openfirmware October 2008

openfirmware@openfirmware.info

1 participants
24 discussions

r979 - dev
by svn＠openfirmware.info 05 Oct '08

05 Oct '08

Author: wmb Date: 2008-10-05 11:37:32 +0200 (Sun, 05 Oct 2008) New Revision: 979 Modified: dev/ps2mouse.fth Log: OLPC trac 8451 - further improvements in mouse/trackpad robustness. Sometimes the driver would fail to open if either the first identify command returned "RESEND" or the EC had mouse events queued up. And the driver failed to turn off streaming mode on close, thus causing events to queue up if you touched the touchpad between "test /mouse" runs. Modified: dev/ps2mouse.fth =================================================================== --- dev/ps2mouse.fth 2008-10-04 00:11:03 UTC (rev 978) +++ dev/ps2mouse.fth 2008-10-05 09:37:32 UTC (rev 979) @@ -25,9 +25,35 @@ \ apex? 0= if " get-data" $call-parent ?ack then \ ; +: timed-read ( #ms -- true | data false ) + 0 do + get-data? if unloop false exit then + 1 ms + loop + true +; + +: put-get-data ( cmd -- data ) " put-get-data" $call-parent ; : send-cmd ( cmd -- ) " put-data" $call-parent ; -: cmd ( cmd -- ) " put-get-data" $call-parent ?ack ; +: cmd? ( cmd -- error? ) + dup put-get-data ( cmd response ) + begin ( cmd response ) + case ( cmd response ) + h# fa of drop false exit endof \ ACK + h# fe of dup put-get-data endof \ RESEND - try again + ( cmd response ) + d# 10 timed-read if ( cmd response ) + 2drop true exit + then ( cmd response new-response ) + swap ( cmd new-response response ) + endcase ( cmd new-response ) + again +; +: cmd ( cmd -- ) cmd? drop ; + +\ : cmd ( cmd -- ) " put-get-data" $call-parent ?ack ; + \ This serves the same purpose as "cmd", but it handles a bug in the \ Apex concentrator in which said device neglects to acknowledge certain \ commands. @@ -100,13 +126,6 @@ \ PS2 mouse packets have no framing information, so we must depend on \ timing information for framing : clear-queue ( -- ) begin get-data? while drop repeat ; -: timed-read ( #ms -- true | data false ) - 0 do - get-data? if unloop false exit then - 1 ms - loop - true -; : no-event ( -- 0 0 0 ) clear-queue 0 0 0 ; headers @@ -148,9 +167,7 @@ : identify ( -- true | char false ) \ h# f2 ['] read1 catch dup if nip then - h# f2 send-cmd - - apex-timeout timed-read if true exit else drop then + h# f2 cmd? if true exit then apex-timeout timed-read ; @@ -198,7 +215,7 @@ remote-mode true ; -: close ( -- ) ; +: close ( -- ) stream-off ; headerless \ Filter out large negative motions; they're probably spurious

1 0

r978 - cpu/x86/pc/olpc
by svn＠openfirmware.info 04 Oct '08

04 Oct '08

Author: wmb Date: 2008-10-04 02:11:03 +0200 (Sat, 04 Oct 2008) New Revision: 978 Modified: cpu/x86/pc/olpc/versions.fth Log: Q2E19 Modified: cpu/x86/pc/olpc/versions.fth =================================================================== --- cpu/x86/pc/olpc/versions.fth 2008-10-04 00:10:58 UTC (rev 977) +++ cpu/x86/pc/olpc/versions.fth 2008-10-04 00:11:03 UTC (rev 978) @@ -2,21 +2,21 @@ \ The overall firmware revision macro: FW_MAJOR E -macro: FW_MINOR 18a +macro: FW_MINOR 19 \ The EC microcode -macro: EC_VERSION e15 +macro: EC_VERSION e18c \ Alternate command for getting EC microcode, for testing new versions. \ Temporarily uncomment the line and modify the path as necessary -\ macro: GET_EC cp PQ2D08a.bin ec.img +\ macro: GET_EC cp pq2e18c.img ec.img macro: KEYS mpkeys \ macro: KEYS testkeys \ The wireless LAN module firmware macro: WLAN_RPM ${WLAN_VERSION}.olpc1 -macro: WLAN_VERSION 5.110.22.p18 +macro: WLAN_VERSION 5.110.22.p20 \ The bios_verify image macro: CRYPTO_VERSION 0.2

