Mirror of the gdb-patches mailing list
 help / color / mirror / Atom feed
From: Pedro Alves <palves@redhat.com>
To: Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>,
	Wei-min Pan <weimin.pan@oracle.com>
Cc: gdb-patches@sourceware.org
Subject: Re: Fix gdb 8.1 Solaris/SPARC compilation (PR build/22206)
Date: Thu, 28 Sep 2017 14:19:00 -0000	[thread overview]
Message-ID: <1c43ef2e-00cb-5ede-de6e-1e25fe7e09cd@redhat.com> (raw)
In-Reply-To: <yddr2urq9mi.fsf@CeBiTec.Uni-Bielefeld.DE>


On 09/28/2017 01:40 PM, Rainer Orth wrote:
> Hi Wei-min,
> 
>> As for the issues you raised on gdb ADI support for Solaris/SPARC in the
>> bug report,
>> it's never intended to add that support for Solaris. While that Solaris
>> provided a totally
>> different ADI API was a factor, for example, Linux made the information of
>> ADI block
>> size available in the ELF aux record, one needed to make a syscall on
>> Solaris to get that
>> piece of information, it's certainly doable.  It was never planned,
>> however, and most likely
>> will never be.
> 
> this is extremely unfortunate, Oracle adding ADI support for one OS, but
> not for its own ;-(
> 
> However, what's worse is the way in which ADI support was added: when
> you add code to a file shared between different OSes like
> sparc64-tdep.c, it's your responsibility to make sure that this code at
> least doesn't break other SPARC targets.  Given how much of the code
> there isn't actually shared (and not sharable, it seems), the approach
> you've take seems wrong to me: it should have been properly factored
> between shared (sparc64-tdep.c) and os-private
> (e.g. sparc64-linux-tdep.c) files to avoid such breakage in the first
> place.

To be fair, that is the sort of issue that should have been pointed out
in review.  When I pushed for moving the ADI support out of the nat 
files and into tdep files for cross debugging, I don't think I even
remembered Solaris was a thing...  Not that I have anything against
Solaris, to be clear.  It just didn't cross my mind.

To me, this indicates a few things:

 - It's great that the Solaris port is again seeing activity, and
   we should strive to make sure that SPARC changes consider it.

 - But also someone needs to continually keep an eye on Solaris
   lest it ends up forgotten and broken again.  We need an official
   Solaris port maintainer.  Any takers?  :-)

 - It'd be desirable to have a SPARC Solaris machine build slave
   in the GDB buildbot, so that we can offload part of the work to 
   bots.  I wonder whether Oracle can help with this?  Might be
   difficult with the whole Solaris situation...

Thanks,
Pedro Alves


  reply	other threads:[~2017-09-28 14:19 UTC|newest]

Thread overview: 9+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2017-09-26  9:33 Rainer Orth
2017-09-26 11:33 ` Pedro Alves
2017-09-26 13:04   ` Rainer Orth
2017-09-26 16:06     ` Wei-min Pan
2017-09-27  9:16       ` Rainer Orth
2017-09-27 20:43         ` Wei-min Pan
2017-09-28 12:40           ` Rainer Orth
2017-09-28 14:19             ` Pedro Alves [this message]
2017-09-28 14:37               ` Rainer Orth

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=1c43ef2e-00cb-5ede-de6e-1e25fe7e09cd@redhat.com \
    --to=palves@redhat.com \
    --cc=gdb-patches@sourceware.org \
    --cc=ro@CeBiTec.Uni-Bielefeld.DE \
    --cc=weimin.pan@oracle.com \
    /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