[yocto-ab] Fwd: Re: eglibc vs glibc
Khem Raj
raj.khem at gmail.com
Wed Dec 18 12:56:48 PST 2013
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
Thanks Sean for pointin out that my message went encrypted I was using
a less frequently used machine to reply. Here is forwarded message (
hopefully not crypted this time)
On 12/11/2013 1:53 PM, Hart, Darren wrote:
> On Wed, 2013-12-11 at 21:09 +0000, Richard Purdie wrote:
>> On Wed, 2013-12-11 at 11:59 -0500, William Mills wrote:
>>> From Linaro I understand that eglibc is nearing its last
>>> release. Linaro is planning the transition from eglibc to
>>> glibc. They are asking what gaps remain in glibc that are
>>> important to
eglibc
>>> users. We realize that embedded is the most likely where any
>>> gaps exist. Linaro toolchain builds eglibc but they turn on all
>>> the options so
they
>>> do not have a lot of experience on scaling down.
>>>
>>> Has oe-core already done or planned the transition to glibc?
>>
>> It has not done it and the transition is not planned as yet.
>>
>>> Have we identified any gaps?
>>
>> The biggest issue is the lack of config options to cut down the
>> size of eglibc.
>>
>>> Has any one tried yp-tiny with glibc?
>>
>> No, it won't work due to the issue above.
>
> An open work item here is to complete research into the limitations
> of uclibc to determine if the granular componentization of eglibc
> is really needed, or if uclibc might be a more appropriate option
> for poky-tiny.
I would say yes thats a good option to consider for tiny systems.
>
> Alternatively, given the complexity of working with the eglibc
> config options and integrating it into a product, another option
> would be to only make the division at individual libraries and use
> uclibc when truly tiny is required.
>
> Ultimately, I think this alternative is where we will land.
> Especially if the resources can't be scratched together to do the
> port ourselves.
>
> My 0.02, for whatever it's worth.
>
> -- Darren
>
>>
>>> Linaro toolchain group is asking if there are any gaps that
>>> need
to be
>>> closed. I don't know enough to guide them.
>>
>> The discussion was taken upstream:
>>
>> http://www.eglibc.org/archives/patches/msg01276.html
>>
>> but they've said no unless someone steps up to do the work. We're
>> kind of in stalemate at the moment as I understand it. Khem or
>> Mark may know more about the current status but I don't think
>> much as changed since those emails.
>>
The changes w.r.t kconfigs is on our plate yes and I have plans but less
of time towards it lately but I would make sure that it does not get
dropped once eglibc is merged into glibc. Some pieces of this
feature is outside of eglibc at present.
>> Cheers,
>>
>> Richard
>>
>>
>>
>>
>
- -------- Original Message --------
Subject: Re: [yocto-ab] eglibc vs glibc
Date: Wed, 18 Dec 2013 13:53:39 -0600
From: Sean Hudson <sean_hudson at mentor.com>
Organization: Mentor Graphics - Embedded Software Division
To: Khem Raj <raj.khem at gmail.com>
Khem,
Did you mean to send this to the AB list? It shows that only Bill and
Darren can decrypt it.
Regards,
Sean
Sean Hudson | Embedded Linux Architect & Member of Technical Staff
Mentor Graphics – Embedded Systems Division™
On 12/11/2013 09:45 PM, Khem Raj wrote:
> On 12/11/2013 1:53 PM, Hart, Darren wrote:
>> On Wed, 2013-12-11 at 21:09 +0000, Richard Purdie wrote:
>>> On Wed, 2013-12-11 at 11:59 -0500, William Mills wrote:
>>>> From Linaro I understand that eglibc is nearing its last
>>>> release. Linaro is planning the transition from eglibc to
>>>> glibc. They are asking what gaps remain in glibc that are
>>>> important to eglibc users. We realize that embedded is the
>>>> most likely where any gaps exist. Linaro toolchain builds
>>>> eglibc but they turn on all the options so they do not have a
>>>> lot of experience on scaling down.
>>>>
>>>> Has oe-core already done or planned the transition to glibc?
>>>
>>> It has not done it and the transition is not planned as yet.
>>>
>>>> Have we identified any gaps?
>>>
>>> The biggest issue is the lack of config options to cut down the
>>> size of eglibc.
>>>
>>>> Has any one tried yp-tiny with glibc?
>>>
>>> No, it won't work due to the issue above.
>>
>> An open work item here is to complete research into the
>> limitations of uclibc to determine if the granular
>> componentization of eglibc is really needed, or if uclibc might
>> be a more appropriate option for poky-tiny.
>
> I would say yes thats a good option to consider for tiny systems.
>
>>
>> Alternatively, given the complexity of working with the eglibc
>> config options and integrating it into a product, another option
>> would be to only make the division at individual libraries and
>> use uclibc when truly tiny is required.
>>
>> Ultimately, I think this alternative is where we will land.
>> Especially if the resources can't be scratched together to do the
>> port ourselves.
>>
>> My 0.02, for whatever it's worth.
>>
>> -- Darren
>>
>>>
>>>> Linaro toolchain group is asking if there are any gaps that
>>>> need to be closed. I don't know enough to guide them.
>>>
>>> The discussion was taken upstream:
>>>
>>> http://www.eglibc.org/archives/patches/msg01276.html
>>>
>>> but they've said no unless someone steps up to do the work.
>>> We're kind of in stalemate at the moment as I understand it.
>>> Khem or Mark may know more about the current status but I don't
>>> think much as changed since those emails.
>>>
>
> The changes w.r.t kconfigs is on our plate yes and I have plans but
> less of time towards it lately but I would make sure that it does
> not get dropped once eglibc is merged into glibc. Some pieces of
> this feature is outside of eglibc at present.
>
>>> Cheers,
>>>
>>> Richard
>>>
>>>
>>>
>>>
>>
>
> _______________________________________________ yocto-ab mailing
> list yocto-ab at yoctoproject.org
> https://lists.yoctoproject.org/listinfo/yocto-ab
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v2.0.22 (GNU/Linux)
Comment: Using GnuPG with Thunderbird - http://www.enigmail.net/
iEYEARECAAYFAlKyDBAACgkQuwUzVZGdMxRucQCbBKZLXIrfigOjcOpXsYCvLwpR
k9AAoI55VWV/QSP2dGkqit5xC4qWN690
=jkzF
-----END PGP SIGNATURE-----
-------------- next part --------------
A non-text attachment was scrubbed...
Name: 0x59545C11.asc
Type: application/pgp-keys
Size: 8598 bytes
Desc: not available
URL: <http://lists.yoctoproject.org/pipermail/yocto-ab/attachments/20131218/4a55b7f0/attachment.key>
More information about the yocto-ab
mailing list