openfirmware November 2008

openfirmware@openfirmware.info

1 participants
19 discussions

r998 - cpu/x86/pc/olpc
by svn＠openfirmware.info 06 Nov '08

06 Nov '08

Author: wmb Date: 2008-11-06 09:49:37 +0100 (Thu, 06 Nov 2008) New Revision: 998 Modified: cpu/x86/pc/olpc/versions.fth Log: Q2E22 redux redux (last checkin had the version number wrong). Modified: cpu/x86/pc/olpc/versions.fth =================================================================== --- cpu/x86/pc/olpc/versions.fth 2008-11-06 08:48:09 UTC (rev 997) +++ cpu/x86/pc/olpc/versions.fth 2008-11-06 08:49:37 UTC (rev 998) @@ -2,7 +2,7 @@ \ The overall firmware revision macro: FW_MAJOR E -macro: FW_MINOR 22a +macro: FW_MINOR 22 \ The EC microcode macro: EC_VERSION e21

1 0

r997 - cpu/x86/pc/olpc
by svn＠openfirmware.info 06 Nov '08

06 Nov '08

Author: wmb Date: 2008-11-06 09:48:09 +0100 (Thu, 06 Nov 2008) New Revision: 997 Modified: cpu/x86/pc/olpc/versions.fth Log: Q2E22 redux. Modified: cpu/x86/pc/olpc/versions.fth =================================================================== --- cpu/x86/pc/olpc/versions.fth 2008-11-06 08:47:40 UTC (rev 996) +++ cpu/x86/pc/olpc/versions.fth 2008-11-06 08:48:09 UTC (rev 997) @@ -2,7 +2,7 @@ \ The overall firmware revision macro: FW_MAJOR E -macro: FW_MINOR 22 +macro: FW_MINOR 22a \ The EC microcode macro: EC_VERSION e21

1 0

r996 - dev/olpc/touchpad
by svn＠openfirmware.info 06 Nov '08

06 Nov '08

Author: wmb Date: 2008-11-06 09:47:40 +0100 (Thu, 06 Nov 2008) New Revision: 996 Modified: dev/olpc/touchpad/touchpad.fth Log: OLPC trac #8940 - fixed glitches in selftest for OLPC ALPS touchpad. The problem was that the display routine was looking at the mode variable to determine how to display the dot, but that variable had sometimes already been switched to the new mode, before the packet from the old mode had been displayed. I also changed the display to overlay cyan and yellow in the center of the screen, with registration to corresponding positions on the touchpad. Modified: dev/olpc/touchpad/touchpad.fth =================================================================== --- dev/olpc/touchpad/touchpad.fth 2008-11-06 05:46:47 UTC (rev 995) +++ dev/olpc/touchpad/touchpad.fth 2008-11-06 08:47:40 UTC (rev 996) @@ -90,15 +90,16 @@ : alps-prefix ( -- ) 3 0 do h# f2 read1 drop loop ; variable mode \ 0 - unknown 1 - GS 2 - PT 3 - mouse +0 instance value packet-type \ Ref: 5.2.10 (3) of Hybrid-GP2B-T-1.pdf : gs-only ( -- ) - mode @ dup 1 = swap 3 = or if exit then + mode @ 3 = if exit then alps-prefix mouse1:1 \ f2 f2 f2 e6 1 mode ! ; : pt-only ( -- ) - mode @ dup 2 = swap 3 = or if exit then + mode @ swap 3 = if exit then alps-prefix mouse2:1 \ f2 f2 f2 e7 2 mode ! ; @@ -172,7 +173,7 @@ \ variable py \ Pen mode y value variable gx \ Glide pad x value variable gy \ Glide pad y value -variable gz \ Glide pad z value, also used fo pen mode x +variable gz \ Glide pad z value, also used for pen mode x variable taps \ Bitmask of tap flags variable switches \ Bitmask of switch flags @@ -233,6 +234,7 @@ \ Ref: 5.2.9 (2) (1) of Hybrid-GP2B-T-1.pdf : decode-pt ( -- ) show-raw? if ." P " then + 2 to packet-type decode-common taps @ 0= if d# 10 exit-pt ! then ; @@ -240,6 +242,7 @@ \ Ref: 5.2.9 (2) (2) of Hybrid-GP2B-T-1.pdf : decode-gs ( -- ) show-raw? if ." G " then + 1 to packet-type \ In case we get out of sync decode-common taps @ 1 and if pt-only then \ Switch to pen mode ; @@ -395,7 +398,7 @@ ; : track ( x y z buttons -- ) - mode @ 2 = if yellow else cyan then pixcolor ! ( x y z but ) + packet-type 2 = if yellow else cyan then pixcolor ! ( x y z but ) dup 3 and 3 = if background load-base ptr ! then dup 1 and if green else black then d# 100 button @@ -406,7 +409,23 @@ show-pressure ( x y ) - relative? if abs>rel then + relative? if ( x y ) + abs>rel ( x' y' ) + else ( x y ) + \ Display cyan dots (GS) in the center of the screen + \ The offset attempts to register the GS with the PT. + \ The offset should be 333, but the PT is expanded, + \ so we have to scale the 333 to match. + packet-type 1 = if ( x y ) + swap d# 400 + ( y x' ) \ GS + swap ( x' y ) + then ( x y ) + \ Expand the PT (0..999) to the screen width (1200) + packet-type 2 = if ( x y ) + swap d# 1200 d# 1000 */ + swap d# 1200 d# 1000 */ + then ( x y ) + then ( x y ) dot ;