1 0

r977 - /
by svn＠openfirmware.info 04 Oct '08

04 Oct '08

Author: wmb Date: 2008-10-04 02:10:58 +0200 (Sat, 04 Oct 2008) New Revision: 977 Modified: .gitignore Log: More gitignore additions Modified: .gitignore =================================================================== --- .gitignore 2008-10-04 00:10:55 UTC (rev 976) +++ .gitignore 2008-10-04 00:10:58 UTC (rev 977) @@ -16,3 +16,4 @@ tags *~ headers +*.md5

1 0

r976 - in cpu/x86/pc: . olpc
by svn＠openfirmware.info 04 Oct '08

04 Oct '08

Author: wmb Date: 2008-10-04 02:10:55 +0200 (Sat, 04 Oct 2008) New Revision: 976 Modified: cpu/x86/pc/linux.fth cpu/x86/pc/olpc/gui.fth cpu/x86/pc/olpc/security.fth Log: This change encompasses three not-quite-related issues: a) The new ramdisk-moving code, needed for recent Linux kernels, was breaking both old kernels and new kernels when booted in secure mode. b) The secure mode code was leaving the display in freeze-when-Linux-starts mode when a developer key was found - sort of a "half-pretty" situation. c) Secure startup didn't give Linux a command line, so you couldn't watch the kernel startup messages on the serial port. Modified: cpu/x86/pc/linux.fth =================================================================== --- cpu/x86/pc/linux.fth 2008-10-04 00:10:49 UTC (rev 975) +++ cpu/x86/pc/linux.fth 2008-10-04 00:10:55 UTC (rev 976) @@ -42,8 +42,13 @@ : code16-size ( -- #bytes ) load-base h# 1f1 + c@ 1+ d# 512 * ; 0 value cmdline-offset +0 value linux-memtop + \ Find the end of the largest piece of memory : memory-limit ( -- limit ) + \ If we have already loaded a RAMdisk in high memory, its base is the memory limit + ramdisk-adr ?dup if exit then + " /memory" find-package 0= abort" No /memory node" ( phandle ) " available" rot get-package-property abort" No memory node available property" ( $ ) \ Find the memory piece that starts at 1 Meg @@ -92,11 +97,11 @@ : set-parameters ( cmdline$ -- ) screen-info linux-params /screen-info move \ Ostensibly screen info - memory-limit ( #bytes ) + linux-memtop ( #bytes ) d# 1023 invert and d# 1024 / ( #kbytes ) d# 1024 - h# 002 +lp w! \ Kbytes of extended (not the 1st meg) memory - memory-limit ( #bytes ) + linux-memtop ( #bytes ) d# 1023 invert and d# 1024 / ( #kbytes ) d# 1024 - h# 1e0 +lp l! \ Alternate amount of extended memory @@ -156,6 +161,27 @@ ' ramdisk-buf " ramdisk" chosen-string defer load-ramdisk +: place-ramdisk ( adr len -- ) + to /ramdisk ( adr ) + + \ Move ramdisk to top of memory for new kernels. In principle, + \ this should work for older kernels too, but for some reason it doesn't. + h# 206 +lp w@ h# 207 < if ( adr ) + dup load-base <> if ( adr ) + load-base tuck /ramdisk move ( adr' ) + then ( adr ) + memory-limit ( adr memtop ) + else + \ The initrd_addr_max field appeared in boot protocol 2.03 + h# 22c +lp l@ ( adr ramdisk-limit ) + ?dup if 1+ else h# 8000.0000 then ( adr ramdisk-limit ) + + memory-limit umin /ramdisk - ( adr new-ramdisk-adr ) + tuck /ramdisk move ( new-ramdisk-adr ) + dup + then ( ramdisk-adr memtop ) + to linux-memtop to ramdisk-adr +; : $load-ramdisk ( name$ -- ) 0 to /ramdisk ( name$ ) @@ -168,18 +194,7 @@ r> to load-path ( throw-code ) throw - loaded to /ramdisk ( adr ) - - \ Move ramdisk to top of memory - - \ The initrd_addr_max field appeared in boot protocol 2.03 - h# 22c +lp l@ ( adr ramdisk-limit ) - ?dup if 1+ else h# 8000.0000 then ( adr ramdisk-limit ) - - memory-limit umin /ramdisk - ( adr new-ramdisk-adr ) - dup to ramdisk-adr ( adr new-ramdisk-adr ) - - /ramdisk move ( ) + loaded place-ramdisk ; : cv-load-ramdisk ( -- ) " ramdisk" eval dup 0= if 2drop exit then ( name$ ) Modified: cpu/x86/pc/olpc/gui.fth =================================================================== --- cpu/x86/pc/olpc/gui.fth 2008-10-04 00:10:49 UTC (rev 975) +++ cpu/x86/pc/olpc/gui.fth 2008-10-04 00:10:55 UTC (rev 976) @@ -398,7 +398,7 @@ ; : linux-hook-freeze [ ' linux-hook behavior compile, ] - 0 " set-source" $call-screen + dcon-freeze ; : freeze ( -- ) ['] linux-hook-freeze to linux-hook ; : unfreeze ( -- ) ['] linux-hook-unfreeze to linux-hook ; Modified: cpu/x86/pc/olpc/security.fth =================================================================== --- cpu/x86/pc/olpc/security.fth 2008-10-04 00:10:49 UTC (rev 975) +++ cpu/x86/pc/olpc/security.fth 2008-10-04 00:10:55 UTC (rev 976) @@ -562,9 +562,7 @@ " RD found - " ?lease-debug img$ sig$ sha-valid? if show-unlock - load-base to ramdisk-adr - img$ dup to /ramdisk ( adr len ) - load-base swap move ( ) + img$ place-ramdisk exit else show-unlock @@ -762,6 +760,10 @@ 0<> ; +: set-cmdline ( -- ) + " console=ttyS0,115200 console=tty0 fbcon=font:SUN12x22" args-buf place-cstr drop +; + : load-from-list ( list$ -- devkey? ) " dev /jffs2-file-system ' ?unfreeze to scan-callout dend" eval @@ -780,7 +782,7 @@ has-developer-key? if ( list$ ) 2drop ( ) true to security-off? - visible + visible unfreeze show-unlock true exit then ( list$ ) @@ -790,6 +792,7 @@ 2drop ( ) ['] secure-load-ramdisk to load-ramdisk " init-program" $find if + set-cmdline execute show-going go then show-x @@ -826,7 +829,7 @@ ?force-secure - secure? 0= if unfreeze visible exit then + secure? 0= if visible unfreeze exit then button-check game-key? if unfreeze visible banner @@ -834,7 +837,7 @@ freeze dcon-freeze then - persistent-devkey? if true to security-off? visible exit then + persistent-devkey? if true to security-off? visible unfreeze exit then get-my-sn if " No serial number" .security-failure then

