GDB 10.1 release: branching early July?

GDB 10.1 release: branching early July?

[Joel Brobecker]

Hi everyone,

Now that the GDB 9.2 release has been out for a few weeks, what
do you think about turning our attention to starting the GDB 10
release cycle?  I looked at the NEWS file, and there is a decent
number of new features.

Based on that, what do you think of aiming for branching during
the weekend of July 4th? That's 3 weeks from now. We would then
be creating the first pre-release at the same time if possible,
and then try to aim for a release 2-3 weeks afterwards.

Are there any changes that you think we should have in master
before we create the branch?

This is what is current in the 10.1 list on Bugzilla:

  * [unassigned]
    <PR gdb/25412> thread_info with duplicate ptid added to inferior thread list
    https://sourceware.org/bugzilla/show_bug.cgi?id=25412

    Reported by Simon, who started investigating a bit.
    Simon says the issue appeared after the "multi-target" patch
    series went in (5b6d1e4fa by Pedro).

    Simon - is that still an issue?

  * [RainerO]
    <PR gdb/25939> [10 regression] run fails with ICE on Solaris

    Not sure if this is still an issue or not. Rainer did push some
    patches to master back in May. To be confirmed.

I also know of the following issue which I think should get fixed
before we branch:

  * [AndrewB/TomT]
    QEMU / GDB compatibility on RISCV64 ELF (failure to fetch some registers)
    https://sourceware.org/pipermail/gdb-patches/2020-June/169393.html

Anything else you think we should include?

Thank you!
-- 
Joel

Re: GDB 10.1 release: branching early July?

[Simon Marchi]

On 2020-06-13 9:56 p.m., Joel Brobecker wrote:
> Hi everyone,
> 
> Now that the GDB 9.2 release has been out for a few weeks, what
> do you think about turning our attention to starting the GDB 10
> release cycle?  I looked at the NEWS file, and there is a decent
> number of new features.
> 
> Based on that, what do you think of aiming for branching during
> the weekend of July 4th? That's 3 weeks from now. We would then
> be creating the first pre-release at the same time if possible,
> and then try to aim for a release 2-3 weeks afterwards.

That's fine with me.

> Are there any changes that you think we should have in master
> before we create the branch?
> 
> This is what is current in the 10.1 list on Bugzilla:
> 
>   * [unassigned]
>     <PR gdb/25412> thread_info with duplicate ptid added to inferior thread list
>     https://sourceware.org/bugzilla/show_bug.cgi?id=25412
> 
>     Reported by Simon, who started investigating a bit.
>     Simon says the issue appeared after the "multi-target" patch
>     series went in (5b6d1e4fa by Pedro).
> 
>     Simon - is that still an issue?

Yes, Pedro has a patch series to address it here:

  https://sourceware.org/pipermail/gdb-patches/2020-April/167578.html

We need to get back to it and get it merged.

Simon

Re: GDB 10.1 release: branching early July?

[Rainer Orth]

Hi Joel,

> This is what is current in the 10.1 list on Bugzilla:

>   * [RainerO]
>     <PR gdb/25939> [10 regression] run fails with ICE on Solaris
>
>     Not sure if this is still an issue or not. Rainer did push some
>     patches to master back in May. To be confirmed.

unfortunately, it is, and I didn't have much success debugging it.  This
makes gdb useless on Solaris and was introduced by the multi-target
merge.  I'm completely unfamiliar with that code and will need all the
help I can get to get this resolved.

	Rainer

-- 
-----------------------------------------------------------------------------
Rainer Orth, Center for Biotechnology, Bielefeld University

Re: GDB 10.1 release: branching early July?

[Hannes Domani]

 Am Sonntag, 14. Juni 2020, 03:56:26 MESZ hat Joel Brobecker <brobecker@adacore.com> Folgendes geschrieben:

> Hi everyone,
>
> Now that the GDB 9.2 release has been out for a few weeks, what
> do you think about turning our attention to starting the GDB 10
> release cycle?  I looked at the NEWS file, and there is a decent
> number of new features.
>
> Based on that, what do you think of aiming for branching during
> the weekend of July 4th? That's 3 weeks from now. We would then
> be creating the first pre-release at the same time if possible,
> and then try to aim for a release 2-3 weeks afterwards.
>
> Are there any changes that you think we should have in master
> before we create the branch?
>
> This is what is current in the 10.1 list on Bugzilla:
>
>   * [unassigned]
>     <PR gdb/25412> thread_info with duplicate ptid added to inferior thread list
>     https://sourceware.org/bugzilla/show_bug.cgi?id=25412
>
>     Reported by Simon, who started investigating a bit.
>     Simon says the issue appeared after the "multi-target" patch
>     series went in (5b6d1e4fa by Pedro).
>
>     Simon - is that still an issue?
>
>   * [RainerO]
>     <PR gdb/25939> [10 regression] run fails with ICE on Solaris
>
>     Not sure if this is still an issue or not. Rainer did push some
>     patches to master back in May. To be confirmed.
>
> I also know of the following issue which I think should get fixed
> before we branch:
>
>   * [AndrewB/TomT]
>     QEMU / GDB compatibility on RISCV64 ELF (failure to fetch some registers)
>     https://sourceware.org/pipermail/gdb-patches/2020-June/169393.html
>
> Anything else you think we should include?

