[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Minios-devel] [UNIKRAFT PATCHv4 16/43] uk/arch: Implement ukarch_find_lsbit for Arm64


  • To: Sharan Santhanam <sharan.santhanam@xxxxxxxxx>, Julien Grall <Julien.Grall@xxxxxxx>, "minios-devel@xxxxxxxxxxxxx" <minios-devel@xxxxxxxxxxxxx>
  • From: Wei Chen <Wei.Chen@xxxxxxx>
  • Date: Mon, 16 Jul 2018 03:06:35 +0000
  • Accept-language: en-US
  • Authentication-results: spf=none (sender IP is ) smtp.mailfrom=Wei.Chen@xxxxxxx;
  • Delivery-date: Mon, 16 Jul 2018 03:06:50 +0000
  • List-id: Mini-os development list <minios-devel.lists.xenproject.org>
  • Spamdiagnosticmetadata: NSPM
  • Spamdiagnosticoutput: 1:99
  • Thread-index: AQHUFQhp9Y0BNxdlDU6yyJQ+s26L9aSLmV4AgAFQiICAAAgpAIAAMm4AgAQShoA=
  • Thread-topic: [Minios-devel] [UNIKRAFT PATCHv4 16/43] uk/arch: Implement ukarch_find_lsbit for Arm64


> -----Original Message-----
> From: Sharan Santhanam <sharan.santhanam@xxxxxxxxx>
> Sent: 2018年7月13日 20:45
> To: Julien Grall <Julien.Grall@xxxxxxx>; Wei Chen <Wei.Chen@xxxxxxx>; minios-
> devel@xxxxxxxxxxxxx
> Subject: Re: [Minios-devel] [UNIKRAFT PATCHv4 16/43] uk/arch: Implement
> ukarch_find_lsbit for Arm64
>
> Hello,
>
>
> On 07/13/2018 11:44 AM, Julien Grall wrote:
> > Hi,
> >
> > On 13/07/18 10:20, Wei Chen wrote:
> >>>> +#ifndef __UKARCH_ATOMIC_H__
> >>>> +#error Do not include this header directly
> >>>> +#endif
> >>>> +
> >>>
> >>> For finding the lsbit shouldn't we use the builtin function
> >>> __builtin_ffs, __builtin_ffsl?
> >
> > AFAICT __builtin_ffs is GNU-ism. Do we really want to tie ourself to it?
>
> Yes, this is a good point I had overlooked previously. On further
> looking into it, this builtin function is supported on gcc[1] and
> clang[2]. We may need to consider what other compiler tool chains do we
> need to be compatible with.
>
> > Also, do you know if that is supported correctly on all GCC versions we
> > targets?
> The __builitin_ffs were introduced as a part of the gcc version3.3
> but the earliest documentation where it is been described is in 3.4 [1].
>
>
> >>>
> >>
> >> When I was implementing the ukarch_find_lsbit, I found both Arm32 and
> >> X86_64 were not using the builtin-functions. X86_64 is using "bsfq",
> >> Arm32 is using the same instructions as Arm64.
>
> I agree in terms of consistency, we may use the current implementation.
>
> I wanted to clarify if we made an explicit decision to provide our own
> implementation for the ffs.
>

Later, in other patch series, maybe we can have some method to detect the
__builitin_ffs supported status. If it's supported by current compiler we
can use it, otherwise we will use current implementation.

How do you think about it?

> > To be honest, I would keep the implement as it is.
> >
> > Cheers,
> >
>
> [1]gcc:
> https://gcc.gnu.org/onlinedocs/gcc-3.4.6/gcc/Other-Builtins.html#Other-
> Builtins
>
> [2] Clang:
> https://github.com/llvm-
> mirror/clang/blob/release_26/include/clang/Basic/Builtins.def
>
>
> Thanks & Regards
> Sharan
IMPORTANT NOTICE: The contents of this email and any attachments are 
confidential and may also be privileged. If you are not the intended recipient, 
please notify the sender immediately and do not disclose the contents to any 
other person, use it for any purpose, or store or copy the information in any 
medium. Thank you.
_______________________________________________
Minios-devel mailing list
Minios-devel@xxxxxxxxxxxxxxxxxxxx
https://lists.xenproject.org/mailman/listinfo/minios-devel

 


Rackspace

Lists.xenproject.org is hosted with RackSpace, monitoring our
servers 24x7x365 and backed by RackSpace's Fanatical Support®.