1 0

r995 - cpu/x86/pc/olpc
by svn＠openfirmware.info 06 Nov '08

06 Nov '08

Author: wmb Date: 2008-11-06 06:46:47 +0100 (Thu, 06 Nov 2008) New Revision: 995 Modified: cpu/x86/pc/olpc/versions.fth Log: Q2E22 Modified: cpu/x86/pc/olpc/versions.fth =================================================================== --- cpu/x86/pc/olpc/versions.fth 2008-11-05 02:30:57 UTC (rev 994) +++ cpu/x86/pc/olpc/versions.fth 2008-11-06 05:46:47 UTC (rev 995) @@ -2,7 +2,7 @@ \ The overall firmware revision macro: FW_MAJOR E -macro: FW_MINOR 21 +macro: FW_MINOR 22 \ The EC microcode macro: EC_VERSION e21

1 0

r994 - ofw/fs/jffs2
by svn＠openfirmware.info 05 Nov '08

05 Nov '08

Author: wmb Date: 2008-11-05 03:30:57 +0100 (Wed, 05 Nov 2008) New Revision: 994 Modified: ofw/fs/jffs2/jffs2.fth Log: OLPC trac 8932 - non-secure OFW reflash from NAND fails. The actual problem was a bug in the JFFS2 file reader, which used the wrong buffer for reconstructing the file in some cases. Modified: ofw/fs/jffs2/jffs2.fth =================================================================== --- ofw/fs/jffs2/jffs2.fth 2008-11-04 00:57:29 UTC (rev 993) +++ ofw/fs/jffs2/jffs2.fth 2008-11-05 02:30:57 UTC (rev 994) @@ -51,7 +51,7 @@ : next-tdirent ( -- adr ) 'next-tdirent @ ; 0 instance value tdirents -0 instance value file-buf \ Buffer for constructing file +-1 instance value file-buf \ Buffer for constructing file 0 instance value file-size \ Actual size of file 0 ( instance ) value /empty-scan \ When to give up looking for nodes

1 0

r993 - dev
by svn＠openfirmware.info 04 Nov '08

04 Nov '08

Author: wmb Date: 2008-11-04 01:57:29 +0100 (Tue, 04 Nov 2008) New Revision: 993 Modified: dev/ps2mouse.fth Log: Added "stream-poll?" and "adaptive-poll?" mouse methods so that client programs can track the mouse efficiently. Modified: dev/ps2mouse.fth =================================================================== --- dev/ps2mouse.fth 2008-11-03 17:32:11 UTC (rev 992) +++ dev/ps2mouse.fth 2008-11-04 00:57:29 UTC (rev 993) @@ -66,7 +66,7 @@ : mouse2:1 h# e7 cmd ; : stream-on h# f4 cmd ; : stream-off h# f5 cmd ; -: stream-mode h# ea cmd ; +: stream-mode h# ea cmd stream-on ; \ The Apex concentrator refuses to send motion and button events \ unless stream data is enabled! @@ -104,7 +104,7 @@ ; [then] -\ Translate RML to RML +\ Translate MRL to RML create buttons 0 c, 1 c, 4 c, 5 c, 2 c, 3 c, 6 c, 7 c, \ create Mitsumi \ Comment out this line if you don't support Mitsumi mouse @@ -128,7 +128,123 @@ : clear-queue ( -- ) begin get-data? while drop repeat ; : no-event ( -- 0 0 0 ) clear-queue 0 0 0 ; +0 instance value mouse-byte# +instance variable mouse-bytes +0 instance value mouse-timestamp +2 instance value poll-delay + +: save-byte ( byte -- ) + mouse-bytes @ 8 lshift or mouse-bytes ! ( ) + mouse-byte# 1+ to mouse-byte# ( ) + get-msecs to mouse-timestamp + 2 to poll-delay +; +: out-of-packet ( -- ) + 0 to mouse-byte# + 0 to mouse-timestamp +; + +: stream-event ( byte -- false | x y buttons true ) + mouse-byte# case ( byte ) + 0 of + \ Discard if framing error + dup 8 and if save-byte else drop then ( ) + false exit + endof + + 1 of save-byte false exit then + endcase ( byte3 ) + + \ Fall through if already have 2 mouse bytes (this is the third) + out-of-packet + + mouse-bytes @ wbsplit ( byte3-ylow byte2-xlow byte1-stat ) + + dup 7 and buttons + c@ >r ( ylow xlow stat ) + swap over h# 10 ?negative -rot ( x ylow stat ) + h# 20 ?negative r> ( x y buttons ) + + true exit +; + +\ Time-based resynchronization. If we are in the middle of a packet, +\ but we poll and see no bytes and the interval since the last byte +\ was longer than the possible time between bytes within a packet, +\ we discard the queued-up bytes. +: ?reframe ( -- ) + mouse-byte# if + \ We don't trust time intervals <= ms/tick because the timestamp + \ could have happened just before a tick and the next sample + \ right after a tick, so the actual elapsed time could be almost + \ zero, but the reported time difference would be ms/tick . + get-msecs mouse-timestamp - ms/tick - 0 max d# 5 > if + out-of-packet + then + then +; headers + +: stream-poll? ( -- false | x y buttons true ) + begin get-data? while ( byte ) + \ If we have a packet, retry soon because there is likely to be another + \ right on its heels. The packet generation rate is 100/sec, but some + \ could be queued up, so we don't wait a full 10 ms + stream-event ?dup if exit then + repeat ( ) + + ?reframe + false +; + +\ delay-ms is a suggested delay until the next poll. It tells you when +\ a call to stream-poll? is likely to have something worthwhile to do. +\ In the middle of an incoming mouse packet, delay-ms will be a couple of +\ milliseconds, a bit longer than the interval between successive PS/2 bytes. +\ Just after receiving a packet, it will be a bit more than 10 mS, since +\ another packet is likely to follow (the standard mouse packet rate is +\ 100 packets/second). When idle, the delay is 56 milliseconds, just +\ longer than a worst-case PC tick interval, but short enough that a +\ human won't notice a delay. +\ This strikes a good balance between responsive, accurate mouse tracking +\ and minimal polling overhead. + +: adaptive-poll? ( -- false delay-ms | x y buttons true delay-ms ) + begin get-data? while ( byte ) + \ If we have a packet, retry soon because there is likely to be another + \ right on its heels. The packet generation rate is 100/sec, but some + \ could be queued up, so we don't wait a full 10 ms + stream-event ?dup if poll-delay exit then + repeat ( ) + + ?reframe + false poll-delay ( false delay-ms ) + + dup 2 = if 9 else d# 32 then to poll-delay ( false delay-ms ) +; +0 [if] +: adaptive-track ( -- ) + stream-mode + begin + adaptive-poll? >r if ( x y stat r: next-poll ) + (cr . 4 .r 4 .r ( r: next-poll ) + then ( r: next-poll ) + r> ms ( ) + key? until ( ) + stream-off +; +: stream-track ( -- ) + stream-mode + begin + stream-poll? if ( x y stat ) + (cr . 4 .r 4 .r ( ) + then ( ) + 5 ms ( ) + key? until ( ) + stream-off +; + +[then] + : poll ( -- x y buttons ) clear-queue

