Quantcast

Bug#850887: Decide proper solution for binutils' mips* bug

Previous Topic Next Topic
 
classic Classic list List threaded Threaded
13 messages Options
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Bug#850887: Decide proper solution for binutils' mips* bug

Lisandro Damián Nicanor Pérez Meyer-2
Package: tech-ctte
Severity: normal

Hi! Before stating the issue at hand I would like to address the fact that
this issue needs a quick solution as many packages are not able to
enter testing and we are near the freeze.

I would like the TC to address a solution for bug #844227. It's a binutils
bug that makes lots of packages FTBFS on mips*.

There are currently two proposed patches upstream, but upstream would like
to find a generic solution before considering accepting them, and encouraged
us to use either of them in the meantime [comment12].

[comment12] <https://sourceware.org/bugzilla/show_bug.cgi?id=20828#c12>

I proposed a patch that exclusively works on mips* to leave aside possible
bugs for other archs, see [patch].

[patch] <https://bugs.debian.org/cgi-bin/bugreport.cgi?
att=1;bug=844227;filename=844227_workaround_v2.patch;msg=144>

Yes, I happened to miss Matthias' request to not upload the patch due to a bad
spam filter, and I would like to use this opportunity to present once again
my apologies for that.

Problem is that we still have the bug affecting us.
binutils maintainer states in [m149]:

  [..] I'm fine to apply work arounds for port architectures, but not
  for release architectures (I didn't decide on this status).  The binutils
update
  plan was announced last June [1], and I plan to stick to it.  At least one
of
  the mips toolchain maintainers (out of the five who committed to in the
  architecture qualification process) seems to address RC issues, and
according to
  the upstream issue, there's work in progress.

[m149] <https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=844227#149>

While I do totally agree with Matthias that an upstream-approved fix is indeed
the best possible way out of this, in the meantime we have been dealing with
this bug almost two months now, and it's hindering the effort of many other
package maintainers not being able to get their packages into testing on
time for the release.

I would also like to note that I could not fully test my proposed patch
because
I can build the package on a porterbox but not use it later to build other
stuff. Of course I'll be more than happy to do this if there is a way for me
to do it.

So my question is: how can we fix this issue? Should we ask binutils
maintainer
to apply the patch (or let someone else do it), should we leave this matter
as it is or should we take any other further way of action?

Kinds regards, Lisandro.

-- System Information:
Debian Release: stretch/sid
  APT prefers unstable
  APT policy: (990, 'unstable'), (500, 'unstable-debug'), (500, 'testing-
debug'), (500, 'buildd-unstable'), (500, 'testing'), (500, 'stable'), (1,
'experimental-debug')
Architecture: amd64 (x86_64)
Foreign Architectures: i386

Kernel: Linux 4.8.0-2-amd64 (SMP w/2 CPU cores)
Locale: LANG=es_AR.UTF-8, LC_CTYPE=es_AR.UTF-8 (charmap=UTF-8)
Shell: /bin/sh linked to /bin/bash
Init: systemd (via /run/systemd/system)

Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Bug#850887: Decide proper solution for binutils' mips* bug

Sam Hartman-3

Hi.
I'd really appreciate comments from debian-release on this issue.
Would debian-release like us to take this up?
If so, I have a proposal for how to fast-track this situation, but I am
only comfortable doing that if the release team is involved.

Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Bug#850887: Decide proper solution for binutils' mips* bug

Lisandro Damián Nicanor Pérez Meyer-2
In reply to this post by Lisandro Damián Nicanor Pérez Meyer-2
I would like to mention that in my previous mail I forgot another possible fix
for this bug: going back to binutils 2.27. The current uploads are snapshots
of the next version, 2.28, and whereas the bug in question seems to habe been
there all the time it started to affect us since a commit in the to-be 2.28
version.

I've asked Matthias in [msg217] if there are any technical reasons to really
go for 2.28 instead of 2.27. So far I did not get a reply to that.

[msg217] <https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=844227#217>

I've taken a look at binutil's [changelog] and there seems to be Debian-
related bugs closed there, but it's not up to me to say wether the mips* bug
has more precedence or not over the others.

[changelog] <http://metadata.ftp-master.debian.org/changelogs/main/b/binutils/
binutils_2.27.90.20170109-1_changelog>

Thanks, Lisandro.

--
Lisandro Damián Nicanor Pérez Meyer
http://perezmeyer.com.ar/
http://perezmeyer.blogspot.com/

signature.asc (849 bytes) Download Attachment
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Bug#850887: TC Involvement: MIPS and binutils

Sam Hartman-3
In reply to this post by Sam Hartman-3


Hi.

As you are probably aware, the question of what to do about linking on
mips and stretch has been referred to the TC.
There's a reasonable probability that we're going to want to move very
quickly on this issue, and I wanted to reach out to you and see how we
could best work with you to collect your input.

I'd be happy to set up an IRC discussion, to set up a phone call, etc.
I think that might work better than an email discussion, because the
email discussion might involve a number of round trips.
I'd be happy to work one-on-one and summarize results/provide logs back
to the entire TC, or to set up something open to as many people as we
can.
Also if there's a TC member you'd rather work with than me, I'm sure
we'd be happy to facilitate this.

I'm hoping that you will be able to quickly work with us to understand
this issue and your position.

Thanks for your consideration,

--Sam

Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Bug#850887: TC Involvement: MIPS and binutils

Lisandro Damián Nicanor Pérez Meyer-2
On miércoles, 11 de enero de 2017 09:39:25 ART Sam Hartman wrote:

> Hi.
>
> As you are probably aware, the question of what to do about linking on
> mips and stretch has been referred to the TC.
> There's a reasonable probability that we're going to want to move very
> quickly on this issue, and I wanted to reach out to you and see how we
> could best work with you to collect your input.
>
> I'd be happy to set up an IRC discussion, to set up a phone call, etc.
> I think that might work better than an email discussion, because the
> email discussion might involve a number of round trips.
> I'd be happy to work one-on-one and summarize results/provide logs back
> to the entire TC, or to set up something open to as many people as we
> can.
> Also if there's a TC member you'd rather work with than me, I'm sure
> we'd be happy to facilitate this.
>
> I'm hoping that you will be able to quickly work with us to understand
> this issue and your position.
Hi Sam! I think an IRC discussion will be the best choice here as my phone
lines are really not reliable at all :-(

I'll be online from 17:30 UTC onwards, nick lisandro on freenode.

I can ping you if I get to IRC sooner.

Kinds regards, Lisandro.

--
Why should I care about posterity?
What's posterity ever done for me?
  -- Groucho Marx

Lisandro Damián Nicanor Pérez Meyer
http://perezmeyer.com.ar/
http://perezmeyer.blogspot.com/

signature.asc (849 bytes) Download Attachment
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Bug#850887: TC Involvement: MIPS and binutils

Lisandro Damián Nicanor Pérez Meyer-2
On miércoles, 11 de enero de 2017 11:49:48 ART Lisandro Damián Nicanor Pérez
Meyer wrote:
[snip]
> I'll be online from 17:30 UTC onwards, nick lisandro on freenode.

And also oftc, of course.

--
Lisandro Damián Nicanor Pérez Meyer
http://perezmeyer.com.ar/
http://perezmeyer.blogspot.com/

signature.asc (849 bytes) Download Attachment
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Bug#850887: TC Involvement: MIPS and binutils

Sam Hartman-3
In reply to this post by Lisandro Damián Nicanor Pérez Meyer-2
>>>>> "Lisandro" == Lisandro Damián Nicanor Pérez Meyer <[hidden email]> writes:

    Lisandro> On miércoles, 11 de enero de 2017 09:39:25 ART Sam Hartman wrote:
    >> Hi.
    >>
    >> As you are probably aware, the question of what to do about
    >> linking on mips and stretch has been referred to the TC.  There's
    >> a reasonable probability that we're going to want to move very
    >> quickly on this issue, and I wanted to reach out to you and see
    >> how we could best work with you to collect your input.
    >>
    >> I'd be happy to set up an IRC discussion, to set up a phone call,
    >> etc.  I think that might work better than an email discussion,
    >> because the email discussion might involve a number of round
    >> trips.  I'd be happy to work one-on-one and summarize
    >> results/provide logs back to the entire TC, or to set up
    >> something open to as many people as we can.  Also if there's a TC
    >> member you'd rather work with than me, I'm sure we'd be happy to
    >> facilitate this.
    >>
    >> I'm hoping that you will be able to quickly work with us to
    >> understand this issue and your position.

    Lisandro> Hi Sam! I think an IRC discussion will be the best choice
    Lisandro> here as my phone lines are really not reliable at all :-(

    Lisandro> I'll be online from 17:30 UTC onwards, nick lisandro on
    Lisandro> freenode.

that was to doko not you.
I'd be happy to chat, but you've articulated your position fairly well.
If there's stuff not in the bug you'd like me to know about I'd be happy
to set things up, but from the bug logs, your position seems fairly
simple.
Let's see if my summary is accurate:

* This bug is creating a number of ftbfses, particularly for
  larger//more complex libraries on mips.

* You have a preferred minimal work-around you tried to upload

* Doko requested you not upload something until it was patched upstream.

* You want a solution sooner than that.

Is that approximately correct?

Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Bug#850887: TC Involvement: MIPS and binutils

Lisandro Damián Nicanor Pérez Meyer-2
On miércoles, 11 de enero de 2017 10:35:37 ART Sam Hartman wrote:
[snip]
> that was to doko not you.

Usual me I'm afraid. I suscribed to the bug and forgot to see the To field.

> I'd be happy to chat, but you've articulated your position fairly well.
> If there's stuff not in the bug you'd like me to know about I'd be happy
> to set things up, but from the bug logs, your position seems fairly
> simple.
> Let's see if my summary is accurate:
>
> * This bug is creating a number of ftbfses, particularly for
>   larger//more complex libraries on mips.
>
> * You have a preferred minimal work-around you tried to upload
>
> * Doko requested you not upload something until it was patched upstream.
>
> * You want a solution sooner than that.
>
> Is that approximately correct?
Almost, and at least it's good that this happened. I don't have any
preferences between both proposed workarounds/patches, I simply took the
easiest one and added code to limit it to mips*. As long as we fix this I have
no real position on how we do it.

Thanks for your work!

--
http://www.tiraecol.net/modules/comic/comic.php?content_id=162

Lisandro Damián Nicanor Pérez Meyer
http://perezmeyer.com.ar/
http://perezmeyer.blogspot.com/

signature.asc (849 bytes) Download Attachment
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Bug#850887: Decide proper solution for binutils' mips* bug

Julien Cristau-6
In reply to this post by Sam Hartman-3
On Tue, Jan 10, 2017 at 16:56:05 -0500, Sam Hartman wrote:

>
> Hi.
> I'd really appreciate comments from debian-release on this issue.
> Would debian-release like us to take this up?
> If so, I have a proposal for how to fast-track this situation, but I am
> only comfortable doing that if the release team is involved.
>
Hi Sam,

I was actually about to involve the TC (from a release management
perspective) when Lisandro did.

Here's a timeline I think summarizes events ttbomk:
- on 31 October 2016, the release team finalized the list of release
  architectures for stretch, including mips, mipsel and mips64el
- on 2 November 2016, the binutils maintainer switch from the upstream
  2.27 branch to upstream trunk, which caused a number of regressions
- on 5 November 2016 was the transition freeze for stretch, which is
  intended to reduce the amount of churn affecting many packages at once
- one of the regressions is still unfixed to this day, and blocks a
  number of package migrations to testing, including library transitions
  and RC bug fixes, to the point that if it doesn't get fixed in the
  next few days the options are to either delay the stretch freeze
  (planned for 5 February) or drop three architectures from stretch; I
  feel like a freeze delay might end up being necessary due to this bug
  anyway, even if it does get fixed now
- early this week Lisandro finally NMUed with a patch for this bug, only
  to be promptly reverted by the maintainer

I think it's way past time we fixed this, to avoid any further harm to
the stretch release.  That may mean reverting to the 2.27 branch,
reverting specific changes from our 2.28 branch snapshot, or applying a
proposed fix ahead of upstream, I'm not picky about specifics.  Help
from the TC in getting to a quick resolution would be very much welcome.

Thanks,
Julien

signature.asc (849 bytes) Download Attachment
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Bug#850887: Decide proper solution for binutils' mips* bug

Lisandro Damián Nicanor Pérez Meyer-2
In reply to this post by Sam Hartman-3
I would like to point out that it would be preferable if, in case a patch is
preferable over going back to the last know version to work, either Matthias
or a mips porter points out which of the two proposed patches is preferable.

For the time being I'm testing the patch I submited to the bug, but I have no
preference over any of them (nor technical grounds to discuss).

Thanks, Lisandro.

--
Lisandro Damián Nicanor Pérez Meyer
http://perezmeyer.com.ar/
http://perezmeyer.blogspot.com/

signature.asc (849 bytes) Download Attachment
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Bug#850887: Decide proper solution for binutils' mips* bug

Sam Hartman-3
As a FYI, Matthias wrote to me in IRC just now indicating that  he plans
to upload a patch in the next couple of days.
(He needs to get to the location where he has the right environment
before preparing the upload).

As such, I'm planning on holding off on calling for any votes.

Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Bug#850887: Decide proper solution for binutils' mips* bug

James Cowgill-4
In reply to this post by Lisandro Damián Nicanor Pérez Meyer-2
Hi,

On 12/01/17 14:54, Lisandro Damián Nicanor Pérez Meyer wrote:
> I would like to point out that it would be preferable if, in case a patch is
> preferable over going back to the last know version to work, either Matthias
> or a mips porter points out which of the two proposed patches is preferable.
>
> For the time being I'm testing the patch I submited to the bug, but I have no
> preference over any of them (nor technical grounds to discuss).

