libc-alpha/2017-04/msg00461.html
Reply | Threaded
Open this post in threaded view
|

Re: Status of pwritev2/preadv2 in glibc?

Adhemerval Zanella-2
In reply to this post by Stephen Bates


On 24/04/2017 14:55, Stephen  Bates wrote:

>
>> So far, no one has submitted a patch.
>
> OK, unless I hear that someone else is working on one I will take a look at this.
>
>> I hope the off_t parameter is passed exactly the same way as for pwritev
>> and its 64-bit variant, for all architectures.
>
> Duly noted.
>
> For the kernel peeps I think this makes a case for revisiting the “big hammer” control for IO polling since it is going to be a while before applications can utilize the preadv2/pwritev2 approach… I will take a look at that too.
>
> Stephen
>
>

I am working on it btw [1]

[1] https://sourceware.org/ml/libc-alpha/2017-04/msg00461.html

PS: resending with cc to all.
Reply | Threaded
Open this post in threaded view
|

Re: Status of pwritev2/preadv2 in glibc?

Stephen  Bates

> I am working on it btw [1]

Thanks for working on this Adhemerval!

> PS: resending with cc to all.

Can you cc linux-block when you submit the patchset?

Stephen

Reply | Threaded
Open this post in threaded view
|

Re: Status of pwritev2/preadv2 in glibc?

Adhemerval Zanella-2
On 25/04/2017 18:23, Stephen  Bates wrote:
>
>> I am working on it btw [1]
>
> Thanks for working on this Adhemerval!
>
>> PS: resending with cc to all.
>
> Can you cc linux-block when you submit the patchset?

Sure, I am just finishing the sanity tests on the supported architectures.