1 0

r992 - ofw/core
by svn＠openfirmware.info 03 Nov '08

03 Nov '08

Author: wmb Date: 2008-11-03 18:32:11 +0100 (Mon, 03 Nov 2008) New Revision: 992 Modified: ofw/core/bootparm.fth Log: Increased the size of args-buf to 256 to handle long Linux cmdlines. Modified: ofw/core/bootparm.fth =================================================================== --- ofw/core/bootparm.fth 2008-11-01 01:11:05 UTC (rev 991) +++ ofw/core/bootparm.fth 2008-11-03 17:32:11 UTC (rev 992) @@ -66,7 +66,7 @@ headers ' path-buf " bootpath" chosen-string headerless -d# 128 buffer: args-buf +d# 256 buffer: args-buf headers ' args-buf " bootargs" chosen-string @@ -82,6 +82,13 @@ ' path-buf is load-path headerless +: limit-255 ( adr len -- adr len ) + dup d# 255 > if + ." Warning: limiting string length to 255 characters" cr + drop d# 255 + then +; + : (boot-read) ( adr len -- ) opened-ih if ( adr len ) 2drop opened-ih 0 to opened-ih ( ihandle ) @@ -92,7 +99,7 @@ ( print-probe-list ) true abort" "r"nCan't open boot device"r"n" then ( fileid ) - dup ihandle>devname load-path place-cstr drop ( fileid ) + dup ihandle>devname limit-255 load-path place-cstr drop ( fileid ) >r ( ) load-started 0 !load-size load-base ( load-adr ) @@ -196,7 +203,7 @@ ." Arguments: " 2over type ( file-str device-str ) )collect ?show-message ( file-str device-str ) - 2swap args-buf place-cstr drop ( device-str ) + 2swap limit-255 args-buf place-cstr drop ( device-str ) boot-read ( ) ;

1 0

r991 - cpu/x86/pc/olpc
by svn＠openfirmware.info 01 Nov '08

01 Nov '08

Author: wmb Date: 2008-11-01 02:11:05 +0100 (Sat, 01 Nov 2008) New Revision: 991 Modified: cpu/x86/pc/olpc/copynand.fth cpu/x86/pc/olpc/nandstat.fth Log: Q2E21 fixup - scan-nand was misclassifying erased-without-cleanmarkers blocks. Modified: cpu/x86/pc/olpc/copynand.fth =================================================================== --- cpu/x86/pc/olpc/copynand.fth 2008-11-01 00:05:41 UTC (rev 990) +++ cpu/x86/pc/olpc/copynand.fth 2008-11-01 01:11:05 UTC (rev 991) @@ -382,7 +382,7 @@ \ See if the whole thing is really completely erased load-base working-page nand-pages/block ( adr block# #blocks ) - " read-blocks" $call-nand nand-pages/block <> if 1 exit then + " read-pages" $call-nand nand-pages/block <> if 1 exit then \ Not completely erased load-base /nand-block written? if 5 exit then Modified: cpu/x86/pc/olpc/nandstat.fth =================================================================== --- cpu/x86/pc/olpc/nandstat.fth 2008-11-01 00:05:41 UTC (rev 990) +++ cpu/x86/pc/olpc/nandstat.fth 2008-11-01 01:11:05 UTC (rev 991) @@ -159,7 +159,7 @@ \ See if the whole thing is really completely erased load-base working-page nand-pages/block ( adr block# #blocks ) - " read-blocks" $call-nand nand-pages/block <> if 1 exit then + " read-pages" $call-nand nand-pages/block <> if 1 exit then \ Not completely erased load-base load-base h# 100000 + /nand-block comp if 5 exit then

1 0

r990 - cpu/x86/pc/olpc
by svn＠openfirmware.info 01 Nov '08

01 Nov '08

Author: wmb Date: 2008-11-01 01:05:41 +0100 (Sat, 01 Nov 2008) New Revision: 990 Modified: cpu/x86/pc/olpc/versions.fth Log: Q2E21 Modified: cpu/x86/pc/olpc/versions.fth =================================================================== --- cpu/x86/pc/olpc/versions.fth 2008-10-31 17:18:42 UTC (rev 989) +++ cpu/x86/pc/olpc/versions.fth 2008-11-01 00:05:41 UTC (rev 990) @@ -2,10 +2,10 @@ \ The overall firmware revision macro: FW_MAJOR E -macro: FW_MINOR 20 +macro: FW_MINOR 21 \ The EC microcode -macro: EC_VERSION e18c +macro: EC_VERSION e21 \ Alternate command for getting EC microcode, for testing new versions. \ Temporarily uncomment the line and modify the path as necessary

1 0

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

openfirmware November 2008