public inbox for binutils@sourceware.org
 help / color / mirror / Atom feed
* Restarting linking from write phase (PR 24373)
       [not found] <DB6PR0802MB2309ADD39E513C8A413AE43DFF310@DB6PR0802MB2309.eurprd08.prod.outlook.com>
@ 2019-05-08 10:38 ` Tamar Christina
  2019-06-06 11:46   ` Nick Clifton
  0 siblings, 1 reply; 3+ messages in thread
From: Tamar Christina @ 2019-05-08 10:38 UTC (permalink / raw)
  To: binutils; +Cc: nd

Resending to mailist

________________________________________
From: Tamar Christina
Sent: Tuesday, May 7, 2019 10:52:06 AM
To: binutils@sourceware.org
Cc: nd
Subject: Restarting linking from write phase (PR 24373)

Hi All,

I'm looking at PR 24373 and am trying to think of  a way to implement the space saving if possible but if not to use the existing method as a fall back.

Essentially the problem is that in order to have the space savings I need to know that all addresses are within the 1MB range for an ADR.
This means I can only know this at the very very end of the linking phase when we're about to start writing out the sections (after GC sections et all).
Now If turns out this is not the case I have to then resize the text section and change the alignment of it.  Which of course means recalculating all the VMAs.

So my question is, does anyone know if it's possible to do this? Or should I just add a new option that assumes the space saving is possible and errors out if it turns out an address is out of range.
In principle I don't have to restart the entire linking process, just the calculation one, as symbol map  doesn't change.

Thanks,
Tamar

^ permalink raw reply	[flat|nested] 3+ messages in thread

* Re: Restarting linking from write phase (PR 24373)
  2019-05-08 10:38 ` Restarting linking from write phase (PR 24373) Tamar Christina
@ 2019-06-06 11:46   ` Nick Clifton
  2019-06-07  8:11     ` Tamar Christina
  0 siblings, 1 reply; 3+ messages in thread
From: Nick Clifton @ 2019-06-06 11:46 UTC (permalink / raw)
  To: Tamar Christina, binutils; +Cc: nd

Hi Tamar,

> So my question is, does anyone know if it's possible to do this? Or should I just add a new option that assumes the space saving is possible and errors out if it turns out an address is out of range.

Given the silence so far, I would guess that restarting the link is not
really possible.  (Certainly that is my feeling for the situation).  
Hence going with the assumption option is probably the best way forward.

Cheers
  Nick


^ permalink raw reply	[flat|nested] 3+ messages in thread

* Re: Restarting linking from write phase (PR 24373)
  2019-06-06 11:46   ` Nick Clifton
@ 2019-06-07  8:11     ` Tamar Christina
  0 siblings, 0 replies; 3+ messages in thread
From: Tamar Christina @ 2019-06-07  8:11 UTC (permalink / raw)
  To: nickc, binutils; +Cc: nd

Thanks Nick,

I have indeed gone with that approach, but if someone has a way to get the best of both worlds I'd be more than happy to revisit it.

Cheers,
Tamar

________________________________________
From: Nick Clifton <nickc@redhat.com>
Sent: Thursday, June 6, 2019 12:46 PM
To: Tamar Christina; binutils@sourceware.org
Cc: nd
Subject: Re: Restarting linking from write phase (PR 24373)

Hi Tamar,

> So my question is, does anyone know if it's possible to do this? Or should I just add a new option that assumes the space saving is possible and errors out if it turns out an address is out of range.

Given the silence so far, I would guess that restarting the link is not
really possible.  (Certainly that is my feeling for the situation).
Hence going with the assumption option is probably the best way forward.

Cheers
  Nick


^ permalink raw reply	[flat|nested] 3+ messages in thread

end of thread, other threads:[~2019-06-07  8:11 UTC | newest]

Thread overview: 3+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
     [not found] <DB6PR0802MB2309ADD39E513C8A413AE43DFF310@DB6PR0802MB2309.eurprd08.prod.outlook.com>
2019-05-08 10:38 ` Restarting linking from write phase (PR 24373) Tamar Christina
2019-06-06 11:46   ` Nick Clifton
2019-06-07  8:11     ` Tamar Christina

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).