Both patches posted in the upstream bug should work. The first one fixes
a bug in the MIPS back end so that local symbols are sorted before
global symbols. This is probably the safer (although larger) patch
because it only touches the MIPS back end to try and bring it into line
with other architectures. The second patch prevents the questionable
local symbols from every appearing (so no sorting is necessary). This
should also be correct, although it will visibly change the contents of
the dynamic symbol table on all arches so I am slightly more
apprehensive because of that.

Side note: the patch you uploaded is not totally correct because it
isn't applied when building cross binutils (__mips__ will not be defined
there).

Thanks,
James


signature.asc (836 bytes) Download Attachment
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Bug#850887: marked as done (Decide proper solution for binutils' mips* bug)

Debian Bug Tracking System
In reply to this post by Lisandro Damián Nicanor Pérez Meyer-2
Your message dated Wed, 19 Apr 2017 15:27:13 -0400
with message-id <[hidden email]>
and subject line MIPS binutils for stretch
has caused the Debian Bug report #850887,
regarding Decide proper solution for binutils' mips* bug
to be marked as done.

This means that you claim that the problem has been dealt with.
If this is not the case it is now your responsibility to reopen the
Bug report if necessary, and/or fix the problem forthwith.

(NB: If you are a system administrator and have no idea what this
message is talking about, this may indicate a serious mail system
misconfiguration somewhere. Please contact [hidden email]
immediately.)


--
850887: http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=850887
Debian Bug Tracking System
Contact [hidden email] with problems

Package: tech-ctte
Severity: normal

Hi! Before stating the issue at hand I would like to address the fact that
this issue needs a quick solution as many packages are not able to
enter testing and we are near the freeze.

I would like the TC to address a solution for bug #844227. It's a binutils
bug that makes lots of packages FTBFS on mips*.

There are currently two proposed patches upstream, but upstream would like
to find a generic solution before considering accepting them, and encouraged
us to use either of them in the meantime [comment12].

[comment12] <https://sourceware.org/bugzilla/show_bug.cgi?id=20828#c12>

I proposed a patch that exclusively works on mips* to leave aside possible
bugs for other archs, see [patch].

[patch] <https://bugs.debian.org/cgi-bin/bugreport.cgi?
att=1;bug=844227;filename=844227_workaround_v2.patch;msg=144>

Yes, I happened to miss Matthias' request to not upload the patch due to a bad
spam filter, and I would like to use this opportunity to present once again
my apologies for that.

Problem is that we still have the bug affecting us.
binutils maintainer states in [m149]:

  [..] I'm fine to apply work arounds for port architectures, but not
  for release architectures (I didn't decide on this status).  The binutils
update
  plan was announced last June [1], and I plan to stick to it.  At least one
of
  the mips toolchain maintainers (out of the five who committed to in the
  architecture qualification process) seems to address RC issues, and
according to
  the upstream issue, there's work in progress.

[m149] <https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=844227#149>

While I do totally agree with Matthias that an upstream-approved fix is indeed
the best possible way out of this, in the meantime we have been dealing with
this bug almost two months now, and it's hindering the effort of many other
package maintainers not being able to get their packages into testing on
time for the release.

I would also like to note that I could not fully test my proposed patch
because
I can build the package on a porterbox but not use it later to build other
stuff. Of course I'll be more than happy to do this if there is a way for me
to do it.

So my question is: how can we fix this issue? Should we ask binutils
maintainer
to apply the patch (or let someone else do it), should we leave this matter
as it is or should we take any other further way of action?

Kinds regards, Lisandro.

-- System Information:
Debian Release: stretch/sid
  APT prefers unstable
  APT policy: (990, 'unstable'), (500, 'unstable-debug'), (500, 'testing-
debug'), (500, 'buildd-unstable'), (500, 'testing'), (500, 'stable'), (1,
'experimental-debug')
Architecture: amd64 (x86_64)
Foreign Architectures: i386

Kernel: Linux 4.8.0-2-amd64 (SMP w/2 CPU cores)
Locale: LANG=es_AR.UTF-8, LC_CTYPE=es_AR.UTF-8 (charmap=UTF-8)
Shell: /bin/sh linked to /bin/bash
Init: systemd (via /run/systemd/system)


The committee would like to thank everyone who was involved in this
discussion for helping us get to a mips binutils that is working for
stretch.

It sounded like Matthias  wanted to add some extra thoughts going
forward.  In closing the bug we're not closing off any future discussion
simply noting that this is no longer an issue before the committee.

As an individual, I note that this is the second significant issue
brought before the TC in the stretch time frame with regard to the
toolchain.
Communication was somewhat challenging on both issues.
The project as a whole might want to consider how we can better balance
these issues and get folks working on the toolchain the resources they
need both to produce an effective toolchain and to communicate with the
rest of the project.
The committee is of course available for a resource if useful in such a
discussion.
Loading...