Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]


Groups > linux.kernel > #1650963 > unrolled thread

Re: [PATCH v2] arm: eBPF JIT compiler

Started byAndrew Lunn <andrew@lunn.ch>
First post2017-05-26 01:30 +0200
Last post2017-05-30 21:00 +0200
Articles 5 — 2 participants

Back to article view | Back to linux.kernel

This discussion starts older than the indexed window; earlier articles aren't shown. The article labeled Started by below is the oldest one visible, not the original post.


Contents

  Re: [PATCH v2] arm: eBPF JIT compiler Andrew Lunn <andrew@lunn.ch> - 2017-05-26 01:30 +0200
    Re: [PATCH v2] arm: eBPF JIT compiler Shubham Bansal <illusionist.neo@gmail.com> - 2017-05-26 01:40 +0200
      Re: [PATCH v2] arm: eBPF JIT compiler Shubham Bansal <illusionist.neo@gmail.com> - 2017-05-26 01:40 +0200
        Re: [PATCH v2] arm: eBPF JIT compiler Shubham Bansal <illusionist.neo@gmail.com> - 2017-05-26 19:00 +0200
          Re: [PATCH v2] arm: eBPF JIT compiler Shubham Bansal <illusionist.neo@gmail.com> - 2017-05-30 21:00 +0200

#1650963 — Re: [PATCH v2] arm: eBPF JIT compiler

FromAndrew Lunn <andrew@lunn.ch>
Date2017-05-26 01:30 +0200
SubjectRe: [PATCH v2] arm: eBPF JIT compiler
Message-ID<tLamB-6m2-5@gated-at.bofh.it>
> Tested on ARMv7 with QEMU by me (Shubham Bansal).
> Tested on ARMv5 by Andrew Lunn (andrew@lunn.ch).
> Expected to work on ARMv6 as well, as its a part ARMv7 and part ARMv5.
> Although, a proper testing is not done for ARMv6.
> 
> Both of these testing are done with and without CONFIG_FRAME_POINTER
> separately for LITTLE ENDIAN machine.

Nope. I only tested it in the default configuration of mvebu_v5_defconfig.
Please change the 'Both' to 'Some'.

      Andrew

[toc] | [next] | [standalone]


#1650976

FromShubham Bansal <illusionist.neo@gmail.com>
Date2017-05-26 01:40 +0200
Message-ID<tLawi-6pJ-27@gated-at.bofh.it>
In reply to#1650963
Hi Andrew,

Oh. I didn't knew. Can you test it now and confirm it? I think it will
work and wouldn't take much of the time.

Please.
Best,
Shubham Bansal


On Fri, May 26, 2017 at 4:53 AM, Andrew Lunn <andrew@lunn.ch> wrote:
>> Tested on ARMv7 with QEMU by me (Shubham Bansal).
>> Tested on ARMv5 by Andrew Lunn (andrew@lunn.ch).
>> Expected to work on ARMv6 as well, as its a part ARMv7 and part ARMv5.
>> Although, a proper testing is not done for ARMv6.
>>
>> Both of these testing are done with and without CONFIG_FRAME_POINTER
>> separately for LITTLE ENDIAN machine.
>
> Nope. I only tested it in the default configuration of mvebu_v5_defconfig.
> Please change the 'Both' to 'Some'.
>
>       Andrew

[toc] | [prev] | [next] | [standalone]


#1650977

FromShubham Bansal <illusionist.neo@gmail.com>
Date2017-05-26 01:40 +0200
Message-ID<tLawi-6pJ-29@gated-at.bofh.it>
In reply to#1650976
Just to add. It a very very small part which deal with
CONFIG_FRAME_POINTER just one move instruction.
Best,
Shubham Bansal


On Fri, May 26, 2017 at 5:04 AM, Shubham Bansal
<illusionist.neo@gmail.com> wrote:
> Hi Andrew,
>
> Oh. I didn't knew. Can you test it now and confirm it? I think it will
> work and wouldn't take much of the time.
>
> Please.
> Best,
> Shubham Bansal
>
>
> On Fri, May 26, 2017 at 4:53 AM, Andrew Lunn <andrew@lunn.ch> wrote:
>>> Tested on ARMv7 with QEMU by me (Shubham Bansal).
>>> Tested on ARMv5 by Andrew Lunn (andrew@lunn.ch).
>>> Expected to work on ARMv6 as well, as its a part ARMv7 and part ARMv5.
>>> Although, a proper testing is not done for ARMv6.
>>>
>>> Both of these testing are done with and without CONFIG_FRAME_POINTER
>>> separately for LITTLE ENDIAN machine.
>>
>> Nope. I only tested it in the default configuration of mvebu_v5_defconfig.
>> Please change the 'Both' to 'Some'.
>>
>>       Andrew

[toc] | [prev] | [next] | [standalone]


#1651522

FromShubham Bansal <illusionist.neo@gmail.com>
Date2017-05-26 19:00 +0200
Message-ID<tLqKK-8no-11@gated-at.bofh.it>
In reply to#1650977
Hi Andrew,

Did you get the time to test the code with CONFIG_FRAME_POINTER? It
would be great if you could check if it works on ARMv5 so that Russell
can look at the patch.

Thanks.
Best,
Shubham Bansal


On Fri, May 26, 2017 at 5:06 AM, Shubham Bansal
<illusionist.neo@gmail.com> wrote:
> Just to add. It a very very small part which deal with
> CONFIG_FRAME_POINTER just one move instruction.
> Best,
> Shubham Bansal
>
>
> On Fri, May 26, 2017 at 5:04 AM, Shubham Bansal
> <illusionist.neo@gmail.com> wrote:
>> Hi Andrew,
>>
>> Oh. I didn't knew. Can you test it now and confirm it? I think it will
>> work and wouldn't take much of the time.
>>
>> Please.
>> Best,
>> Shubham Bansal
>>
>>
>> On Fri, May 26, 2017 at 4:53 AM, Andrew Lunn <andrew@lunn.ch> wrote:
>>>> Tested on ARMv7 with QEMU by me (Shubham Bansal).
>>>> Tested on ARMv5 by Andrew Lunn (andrew@lunn.ch).
>>>> Expected to work on ARMv6 as well, as its a part ARMv7 and part ARMv5.
>>>> Although, a proper testing is not done for ARMv6.
>>>>
>>>> Both of these testing are done with and without CONFIG_FRAME_POINTER
>>>> separately for LITTLE ENDIAN machine.
>>>
>>> Nope. I only tested it in the default configuration of mvebu_v5_defconfig.
>>> Please change the 'Both' to 'Some'.
>>>
>>>       Andrew

[toc] | [prev] | [next] | [standalone]


#1653494

FromShubham Bansal <illusionist.neo@gmail.com>
Date2017-05-30 21:00 +0200
Message-ID<tMUx3-2xi-9@gated-at.bofh.it>
In reply to#1651522
Hi Russell,

I tried everything I could to run a ARMv6 machine in last 4 days but I
am not able to. If nobody is willing to test the code no there
machine, may be we should not merge the code.
I have been more than happy to assist anyone who want to test it but
it looks like other people don't need it. I guess, it was a wasted
effort.

I don't have anything more to offer in regarding to this patch.
Thank you.

Best,
Shubham Bansal


On Fri, May 26, 2017 at 10:27 PM, Shubham Bansal
<illusionist.neo@gmail.com> wrote:
> Hi Andrew,
>
> Did you get the time to test the code with CONFIG_FRAME_POINTER? It
> would be great if you could check if it works on ARMv5 so that Russell
> can look at the patch.
>
> Thanks.
> Best,
> Shubham Bansal
>
>
> On Fri, May 26, 2017 at 5:06 AM, Shubham Bansal
> <illusionist.neo@gmail.com> wrote:
>> Just to add. It a very very small part which deal with
>> CONFIG_FRAME_POINTER just one move instruction.
>> Best,
>> Shubham Bansal
>>
>>
>> On Fri, May 26, 2017 at 5:04 AM, Shubham Bansal
>> <illusionist.neo@gmail.com> wrote:
>>> Hi Andrew,
>>>
>>> Oh. I didn't knew. Can you test it now and confirm it? I think it will
>>> work and wouldn't take much of the time.
>>>
>>> Please.
>>> Best,
>>> Shubham Bansal
>>>
>>>
>>> On Fri, May 26, 2017 at 4:53 AM, Andrew Lunn <andrew@lunn.ch> wrote:
>>>>> Tested on ARMv7 with QEMU by me (Shubham Bansal).
>>>>> Tested on ARMv5 by Andrew Lunn (andrew@lunn.ch).
>>>>> Expected to work on ARMv6 as well, as its a part ARMv7 and part ARMv5.
>>>>> Although, a proper testing is not done for ARMv6.
>>>>>
>>>>> Both of these testing are done with and without CONFIG_FRAME_POINTER
>>>>> separately for LITTLE ENDIAN machine.
>>>>
>>>> Nope. I only tested it in the default configuration of mvebu_v5_defconfig.
>>>> Please change the 'Both' to 'Some'.
>>>>
>>>>       Andrew

[toc] | [prev] | [standalone]


Back to top | Article view | linux.kernel


csiph-web