1 0

r975 - cpu/x86/pc/olpc
by svn＠openfirmware.info 04 Oct '08

04 Oct '08

Author: wmb Date: 2008-10-04 02:10:49 +0200 (Sat, 04 Oct 2008) New Revision: 975 Modified: cpu/x86/pc/olpc/fw.bth Log: OLPC - installed second-chance memory allocator to fix a problem with multicast NAND updater. Modified: cpu/x86/pc/olpc/fw.bth =================================================================== --- cpu/x86/pc/olpc/fw.bth 2008-10-04 00:10:44 UTC (rev 974) +++ cpu/x86/pc/olpc/fw.bth 2008-10-04 00:10:49 UTC (rev 975) @@ -82,6 +82,7 @@ fload ${BP}/ofw/core/showlist.fth \ Linked list display tool fload ${BP}/ofw/core/allocph1.fth \ S Physical memory allocator fload ${BP}/ofw/core/availpm.fth \ Available memory list +fload ${BP}/ofw/core/allocmor.fth \ Secondary allocator fload ${BP}/cpu/x86/pc/rootnode.fth \ Platform-specific root node changes dev / @@ -625,12 +626,19 @@ ; : mcastnand " boot rom:mcastnand 239.255.1.2 12345 /nandflash" eval + false to already-go? ; +: ucastnand + " boot rom:mcastnand 10.20.0.16,,10.20.0.44 12345 /nandflash" eval + false to already-go? +; : nmcastnand \ Boot from network, for testing " boot http:\\10.20.0.14\mcastnand 239.255.1.2 12345 /nandflash" eval + false to already-go? ; : dmcastnand \ Boot from USB disk, for testing " boot disk:\mcnand 239.255.1.2 12345 /nandflash" eval + false to already-go? ;

1 0

r974 - forth/lib
by svn＠openfirmware.info 04 Oct '08

04 Oct '08

Author: wmb Date: 2008-10-04 02:10:44 +0200 (Sat, 04 Oct 2008) New Revision: 974 Modified: forth/lib/rstrace.fth Log: Generic rstrace command - decode do loop indices and catch frames. Modified: forth/lib/rstrace.fth =================================================================== --- forth/lib/rstrace.fth 2008-10-04 00:10:41 UTC (rev 973) +++ forth/lib/rstrace.fth 2008-10-04 00:10:44 UTC (rev 974) @@ -12,18 +12,57 @@ decimal only forth also hidden also definitions headerless +: @+ ( adr -- adr' n ) dup na1+ swap @ ; : .last-executed ( ip -- ) ip>token token@ ( acf ) dup reasonable-ip? if .name else drop ." ??" then ; -: .traceline ( ipaddr -- ) - push-hex - dup reasonable-ip? - if dup .last-executed ip>token .caller else 9 u.r then cr +: in-catch? ( ip -- flag ) find-cfa ['] catch = ; +: .catch ( rs-adr -- rs-adr' ) + ." Catch frame - SP: " @+ . ." my-self: " @+ . ." handler: " @+ . +; +1 bits/cell 1- lshift constant minus0 +: .do-or-n ( rs-adr n -- rs-adr' ) + over @ reasonable-ip? 0= if ( rs-adr n ) + \ The second number is not an IP so it could be a do loop frame + over na1+ @ reasonable-ip? if ( rs-adr n ) + \ The third entry is a reasonable IP so it could be a do loop frame + \ Make sure it points just past a loop end + over na1+ @ ( rs-adr n n2 ) + ip>token -1 na+ token@ ( rs-adr n xt ) + dup ['] (loop) = swap ['] (+loop) = or if ( rs-adr n ) + \ The two numbers span the +- boundary, so probably a do loop + ." Do loop frame inside " + over na1+ @ ip>token .current-word ( rs-adr n ) + over @ ( rs-adr n n1 ) + ." i: " tuck + . ( rs-adr n1 ) + ." limit: " minus0 + . ( rs-adr ) + 2 na+ exit + then ( rs-adr n ) + then ( rs-adr n ) + then ( rs-adr n ) + 9 u.r +; +: .traceline ( rs-adr -- rs-adr' ) + push-hex ( rs-adr ) + @+ ( rs-adr' ip ) + dup reasonable-ip? if ( rs-adr ip ) + dup in-catch? if ( rs-adr ip ) + drop .catch ( rs-adr' ) + else ( rs-adr ip ) + dup .last-executed ip>token .caller ( rs-adr ) + then ( rs-adr ) + else ( rs-adr ip ) + .do-or-n ( rs-adr ) + then cr ( rs-adr ) pop-base ; -: (rstrace ( bottom-adr top-adr -- ) - do i @ .traceline exit? ?leave /n +loop +: (rstrace ( end-adr start-adr -- ) + begin 2dup u> while ( end-adr adr ) + .traceline ( end-adr adr' ) + exit? if 2drop exit then ( end-adr adr ) + repeat ( end-adr adr ) + 2drop ; headers forth definitions @@ -35,6 +74,7 @@ then ; only forth also definitions + \ LICENSE_BEGIN \ Copyright (c) 2006 FirmWorks \

