headers
svn | 1 Aug 2007 05:11

r507 - ofw/fs/fatfs

Author: wmb
Date: 2007-08-01 05:11:46 +0200 (Wed, 01 Aug 2007)
New Revision: 507

Modified:
   ofw/fs/fatfs/bpb.fth
   ofw/fs/fatfs/fat.fth
Log:
FAT filesystem - OLPC trac 2609.  Fixed FAT32 cluster allocation bug.
The "next cluster" field in the fsinfo structure was being treated as
a definitive free cluster number, instead of as a hint for where to
begin searching.  In addition, it was off by one - that field
typically contains the number of the most recently written cluster,
whereas the code was treating it as the next available cluster number.

Modified: ofw/fs/fatfs/bpb.fth
===================================================================
--- ofw/fs/fatfs/bpb.fth	2007-07-30 21:08:57 UTC (rev 506)
+++ ofw/fs/fatfs/bpb.fth	2007-08-01 03:11:46 UTC (rev 507)
 <at>  <at>  -91,21 +91,11  <at>  <at> 
 d# 16 constant fat16
 d# 32 constant fat32

-: fs-lelop  ( adr op -- )
-   fsinfo  <at>  0<>  if
-      over lel <at>  swap execute swap lel!
-      true fsinfos-dirty c!
-   else
-      2drop
-   then
(Continue reading)

Permalink | Reply | 0 comments |
headers
svn | 1 Aug 2007 05:13

r508 - in dev: . olpc/touchpad

Author: wmb
Date: 2007-08-01 05:13:15 +0200 (Wed, 01 Aug 2007)
New Revision: 508

Modified:
   dev/i8042.fth
   dev/olpc/touchpad/touchpad.fth
Log:
PS2 drivers - Added debugging features to help in debugging
ESD-induced resets of the OLPC touchpad.

Modified: dev/i8042.fth
===================================================================
--- dev/i8042.fth	2007-08-01 03:11:46 UTC (rev 507)
+++ dev/i8042.fth	2007-08-01 03:13:15 UTC (rev 508)
 <at>  <at>  -19,6 +19,12  <at>  <at> 
 1 " #address-cells"  integer-property
 0 " #size-cells"     integer-property

+0 value debug?
+also forth definitions
+: debug-ps2    ( -- )  true  to debug?  ;
+: undebug-ps2  ( -- )  false to debug?  ;
+previous definitions
+
 hex

 [ifndef] $=
 <at>  <at>  -97,7 +103,8  <at>  <at> 
 headerless
(Continue reading)

Permalink | Reply | 0 comments |
headers
svn | 1 Aug 2007 05:15

r509 - in cpu/x86/pc: . biosload

Author: wmb
Date: 2007-08-01 05:15:29 +0200 (Wed, 01 Aug 2007)
New Revision: 509

Modified:
   cpu/x86/pc/biosload/addrs.fth
   cpu/x86/pc/biosload/config.fth
   cpu/x86/pc/biosload/fw.bth
   cpu/x86/pc/biosload/ofw.bth
   cpu/x86/pc/elfhdr.fth
   cpu/x86/pc/resetend.fth
Log:
Patch from Jens Freimann to support running OFW under LinuxBIOS on QEMU.

Modified: cpu/x86/pc/biosload/addrs.fth
===================================================================
--- cpu/x86/pc/biosload/addrs.fth	2007-08-01 03:13:15 UTC (rev 508)
+++ cpu/x86/pc/biosload/addrs.fth	2007-08-01 03:15:29 UTC (rev 509)
 <at>  <at>  -9,9 +9,15  <at>  <at> 
 \ copies it to get it out of the way of things like OS load areas.

 [ifdef] linuxbios-loaded
-h# fff8.0000 constant dropin-base  \ Location of payload in FLASH
-h#   08.0000 constant dropin-size
-dropin-base h# 80 + h# 20 +  constant ResetBase	\ Location of "reset" dropin in ROM
+  [ifdef] qemu-loaded
+    \ when running in qemu OFW is not in ROM but loaded to RAM by elfboot
+    h# 198.0080 constant dropin-base  \ Location of payload in RAM
+    dropin-base h# 20 +  constant ResetBase	\ Location of "reset" dropin in RAM
+  [else]
(Continue reading)

Permalink | Reply | 0 comments |
headers
svn | 1 Aug 2007 10:25

r510 - cpu/x86/pc/olpc dev/olpc/cafenand

Author: wmb
Date: 2007-08-01 10:25:31 +0200 (Wed, 01 Aug 2007)
New Revision: 510

Modified:
   cpu/x86/pc/olpc/copynand.fth
   dev/olpc/cafenand/badblock.fth
   dev/olpc/cafenand/cafenand.fth
Log:
OLPC copy-nand - OLPC trac 2614.  Made copy-nand abort not only on a
CRC mismatch, but also on a NAND write error.

