Meego Wiki
Views

SDK/Toolchains/Meetings/Notes20110114

From MeeGo wiki
< SDK | Toolchains | Meetings
Revision as of 12:23, 31 January 2011 by Kant (Talk | contribs)
(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)
Jump to: navigation, search

Contents

Date and Time

Friday 2011-01-14 @8.00 - 9.00 UTC

Agenda

  • Status
  • AOB

Status

  • Patch for RPM armv7hl and armv7nhl architecture exist (done by Carsten Munk)

https://build.meego.com/package/view_file?file=rpm-4.8.1-armv7hl-armv7nhl-support.patch&package=rpm&project=devel%3Ahardfp%3Abase&srcmd5=b3d6d743c38893b866ae5325d890cd08

ARM RunFast in glibc? Do we have a patch for glibc-2.11-12-g24c0bf7 and/or glibc-2.12.1 that enables ARM RunFast[1] mode by default anywhere? Would be good to push it along with hardfp while we're at it and getting things tested through. [1] http://infocenter.arm.com/help/index.jsp?topic=/com.arm.doc.ddi0274h/Chdiihcd.html

  • Juha made final patch
  • By default the RunFast is used
  • Floats are improved, doubles not change.
  • Negative: numbers close of 0 but still > 0 are interpreted as 0
  • We can take this together with glibc
  • Juha Kallioinen will submit the patch(?)
  • will be in glibc-2.12.1 (verify with Juha)
  • Comments from Austin:

austin: which version is the patch made against? austin: glibc2.12 building had failed in OBS for a long time under ARM (x86 is ok) austin: and so far, I am afraid we don't have enough time for glibc update. austin: so it is important to have this patch for 2.11

AOB

Comments from Carsten Munk to Communication material

  • The OBS name for this architecture will be 'armv7hl', so you may have to re-add repositories in orer

In "Default set"

  • NEON instructions are not allowed unless guarded, ie checks for NEON capability in /proc/self/auxv

In both:

  • Thumb not allowed (-mno-thumb)

Some questions to Carsten:

  • Why Thumb is not allowed?? Thumb should be allowed.
  • Cortex has been optimized for Thumb

In practice, the cross compilers has float-abi=hard hardcoded so maybe we should downplay that. You can't have mfloat-abi=hard with a softfp libgcc anyway.

Personal tools