1 0

r973 - dev/usb2/device/wlan
by svn＠openfirmware.info 04 Oct '08

04 Oct '08

Author: wmb Date: 2008-10-04 02:10:41 +0200 (Sat, 04 Oct 2008) New Revision: 973 Modified: dev/usb2/device/wlan/wlan.fth Log: OLPC wireless driver - disable multicast when closing, just in case. Modified: dev/usb2/device/wlan/wlan.fth =================================================================== --- dev/usb2/device/wlan/wlan.fth 2008-10-04 00:10:36 UTC (rev 972) +++ dev/usb2/device/wlan/wlan.fth 2008-10-04 00:10:41 UTC (rev 973) @@ -99,6 +99,7 @@ : close ( -- ) opencount @ 1- 0 max opencount ! opencount @ 0= if + disable-multicast link-up? if target-mac$ deauthenticate then ['] 2drop to ?process-eapol end-bulk-in

1 0

r972 - cpu/x86/pc/olpc
by svn＠openfirmware.info 04 Oct '08

04 Oct '08

Author: wmb Date: 2008-10-04 02:10:36 +0200 (Sat, 04 Oct 2008) New Revision: 972 Modified: cpu/x86/pc/olpc/usb8388.bth Log: OLPC usb8388 build script - added both "-b" and plain versions of the md5 check command to make it easier to handle differences in the way that people publish new wireless firmware files. Modified: cpu/x86/pc/olpc/usb8388.bth =================================================================== --- cpu/x86/pc/olpc/usb8388.bth 2008-10-04 00:10:31 UTC (rev 971) +++ cpu/x86/pc/olpc/usb8388.bth 2008-10-04 00:10:36 UTC (rev 972) @@ -11,7 +11,8 @@ " wget -q http://dev.laptop.org/pub/firmware/libertas/usb8388-${WLAN_VERSION}.bin.md5" expand$ $sh -" md5sum usb8388-${WLAN_VERSION}.bin | cmp - usb8388-${WLAN_VERSION}.bin.md5" expand$ $sh +\ " md5sum usb8388-${WLAN_VERSION}.bin | cmp - usb8388-${WLAN_VERSION}.bin.md5" expand$ $sh +" md5sum -b usb8388-${WLAN_VERSION}.bin | cmp - usb8388-${WLAN_VERSION}.bin.md5" expand$ $sh " mv usb8388-${WLAN_VERSION}.bin usb8388.bin" expand$ $sh " rm usb8388-${WLAN_VERSION}.bin.md5" expand$ $sh

