System Software Telecon Minutes

[Mitch Bradley]

System SW Meeting Minutes
2006-11-07

  Jim (moderator),  Mitch (scribe),  Chris Blizzard,  Richard,
  Ray,  Ted,  Chris Ball,  Zephaniah,  Jordan,  Andres
  MFoster,  Dave Woodhouse,  Marcello
  Alan and Ashwath  from Marvell

=== Agenda ===
Priorities:
  - Get all the hardware working
  - Get everything integrated
  - Get everything stable
  - Cull what isn't stable
  - Something to demonstrate the camera and video
  - Stabilize the X setup


Cafe:

David hasn't tested recent hardware mods for power-off SD card detect,
we are not going to change the hardware for B1.

NAND is good, apart from having to erase complete with skip_bbt trick
if you have ever written the FLASH with old LInux

Chris tested NAND, it worked well.  Chris thinks NAND is 3x the speed
of SD under load.  3 MiB/sec vs 1 MiB.  We think we have used the
SD at high speed mode - 48 MHz seen on scope.  Patches for HS mode not
yet in kernel.  That needs to happen soon.

MarkF thinks that SD write-protect will be functional.

Wireless:

USB experts at Marvell still reviewing the situation.

Dan thinks his reflash tool works, Dave and MarkF agree.

Dave has been using the firmware reflash tool to test the
wireless vs. NAND failure mode.

AI: Chris.  Reconfirm that NAND causes problems with wireless in
Cambridge - check reflash tool and the real driver.

Color map - X driver / Xv was messing with the color map.

Jim has seen strange behavior with some touchpad samples.
Touchpad crashes.  Sometimes it just stops working.  Sometimes
it will go vertically but not horzontally.  Really old
touchpads seem to work more reliably.

USB:

Routing has been improved for the build, but can't test that.
The problems continue to persist, especially with wireless.

AMD is working on power-up problems.  May be issues with
devices switching to PS/2 mode if they don't win with USB.

AI: Marcello to upgrade to recent OFW with new EC bits to
see if it helps with the disappearing wireless problem.

The new boards will have several ECOs, including a USB power switch.
The USB switch does not require any immediate BIOS changes.

Audio:

The version of csound that uses /dev/pcm causes DMA errors.
The version that uses ALSA works well.

Use the keyword "relnote" for trac entries that have info
that needs to be release notes.

EC seems to be locked.  Something is leaving it that
way and needs to unlock it.

AI: Mitch:  Make sure EC is unlocked.  trac 314


Blocker bugs:

42 sensor input - probably won't get done.  Rescheduled to Btest-2

65 tmpfs in RAM - fixed

231 btest board fails to boot.  Still open.  Reassign to Jordan

257 sd card detect with power off - to Btest-2

258 partially fixed, still problem with NAND

261 still weeding out
I missed some...

263 on chris's list for tomorrow

266 reassign to J5

268 in build but not tested
269 same

270 AMD still looking at it
272 AMD still looking at it
273 AMD thinks the device is driving 5V.

274 still open - OFW workaround is effective, but data tranfer errors
    are still worrisome.  But this is not the highest priority.
abiword needs more testing

280 video output is delayed after DCON -> GPU transition.
    Jordan has ideas.  The delay needs to 64 ms.
    Jordan thinks we should leave the panel engine up.
    Currently we are not going into DCON mode without suspend to RAM.

csound b1 demo launcher - assign to Dan.  May defer to B2

285 open for moment
286 Marcello working on
288 Marcello working on

291 is in beta 1 but csound blocks it
294 ?
298 reassign to dwmw
301 reassign to Andres
314 reassign to wmb
318 still open
319 open
322 defer
328 reassign to marco
329 Colormap.  Mitch, Jordan, Jim to discuss