For Windows it would be good that win32/25302 is fixed:
Mismatching fstat() function calls in gdb_bfd_open() and cache_bstat()
https://sourceware.org/bugzilla/show_bug.cgi?id=25302

TomT had some patches here:
https://sourceware.org/legacy-ml/gdb-patches/2020-01/msg00386.html


Hannes

Re: GDB 10.1 release: branching early July?

[Philippe Waroquiers]

On Sat, 2020-06-13 at 18:56 -0700, Joel Brobecker wrote:
> Hi everyone,
> 
> Now that the GDB 9.2 release has been out for a few weeks, what
> do you think about turning our attention to starting the GDB 10
> release cycle?  I looked at the NEWS file, and there is a decent
> number of new features.
> 
> Based on that, what do you think of aiming for branching during
> the weekend of July 4th? That's 3 weeks from now. We would then
> be creating the first pre-release at the same time if possible,
> and then try to aim for a release 2-3 weeks afterwards.
> 
> Are there any changes that you think we should have in master
> before we create the branch?
I have a few things that IMO would be nice (or needed) in 10.1:

Default args for alias.
I have just sent the RFAv7 that implements
the prevention of an alias of an alias that has default args,
as suggested by Simon:
https://sourceware.org/pipermail/gdb-patches/2020-June/169467.html

There are 2 things still to look at for the 'exec-file-mismatch':

* Below fixes a regression due to using build-id in exec-file-mismatch:
  https://sourceware.org/pipermail/gdb-patches/2020-June/169467.html

* an opened PR for which some feedback is needed
   https://sourceware.org/bugzilla/show_bug.cgi?id=25475
  Note that possibly fixing this might change the user visible CLI.

Thanks
Philippe

NB: I also have an RFC Improve handling of breakpoints with multiple locations.
https://sourceware.org/pipermail/gdb-patches/2020-May/168921.html
but no feedback received yet, so better for GDB 11.


> 
> This is what is current in the 10.1 list on Bugzilla:
> 
>   * [unassigned]
>     <PR gdb/25412> thread_info with duplicate ptid added to inferior thread list
>     https://sourceware.org/bugzilla/show_bug.cgi?id=25412
> 
>     Reported by Simon, who started investigating a bit.
>     Simon says the issue appeared after the "multi-target" patch
>     series went in (5b6d1e4fa by Pedro).
> 
>     Simon - is that still an issue?
> 
>   * [RainerO]
>     <PR gdb/25939> [10 regression] run fails with ICE on Solaris
> 
>     Not sure if this is still an issue or not. Rainer did push some
>     patches to master back in May. To be confirmed.
> 
> I also know of the following issue which I think should get fixed
> before we branch:
> 
>   * [AndrewB/TomT]
>     QEMU / GDB compatibility on RISCV64 ELF (failure to fetch some registers)
>     https://sourceware.org/pipermail/gdb-patches/2020-June/169393.html
> 
> Anything else you think we should include?
> 
> Thank you!

Re: GDB 10.1 release: branching early July?

[Philippe Waroquiers]

I have corrected the link for the build-id exec-file-mismatch ...
Philippe

On Sun, 2020-06-14 at 18:53 +0200, Philippe Waroquiers via Gdb-patches wrote:
> On Sat, 2020-06-13 at 18:56 -0700, Joel Brobecker wrote:
> > Hi everyone,
> > 
> > Now that the GDB 9.2 release has been out for a few weeks, what
> > do you think about turning our attention to starting the GDB 10
> > release cycle?  I looked at the NEWS file, and there is a decent
> > number of new features.
> > 
> > Based on that, what do you think of aiming for branching during
> > the weekend of July 4th? That's 3 weeks from now. We would then
> > be creating the first pre-release at the same time if possible,
> > and then try to aim for a release 2-3 weeks afterwards.
> > 
> > Are there any changes that you think we should have in master
> > before we create the branch?
> I have a few things that IMO would be nice (or needed) in 10.1:
> 
> Default args for alias.
> I have just sent the RFAv7 that implements
> the prevention of an alias of an alias that has default args,
> as suggested by Simon:
> https://sourceware.org/pipermail/gdb-patches/2020-June/169467.html
> 
> There are 2 things still to look at for the 'exec-file-mismatch':
> 
> * Below fixes a regression due to using build-id in exec-file-mismatch:
>   https://sourceware.org/pipermail/gdb-patches/2020-June/169467.html