1 0

r971 - dev
by svn＠openfirmware.info 04 Oct '08

04 Oct '08

Author: wmb Date: 2008-10-04 02:10:31 +0200 (Sat, 04 Oct 2008) New Revision: 971 Modified: dev/pckbd.fth Log: OLPC - Added support for new OLPC keyboard controller. Modified: dev/pckbd.fth =================================================================== --- dev/pckbd.fth 2008-10-01 19:05:29 UTC (rev 970) +++ dev/pckbd.fth 2008-10-04 00:10:31 UTC (rev 971) @@ -102,9 +102,17 @@ : clear-out-buf ( -- ) " clear-out-buf" $call-parent ; +: timed-read ( #ms -- true | data false ) + 0 do + get-data? if unloop false exit then + 1 ms + loop + true +; + : ?ack ( response -- ) begin h# fa <> while \ Consume stuff while waiting for ack - d# 10 ms get-data? 0= if exit then + d# 10 timed-read if exit then repeat ; @@ -211,18 +219,36 @@ oem-keymap if oem-keymap /keymap free-mem then 0 to oem-keymap ; + [ifdef] olpc : ?olpc-keyboard ( -- ) " enable-intf" $call-parent - begin get-data? while + begin 5 timed-read 0= while drop true to keyboard-present? - 5 ms repeat keyboard-present? if exit then kbd-reset 0= to keyboard-present? ; +\ For the ENE keyboard controller we have to tell the EC to use +\ a different internal mapping table. OLPC switched from an +\ ALPS to an ENE controller in late 2007. +: olpc-set-ec-keymap ( -- ) + h# f2 cmd + + \ Keyboards return the ID sequence ab XX + 2 timed-read if exit then ( id1 ) + h# ab <> if exit then + + \ The ENE keyboard controller return ab 41 + 2 timed-read if exit then ( id2 ) + h# 41 <> if exit then + + \ This looks like an ENE controller, so flip the mapping table + h# f7 cmd +; + fload ${BP}/cpu/x86/pc/olpc/keymap.fth [then] @@ -328,13 +354,13 @@ " us" set-keyboard \ choose-type ; -: consume ( -- ) begin 5 ms get-data? while drop repeat ; +: consume ( -- ) begin 5 timed-read 0= while drop repeat ; : get-initial-state ( -- ) d# 200 ms \ Give the keyboard time to respond \ 5 ms is less than the standard auto-repeat rate - begin 5 ms get-data? while handle-initial-scan repeat + begin 5 timed-read 0= while handle-initial-scan repeat default-disable-kbd \ Clear last typematic key, if any consume @@ -359,7 +385,7 @@ clear-state [ifdef] ?olpc-keyboard ?olpc-keyboard - keyboard-present? if exit then + keyboard-present? if olpc-set-ec-keymap exit then [else] get-initial-state

1 0

r970 - / clients clients/emacs clients/memtest86 cpu/x86/Linux cpu/x86/pc/olpc/build
by svn＠openfirmware.info 01 Oct '08

01 Oct '08

Author: wmb Date: 2008-10-01 21:05:29 +0200 (Wed, 01 Oct 2008) New Revision: 970 Added: .gitignore clients/.gitignore clients/emacs/.gitignore clients/memtest86/.gitignore cpu/x86/Linux/.gitignore cpu/x86/pc/olpc/build/.gitignore Log: Added some .gitignore files for people who import the SVN repository into git. Added: .gitignore =================================================================== --- .gitignore (rev 0) +++ .gitignore 2008-10-01 19:05:29 UTC (rev 970) @@ -0,0 +1,18 @@ +*.o +*.log +*.di +*.fc +*.aml +*.tag +*.img +*.dic +*.dep +*.version +*.a +*.bin +builton.fth +*.rom +*.elf +tags +*~ +headers Added: clients/.gitignore =================================================================== --- clients/.gitignore (rev 0) +++ clients/.gitignore 2008-10-01 19:05:29 UTC (rev 970) @@ -0,0 +1 @@ +*.o Added: clients/emacs/.gitignore =================================================================== --- clients/emacs/.gitignore (rev 0) +++ clients/emacs/.gitignore 2008-10-01 19:05:29 UTC (rev 970) @@ -0,0 +1 @@ +emacs Added: clients/memtest86/.gitignore =================================================================== --- clients/memtest86/.gitignore (rev 0) +++ clients/memtest86/.gitignore 2008-10-01 19:05:29 UTC (rev 970) @@ -0,0 +1,3 @@ +memtest +memtest_shared +head.s Added: cpu/x86/Linux/.gitignore =================================================================== --- cpu/x86/Linux/.gitignore (rev 0) +++ cpu/x86/Linux/.gitignore 2008-10-01 19:05:29 UTC (rev 970) @@ -0,0 +1,2 @@ +*.o +*forth Added: cpu/x86/pc/olpc/build/.gitignore =================================================================== --- cpu/x86/pc/olpc/build/.gitignore (rev 0) +++ cpu/x86/pc/olpc/build/.gitignore 2008-10-01 19:05:29 UTC (rev 970) @@ -0,0 +1,5 @@ +*.public +multicast-nand/ +build +dsdt.stamp +sourceurl

1 0

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

openfirmware October 2008