Modified: cpu/x86/pc/olpc/copynand.fth
===================================================================
--- cpu/x86/pc/olpc/copynand.fth	2007-08-01 03:15:29 UTC (rev 509)
+++ cpu/x86/pc/olpc/copynand.fth	2007-08-01 08:25:31 UTC (rev 510)
 <at>  <at>  -148,7 +148,8  <at>  <at> 
       (cr i .
       read-image-block
       i ?check-crc
-      load-base " copy-block" nandih $call-method
+      load-base " copy-block" nandih $call-method  ( error? )
+      " Error writing to NAND FLASH" ?nand-abort
    loop

    cr ." Filling with cleanmarkers ..."

Modified: dev/olpc/cafenand/badblock.fth
===================================================================
--- dev/olpc/cafenand/badblock.fth	2007-08-01 03:15:29 UTC (rev 509)
(Continue reading)

Permalink | Reply | 0 comments |
headers
svn | 1 Aug 2007 10:32

r511 - cpu/x86/pc/olpc

Author: wmb
Date: 2007-08-01 10:32:03 +0200 (Wed, 01 Aug 2007)
New Revision: 511

Modified:
   cpu/x86/pc/olpc/versions.fth
Log:
Version Q2C21.

Modified: cpu/x86/pc/olpc/versions.fth
===================================================================
--- cpu/x86/pc/olpc/versions.fth	2007-08-01 08:25:31 UTC (rev 510)
+++ cpu/x86/pc/olpc/versions.fth	2007-08-01 08:32:03 UTC (rev 511)
 <at>  <at>  -2,7 +2,7  <at>  <at> 

 \ The overall firmware revision
 macro: FW_MAJOR C
-macro: FW_MINOR 20
+macro: FW_MINOR 21

 \ The EC microcode
 macro: EC_VERSION c20
Permalink | Reply | 0 comments |
headers
svn | 1 Aug 2007 21:17

r512 - cpu/x86/pc/olpc

Author: rsmith
Date: 2007-08-01 21:17:30 +0200 (Wed, 01 Aug 2007)
New Revision: 512

Modified:
   cpu/x86/pc/olpc/versions.fth
Log:
OLPC: Updater version number to q2c20e

Modified: cpu/x86/pc/olpc/versions.fth
===================================================================
--- cpu/x86/pc/olpc/versions.fth	2007-08-01 08:32:03 UTC (rev 511)
+++ cpu/x86/pc/olpc/versions.fth	2007-08-01 19:17:30 UTC (rev 512)
 <at>  <at>  -2,7 +2,7  <at>  <at> 

 \ The overall firmware revision
 macro: FW_MAJOR C
-macro: FW_MINOR 21
+macro: FW_MINOR 20e

 \ The EC microcode
 macro: EC_VERSION c20
Permalink | Reply | 0 comments |
headers
Blue Swirl | 1 Aug 2007 21:47
Picon

Re: [Qemu-devel] IBM makes AIX 6 beta available for download for everyone

The correct place for these questions would be the OpenBIOS list
(blatantly cross-posted).

On 7/30/07, Andreas Färber <andreas.faerber <at> web.de> wrote:
>
> Am 19.07.2007 um 09:23 schrieb Natalia Portillo:
>
> > El jue, 19-07-2007 a las 00:46 +0200, Andreas Färber escribió:
> >> Am 18.07.2007 um 18:26 schrieb Blue Swirl:
> >>
> >>>>> It looks like open hack'ware is pretty much stalled, so its hard
> >>>>> telling
> >>>>> what progress will be made futurewise.
> >>>>
> >>>> Might it be possible to replace it with OpenBIOS?
> >>>
> >>> Sure, and as both are GPL useful, code can be shared. But some PPC
> >>> developers are needed. It may still be easier to fix the OHW bugs
> >>> than
> >>> start over anew.
> >>
> >> Not sure I understand... I thought Open Firmware was based on
> >> platform independent Forth code?
> >> I'm running on ppc64 and could try to compile something under Linux
> >> in case that's the problem.
> > Yes, OF is basically a Forth-interpreter.
> > Of course this Forth-interpreter is in C and compile to machine
> > code :p
>
> Yes, that is obvious. Not so obvious was what ppc *developers* were
(Continue reading)

Permalink | Reply | 0 comments |
headers
svn | 2 Aug 2007 07:26

r513 - dev/olpc/kb3700

Author: wmb
Date: 2007-08-02 07:26:07 +0200 (Thu, 02 Aug 2007)
New Revision: 513

Modified:
   dev/olpc/kb3700/battery.fth
Log:
OLPC watch-battery - OLPC track 2620 - Use EC commands for reading
battery status, instead of indexed I/O to EC RAM locations.

Modified: dev/olpc/kb3700/battery.fth
===================================================================
--- dev/olpc/kb3700/battery.fth	2007-08-01 19:17:30 UTC (rev 512)
+++ dev/olpc/kb3700/battery.fth	2007-08-02 05:26:07 UTC (rev 513)
 <at>  <at>  -11,16 +11,23  <at>  <at> 
 : bat-b <at>  h# f900 + ec <at>   ;
 : bat-w <at>  dup 1+ bat-b <at>  swap bat-b <at>  bwjoin ;
 : eram-w <at>   h# f400 + dup 1+ ec <at>  swap ec <at>  bwjoin ;
-: uvolt <at>  0 bat-w <at>  d# 9760 d# 32 */ ;
-: cur <at>  2 bat-w <at>  wextend  d# 15625 d# 120 */ ;
 \ Base unit for temperature is 1/256 degrees C
 : >degrees-c 7 rshift 1+ 2/  ;  \ Round to nearest degree 
-: pcb-temp 8 bat-w <at>  >degrees-c  ;
-: bat-temp 6 bat-w <at>  >degrees-c  ;
 : .%  ( n -- )  .d ." %" ;
-: soc     h# 10 bat-b <at>   ;
+\ : uvolt <at>  0 bat-w <at>  d# 9760 d# 32 */ ;
+\ : cur <at>  2 bat-w <at>  wextend  d# 15625 d# 120 */ ;
+\ : pcb-temp 8 bat-w <at>  >degrees-c  ;
+\ : bat-temp 6 bat-w <at>  >degrees-c  ;
(Continue reading)

Permalink | Reply | 0 comments |
headers
svn | 2 Aug 2007 07:35

r514 - cpu/x86/pc/olpc

Author: wmb
Date: 2007-08-02 07:35:04 +0200 (Thu, 02 Aug 2007)
New Revision: 514

Modified:
   cpu/x86/pc/olpc/versions.fth
Log:
q2c21

Modified: cpu/x86/pc/olpc/versions.fth
===================================================================
--- cpu/x86/pc/olpc/versions.fth	2007-08-02 05:26:07 UTC (rev 513)
+++ cpu/x86/pc/olpc/versions.fth	2007-08-02 05:35:04 UTC (rev 514)
 <at>  <at>  -2,7 +2,7  <at>  <at> 

 \ The overall firmware revision
 macro: FW_MAJOR C
-macro: FW_MINOR 20e
+macro: FW_MINOR 21

 \ The EC microcode
 macro: EC_VERSION c20
Permalink | Reply | 0 comments |
headers
svn | 2 Aug 2007 12:23

r515 - dev/olpc/spiflash

Author: wmb
Date: 2007-08-02 12:23:58 +0200 (Thu, 02 Aug 2007)
New Revision: 515

Modified:
   dev/olpc/spiflash/spiui.fth
Log:
SPI FLASH selftest - OLPC trac 2636 - Workaround a hardware design
deficiency by throttling access to the SPI FLASH in the /flash
selftest method.  If the CPU accesses the SPI FLASH too frequently,
the EC apparently can't do its instruction fetches, and the system
resets.  The workaround is to access the SPI FLASH in 64K chunks,
waiting 100 mS between chunks.

Modified: dev/olpc/spiflash/spiui.fth
===================================================================
--- dev/olpc/spiflash/spiui.fth	2007-08-02 05:35:04 UTC (rev 514)
+++ dev/olpc/spiflash/spiui.fth	2007-08-02 10:23:58 UTC (rev 515)
 <at>  <at>  -243,8 +243,17  <at>  <at> 
 0 value rom-va
 : selftest  ( -- error? )
    rom-va 0=  if  rom-pa /flash root-map-in to rom-va  then
-   rom-va flash-buf /flash move

+   \ This is a slower version of "rom-va flash-buf /flash lmove"
+   \ It works around the problem that continuous CPU access to the
+   \ SPI FLASH starves the EC of instruction fetch cycles, often
+   \ causing it to turn off the system.
+   rom-va flash-buf
+   /flash  0  do
(Continue reading)

Permalink | Reply | 0 comments |

Gmane