This is the correct link:
https://sourceware.org/pipermail/gdb-patches/2020-June/169287.html

> 
> * an opened PR for which some feedback is needed
>    https://sourceware.org/bugzilla/show_bug.cgi?id=25475
>   Note that possibly fixing this might change the user visible CLI.
> 
> Thanks
> Philippe
> 
> NB: I also have an RFC Improve handling of breakpoints with multiple locations.
> https://sourceware.org/pipermail/gdb-patches/2020-May/168921.html
> but no feedback received yet, so better for GDB 11.
> 
> 
> > This is what is current in the 10.1 list on Bugzilla:
> > 
> >   * [unassigned]
> >     <PR gdb/25412> thread_info with duplicate ptid added to inferior thread list
> >     https://sourceware.org/bugzilla/show_bug.cgi?id=25412
> > 
> >     Reported by Simon, who started investigating a bit.
> >     Simon says the issue appeared after the "multi-target" patch
> >     series went in (5b6d1e4fa by Pedro).
> > 
> >     Simon - is that still an issue?
> > 
> >   * [RainerO]
> >     <PR gdb/25939> [10 regression] run fails with ICE on Solaris
> > 
> >     Not sure if this is still an issue or not. Rainer did push some
> >     patches to master back in May. To be confirmed.
> > 
> > I also know of the following issue which I think should get fixed
> > before we branch:
> > 
> >   * [AndrewB/TomT]
> >     QEMU / GDB compatibility on RISCV64 ELF (failure to fetch some registers)
> >     https://sourceware.org/pipermail/gdb-patches/2020-June/169393.html
> > 
> > Anything else you think we should include?
> > 
> > Thank you!

Re: GDB 10.1 release: branching early July?

[Joel Brobecker]

> >   * [unassigned]
> >     <PR gdb/25412> thread_info with duplicate ptid added to inferior thread list
> >     https://sourceware.org/bugzilla/show_bug.cgi?id=25412
> > 
> >     Reported by Simon, who started investigating a bit.
> >     Simon says the issue appeared after the "multi-target" patch
> >     series went in (5b6d1e4fa by Pedro).
> > 
> >     Simon - is that still an issue?
> 
> Yes, Pedro has a patch series to address it here:
> 
>   https://sourceware.org/pipermail/gdb-patches/2020-April/167578.html
> 
> We need to get back to it and get it merged.

Thanks for the update and the URL, Simon. This is super helpful
for me to keep track of progress.

-- 
Joel

Re: GDB 10.1 release: branching early July?

[Joel Brobecker]

> >   * [RainerO]
> >     <PR gdb/25939> [10 regression] run fails with ICE on Solaris
> >
> >     Not sure if this is still an issue or not. Rainer did push some
> >     patches to master back in May. To be confirmed.
> 
> unfortunately, it is, and I didn't have much success debugging it.  This
> makes gdb useless on Solaris and was introduced by the multi-target
> merge.  I'm completely unfamiliar with that code and will need all the
> help I can get to get this resolved.

Boooo... Understood. Maybe start a dedicated discussion on gdb-patches
with your current findings? Maybe Pedro can give you some pointers,
to help us debug and fix this thing?

I see that you've already provided a bit of information in the PR
itself (https://sourceware.org/bugzilla/show_bug.cgi?id=25939), but
perhaps we could start with a summary of the situation, and what
you have found so far? For instance, IIUC, this affects debugging
of any program on Solaris. Are you still getting the same error?
Your observation about a large number of threads being created is
interesting, and perhaps deserves to be dug into a little deeper,
considering that you think this wasn't happening with a version
prior to the multitarget patch series?

-- 
Joel

Re: GDB 10.1 release: branching early July?

[Joel Brobecker]

Thanks for the feedback, Hannes.

Tom, what do you think? I remember reviewing those patches internally
at AdaCore, and they looked good to me. I will send my (positive)
feedback on those patches now...

> > Anything else you think we should include?
> 
> For Windows it would be good that win32/25302 is fixed:
> Mismatching fstat() function calls in gdb_bfd_open() and cache_bstat()
> https://sourceware.org/bugzilla/show_bug.cgi?id=25302
> 
> TomT had some patches here:
> https://sourceware.org/legacy-ml/gdb-patches/2020-01/msg00386.html



-- 
Joel

Re: GDB 10.1 release: branching early July?

[Joel Brobecker]

Hi Philippe,

> > I have a few things that IMO would be nice (or needed) in 10.1:
> > 
> > Default args for alias.
> > I have just sent the RFAv7 that implements
> > the prevention of an alias of an alias that has default args,
> > as suggested by Simon:
> > https://sourceware.org/pipermail/gdb-patches/2020-June/169467.html

Being a new feature, it's tough to justify waiting for it.. But
since we are looking at version 7 of the patches series, let's hope
that we can converge quickly on that one. There are some other issues
that look a bit difficult to resolve.

> > There are 2 things still to look at for the 'exec-file-mismatch':
> > 
> > * Below fixes a regression due to using build-id in exec-file-mismatch:
> https://sourceware.org/pipermail/gdb-patches/2020-June/169287.html

Indeed. It sounds useful to wait for that one.

> > * an opened PR for which some feedback is needed
> >    https://sourceware.org/bugzilla/show_bug.cgi?id=25475
> >   Note that possibly fixing this might change the user visible CLI.

This one is, erm, "interesting". I commented on the PR directly.
The short version is that I seem to be seeing the bug as being elsewhere,
so I asked for clarification. Regardless, at the moment, this seems
like this could be categorized as a relatively minor bug.

Still, worth keeping in our radar, since exec-file-mismatch is new
in GDB 10.

> > NB: I also have an RFC Improve handling of breakpoints with multiple locations.
> > https://sourceware.org/pipermail/gdb-patches/2020-May/168921.html
> > but no feedback received yet, so better for GDB 11.

I don't forsee any problems with the changes you propose. On the other
hand, I tend to agree with the general principle that it's safer to
make changes affecting existing output earlier in the release cycle.

-- 
Joel

Re: GDB 10.1 release: branching early July?

[Rainer Orth]

Hi Joel,

>> >   * [RainerO]
>> >     <PR gdb/25939> [10 regression] run fails with ICE on Solaris
>> >
>> >     Not sure if this is still an issue or not. Rainer did push some
>> >     patches to master back in May. To be confirmed.
>> 
>> unfortunately, it is, and I didn't have much success debugging it.  This
>> makes gdb useless on Solaris and was introduced by the multi-target
>> merge.  I'm completely unfamiliar with that code and will need all the
>> help I can get to get this resolved.
>
> Boooo... Understood. Maybe start a dedicated discussion on gdb-patches
> with your current findings? Maybe Pedro can give you some pointers,
> to help us debug and fix this thing?

I guess that's the only option: at the glacial rate I'm currently making
progress, a patch will be ready for GDB 11 at the earliest otherwise.

> I see that you've already provided a bit of information in the PR
> itself (https://sourceware.org/bugzilla/show_bug.cgi?id=25939), but
> perhaps we could start with a summary of the situation, and what
> you have found so far? For instance, IIUC, this affects debugging
> of any program on Solaris. Are you still getting the same error?

I've got to try that first, but assuming that the problem is in procfs.c
(and maybe sol-thread.c), it's unlikely to have gone away in the
meantime.

> Your observation about a large number of threads being created is
> interesting, and perhaps deserves to be dug into a little deeper,
> considering that you think this wasn't happening with a version
> prior to the multitarget patch series?

I've now found where those are coming from: the thread name "gdb worker"
visible in ps -L provided the necessary clue.  They are due to worker
threads now being enabled by default.  On my desktop with 4 cores and
hyperthreading, that's relatively harmless.  However, when I run gdb on
an 8-socket server with 160 cores total, the messages about thread
creation when debugging gdb are killing you ;-)  I traced this to
maint.c (update_thread_pool_size) where the count defaults to
std::thread::hardware_concurrency ().  Maybe a bit over-eager...

	Rainer

-- 
-----------------------------------------------------------------------------
Rainer Orth, Center for Biotechnology, Bielefeld University

Re: GDB 10.1 release: branching early July?

[Joel Brobecker]

> I've now found where those are coming from: the thread name "gdb worker"
> visible in ps -L provided the necessary clue.  They are due to worker
> threads now being enabled by default.  On my desktop with 4 cores and
> hyperthreading, that's relatively harmless.  However, when I run gdb on
> an 8-socket server with 160 cores total, the messages about thread
> creation when debugging gdb are killing you ;-)  I traced this to
> maint.c (update_thread_pool_size) where the count defaults to
> std::thread::hardware_concurrency ().  Maybe a bit over-eager...

Oh my....

I remember that this version of GDB makes parallel minimal reading
the default. I think you can throttle down the number of threads
using:

    maint set worker-threads N

-- 
Joel

Re: GDB 10.1 release: branching early July?

[Tom Tromey]

>>>>> "Joel" == Joel Brobecker <brobecker@adacore.com> writes:

Joel> Thanks for the feedback, Hannes.
Joel> Tom, what do you think? I remember reviewing those patches internally
Joel> at AdaCore, and they looked good to me. I will send my (positive)
Joel> feedback on those patches now...

I was waiting for Pedro's patches; not sure what happened there.
I'll resurrect his patches and combine them with mine.

Tom

Re: GDB 10.1 release: branching early July?

[Pedro Alves]

Hi Joel,

On 6/15/20 11:53 PM, Joel Brobecker wrote:
>>>   * [unassigned]
>>>     <PR gdb/25412> thread_info with duplicate ptid added to inferior thread list
>>>     https://sourceware.org/bugzilla/show_bug.cgi?id=25412
>>>
>>>     Reported by Simon, who started investigating a bit.
>>>     Simon says the issue appeared after the "multi-target" patch
>>>     series went in (5b6d1e4fa by Pedro).
>>>
>>>     Simon - is that still an issue?
>>
>> Yes, Pedro has a patch series to address it here:
>>
>>   https://sourceware.org/pipermail/gdb-patches/2020-April/167578.html
>>
>> We need to get back to it and get it merged.
> 
> Thanks for the update and the URL, Simon. This is super helpful
> for me to keep track of progress.
> 

I'm addressing the review comments of that series today.

Thanks,
Pedro Alves

Re: GDB 10.1 release: branching early July?

[Luis Machado]

On 6/13/20 10:56 PM, Joel Brobecker wrote:
> Hi everyone,
> 
> Now that the GDB 9.2 release has been out for a few weeks, what
> do you think about turning our attention to starting the GDB 10
> release cycle?  I looked at the NEWS file, and there is a decent
> number of new features.
> 
> Based on that, what do you think of aiming for branching during
> the weekend of July 4th? That's 3 weeks from now. We would then
> be creating the first pre-release at the same time if possible,
> and then try to aim for a release 2-3 weeks afterwards.
> 
> Are there any changes that you think we should have in master
> before we create the branch?
> 
> This is what is current in the 10.1 list on Bugzilla:
> 
>    * [unassigned]
>      <PR gdb/25412> thread_info with duplicate ptid added to inferior thread list
>      https://sourceware.org/bugzilla/show_bug.cgi?id=25412
> 
>      Reported by Simon, who started investigating a bit.
>      Simon says the issue appeared after the "multi-target" patch
>      series went in (5b6d1e4fa by Pedro).
> 
>      Simon - is that still an issue?
> 
>    * [RainerO]
>      <PR gdb/25939> [10 regression] run fails with ICE on Solaris
> 
>      Not sure if this is still an issue or not. Rainer did push some
>      patches to master back in May. To be confirmed.
> 
> I also know of the following issue which I think should get fixed
> before we branch:
> 
>    * [AndrewB/TomT]
>      QEMU / GDB compatibility on RISCV64 ELF (failure to fetch some registers)
>      https://sourceware.org/pipermail/gdb-patches/2020-June/169393.html
> 
> Anything else you think we should include?

I just reported one for 10.1, so we can make sure that is fixed before 
branch/release.

https://sourceware.org/bugzilla/show_bug.cgi?id=26175

Re: GDB 10.1 release: branching early July?

[Joel Brobecker]

Hi Luis,

> I just reported one for 10.1, so we can make sure that is fixed before
> branch/release.
> 
> https://sourceware.org/bugzilla/show_bug.cgi?id=26175

Thanks for bringing it to our attention. I agree this is something
we should certainly try to fix. Do you know if someone already
agreed to work on that?

-- 
Joel

Re: GDB 10.1 release: branching early July?

[Luis Machado]

Hi Joel,

On Fri, Jun 26, 2020, 13:26 Joel Brobecker <brobecker@adacore.com> wrote:

> Hi Luis,
>
> > I just reported one for 10.1, so we can make sure that is fixed before
> > branch/release.
> >
> > https://sourceware.org/bugzilla/show_bug.cgi?id=26175
>
> Thanks for bringing it to our attention. I agree this is something
> we should certainly try to fix. Do you know if someone already
> agreed to work on that?
>

No word yet, but I think Tom de Vries will get to it. I suppose he is
either busy or having days off. So I'm giving it a few days so he can chime
in.


>