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

Re: Bug#882490: mips/mipsel: enforce --param ggc-min-expand=10 in dpkg-buildflags

Control: tags -1 moreinfo

[ CCing debian-mips@lists.debian.org for porters input. ]


On Thu, 2017-11-23 at 13:59:01 +0100, Mathieu Malaterre wrote:
> Package: dpkg
> Version:
> User: debian-mips@lists.debian.org
> Usertags: mips mipsel

> On Thu, Nov 23, 2017 at 1:14 PM, James Cowgill <jcowgill@debian.org> wrote:
> > On 23/11/17 11:42, Mathieu Malaterre wrote:
> >> YunQiang,
> >>
> >> Do you know of any drawbacks ?
> >
> > This has been raised before:
> > https://lists.debian.org/debian-mips/2016/10/msg00049.html
> >
> > I think it's generally a good idea. You wouldn't want to "enforce" this
> > - you would set it as an overridable default setting.
> >
> > I think the disadvantages are: compilation speed, I think there is no
> > official way to do this so you would have to hack the gcc source a bit,
> > and deviation from upstream / other distros.

Why isn't this considered a buildd issue, lacking enough memory? Or as
hinted in the discussion, perhaps a leak in gcc?

If this is gcc on MIPS just requiring more memory, why isn't such
value set as the default. Which would be overridable with the same
command-line option?

> It would be nice if the default CFLAGS/CXXFLAGS were updated on mips &
> mipsel (only) to include an additional gcc flag:
> `--param ggc-min-expand=10`
> This has caused multiple FTBFS in the past as for example: 849657 or 879636.

Adding such option, so that it can be easily disabled would require
adding an arch-specific feature in a feature area. Which is pretty
unusual. The problem with requiring the removal via STRIP variables is
that it would need to match the exact value in the default, even if it
might change in the future.

I'm a bit uncomfortable with this request TBH.


Reply to: