From: Mark Hatle <mark.hatle@windriver.com>
To: Martin Jansa <martin.jansa@gmail.com>
Cc: openembedded-core@lists.openembedded.org
Subject: Re: [jethro/master][PATCH 1/1] local.conf.sample: Disable image-prelink by default
Date: Mon, 26 Oct 2015 18:02:22 -0500 [thread overview]
Message-ID: <562EB0FE.9030705@windriver.com> (raw)
In-Reply-To: <20151026195647.GH2554@jama>
On 10/26/15 2:56 PM, Martin Jansa wrote:
> On Mon, Oct 26, 2015 at 02:18:12PM -0500, Mark Hatle wrote:
>> On 10/26/15 1:15 PM, Martin Jansa wrote:
>>> On Mon, Oct 26, 2015 at 10:30:54AM -0500, Mark Hatle wrote:
>>>> Due to problems with the prelinker itself, we need to disable the
>>>> image-prelink by default. This will hopefully be re-enabled in the near
>>>> future.
>>>
>>> Acked-by: Martin Jansa <Martin.Jansa@gmail.com>
>>>
>>> We had to disable it in our builds back in Dylan days.
>>
>> FYI this is working just fine in Dizzy. I've heard reports it works in Fido.
>> This is the first that I know it's universally broken. (It's possible there
>> were some issues in Fido, as I have not verified it either way.. but Dizzy
>> definitely works and passes expected testings.)
>
> The main reason why we disabled it back then was that it was causing
> debug symbols in -dbg packages not matching the prelinked binary in the
> image
>
> The size of dynamic relocation section(rel.dyn) was changed from 0027a0
> to 003b70 by prelinking. Hence, the code section offset was also changed
> from 015550 to 016920.
> So, we can not use the dwarf file (.debug/libc-2.14.1.so) made before
> prelinking.
>
> The number of relocations was 1296 when i use prelinked libraries.
> This value was same on non-prelink. But the number of relative
> relocations has decreased from 2296 to 328.
>
> Then we did some benchmarks on our target and they didn't show any
> significant difference with and without prelink, for 4 stages of our
> boot we got:
>
> 7.06 -> 7.08, 0.28%
> 5.63 -> 5.75, 2.09%
> 10.47 -> 10.53, 0.53%
> 26.11 -> 26.15, 0.15%
>
> So we have disabled it and never looked back.
If I get things working again, I'll look into that. I suspect that was fixed
shortly after, I remebmer there being some debug/dwarf patches going in around
that time frame.
--Mark
> Regards,
>
>>>> Signed-off-by: Mark Hatle <mark.hatle@windriver.com>
>>>> ---
>>>> meta/conf/local.conf.sample | 4 +++-
>>>> 1 file changed, 3 insertions(+), 1 deletion(-)
>>>>
>>>> diff --git a/meta/conf/local.conf.sample b/meta/conf/local.conf.sample
>>>> index 3ae24ab..9d1480e 100644
>>>> --- a/meta/conf/local.conf.sample
>>>> +++ b/meta/conf/local.conf.sample
>>>> @@ -128,7 +128,9 @@ EXTRA_IMAGE_FEATURES = "debug-tweaks"
>>>> # - 'image-swab' to perform host system intrusion detection
>>>> # NOTE: if listing mklibs & prelink both, then make sure mklibs is before prelink
>>>> # NOTE: mklibs also needs to be explicitly enabled for a given image, see local.conf.extended
>>>> -USER_CLASSES ?= "buildstats image-mklibs image-prelink"
>>>> +# NOTE: image-prelink is currently broken due to problems with the prelinker. It is advised
>>>> +# that you do NOT run the prelinker at this time.
>>>> +USER_CLASSES ?= "buildstats image-mklibs"
>>>>
>>>>
>>>> #
>>>> --
>>>> 1.9.3
>>>>
>>>> --
>>>> _______________________________________________
>>>> Openembedded-core mailing list
>>>> Openembedded-core@lists.openembedded.org
>>>> http://lists.openembedded.org/mailman/listinfo/openembedded-core
>>>
>>
>
next prev parent reply other threads:[~2015-10-26 23:02 UTC|newest]
Thread overview: 7+ messages / expand[flat|nested] mbox.gz Atom feed top
2015-10-26 15:30 [jethro/master][PATCH 0/1] Disable the image-prelink by default Mark Hatle
2015-10-26 15:30 ` [jethro/master][PATCH 1/1] local.conf.sample: Disable " Mark Hatle
2015-10-26 18:15 ` Martin Jansa
2015-10-26 19:18 ` Mark Hatle
2015-10-26 19:56 ` Martin Jansa
2015-10-26 23:02 ` Mark Hatle [this message]
2015-10-30 4:48 ` Khem Raj
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=562EB0FE.9030705@windriver.com \
--to=mark.hatle@windriver.com \
--cc=martin.jansa@gmail.com \
--cc=openembedded-core@lists.openembedded.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox