December 2009 - openfirmware - mail.coreboot.org

r1601 - cpu/x86/pc/olpc/via
by svn＠openfirmware.info Dec. 14, 2009

Dec. 14, 2009

Author: pgf Date: 2009-12-15 00:28:13 +0100 (Tue, 15 Dec 2009) New Revision: 1601 Modified: cpu/x86/pc/olpc/via/assy.fth Log: Changed the IP address in the ka-dir$ value. Modified: cpu/x86/pc/olpc/via/assy.fth =================================================================== --- cpu/x86/pc/olpc/via/assy.fth 2009-12-14 22:40:31 UTC (rev 1600) +++ cpu/x86/pc/olpc/via/assy.fth 2009-12-14 23:28:13 UTC (rev 1601) @@ -5,7 +5,7 @@ : swid$ ( -- adr len ) " OFW ASSY test $Revision$" ; \ Location of the files containing KA tag data -: ka-dir$ ( -- adr len ) " http:\\10.0.0.1\ka\" ; +: ka-dir$ ( -- adr len ) " http:\\10.1.0.1\ka\" ; : put-ka-tag ( value$ key$ -- ) 2over 8 min ka-dir$ " %s%s" sprintf ( value$ key$ filename$ )

1 0

r1600 - cpu/x86/pc/olpc/via
by svn＠openfirmware.info Dec. 14, 2009

Dec. 14, 2009

Author: wmb Date: 2009-12-14 23:40:31 +0100 (Mon, 14 Dec 2009) New Revision: 1600 Modified: cpu/x86/pc/olpc/via/runin.fth Log: Via manufacturing tests - the NTP/RTC comparison code was horribly broken. Also changed the IP address in the ka-dir$ value. Modified: cpu/x86/pc/olpc/via/runin.fth =================================================================== --- cpu/x86/pc/olpc/via/runin.fth 2009-12-14 16:05:50 UTC (rev 1599) +++ cpu/x86/pc/olpc/via/runin.fth 2009-12-14 22:40:31 UTC (rev 1600) @@ -3,8 +3,13 @@ visible \ Location of the files containing KA tag data -: ka-dir$ ( -- adr len ) " http:\\10.0.0.1\ka\" ; +: ka-dir$ ( -- adr len ) " http:\\10.1.0.1\ka\" ; +: mfg-ntp-server ( -- name$ ) + " NT" find-tag if ?-null else " 10.1.0.1" then +; +' mfg-ntp-server to ntp-servers + \ The Linux-based runin selftests put this file at int:\runin\olpc.fth \ after they have finished. On the next reboot, OFW thus boots this \ script instead of int:\boot\olpc.fth . This script either displays @@ -177,31 +182,33 @@ kbc-on ; -: mfg-ntp-server ( -- name$ ) - " NT" find-tag if ?-null else " 10.60.0.2" then +d# 4 constant rtc-threshold +: .clocks ( -- ) + ntp>time&date time&date + ." RTC: " .clock ." NTP: " .clock ; -' mfg-ntp-server to ntp-servers - -d# 4 constant rtc-threshold : verify-rtc-date ( -- ) \ XXX check RTC power lost bit - ." Getting time from NTP server " + ." Getting time from NTP server .. " begin ntp-timestamp while ." Retry " repeat ( d.timestamp ) - cr - ntp>time&date unix-seconds> ( ntp-seconds ) - today unix-seconds> ( ntp-seconds rtc-seconds ) - - ( difference ) - dup rtc-threshold > if ( difference ) - page show-fail - ." Clock lost " .d ." seconds since SMT" cr - begin halt again - else - abs rtc-threshold > if - page show-fail - ." Clock gained " .d ." seconds since SMT" cr - begin halt again - then + + ntp>time&date >unix-seconds ( ntp-seconds ) + time&date >unix-seconds ( ntp-seconds rtc-seconds ) + - ( lost-seconds ) + dup rtc-threshold > if ( lost-seconds ) + page show-fail ( lost-seconds ) + ." Clock lost " .d ." seconds since SMT" cr ( ) + .clocks + stall + then ( lost-seconds ) + + abs dup rtc-threshold > if ( gained-seconds ) + page show-fail ( gained-seconds ) + ." Clock gained " .d ." seconds since SMT" cr ( ) + .clocks + stall then + ." NTP and RTC clocks agree." cr ; : put-key:value ( value$ key$ -- ) " %s:%s" sprintf put-key-line ;

1 0

r1599 - cpu/x86/pc/olpc/via
by svn＠openfirmware.info Dec. 14, 2009

Dec. 14, 2009

Author: rsmith Date: 2009-12-14 17:05:50 +0100 (Mon, 14 Dec 2009) New Revision: 1599 Modified: cpu/x86/pc/olpc/via/versions.fth Log: OLPC 1.5: Firmware version q3a24 Modified: cpu/x86/pc/olpc/via/versions.fth =================================================================== --- cpu/x86/pc/olpc/via/versions.fth 2009-12-14 10:22:45 UTC (rev 1598) +++ cpu/x86/pc/olpc/via/versions.fth 2009-12-14 16:05:50 UTC (rev 1599) @@ -2,7 +2,7 @@ \ The overall firmware revision macro: FW_MAJOR A -macro: FW_MINOR 23 +macro: FW_MINOR 24 \ The EC microcode macro: EC_VERSION 1_9_18

1 0

r1598 - dev/hdaudio
by svn＠openfirmware.info Dec. 14, 2009

Dec. 14, 2009

Author: wmb Date: 2009-12-14 11:22:45 +0100 (Mon, 14 Dec 2009) New Revision: 1598 Modified: dev/hdaudio/core.fth Log: HDAudio driver - got rid of remaining memory leak. Modified: dev/hdaudio/core.fth =================================================================== --- dev/hdaudio/core.fth 2009-12-14 10:03:24 UTC (rev 1597) +++ dev/hdaudio/core.fth 2009-12-14 10:22:45 UTC (rev 1598) @@ -452,7 +452,7 @@ : (write-done) ( -- ) stop-stream - free-bdl + teardown-bdl release-sound-buffer uninstall-playback-alarm ; @@ -545,7 +545,7 @@ wait-stream-done stop-stream release-sound-buffer - free-pad-buffer + teardown-bdl /recbuf ;

1 0

r1597 - cpu/x86/pc/olpc/via
by svn＠openfirmware.info Dec. 14, 2009

Dec. 14, 2009

Author: pgf Date: 2009-12-14 11:03:24 +0100 (Mon, 14 Dec 2009) New Revision: 1597 Modified: cpu/x86/pc/olpc/via/assy.fth Log: unfreeze screen, differentiate two identical error messages, and add missing parameter to execute-downloads Modified: cpu/x86/pc/olpc/via/assy.fth =================================================================== --- cpu/x86/pc/olpc/via/assy.fth 2009-12-14 10:01:05 UTC (rev 1596) +++ cpu/x86/pc/olpc/via/assy.fth 2009-12-14 10:03:24 UTC (rev 1597) @@ -1,5 +1,7 @@ \ Manufacturing test boot script for ASSY state +visible + : swid$ ( -- adr len ) " OFW ASSY test $Revision$" ; \ Location of the files containing KA tag data @@ -31,7 +33,7 @@ : check-smt-status ( -- ) " SS" find-tag 0= abort" Board failed SMT !!!" ( adr len ) -null ( adr len' ) - " EN" $= 0= abort" Board failed SMT !!!" ( ) + " EN" $= 0= abort" Board failed SMT !!!!" ( ) ; : fwver$ ( -- adr len ) h# ffff.ffc6 6 ; @@ -207,7 +209,7 @@ cifs-connect assy-tag-exchange cifs-disconnect ." Done" cr - execute-downloads + response$ execute-downloads inject-tags ." Powering off ..." d# 2000 ms

1 0

r1596 - dev/hdaudio
by svn＠openfirmware.info Dec. 14, 2009

Dec. 14, 2009

Author: wmb Date: 2009-12-14 11:01:05 +0100 (Mon, 14 Dec 2009) New Revision: 1596 Modified: dev/hdaudio/core.fth dev/hdaudio/test.fth Log: HDAudio driver and selftest - significant changes to reduce memory leaks, prevent reentrancy in the selftest, and clean up after sound completion. Modified: dev/hdaudio/core.fth =================================================================== --- dev/hdaudio/core.fth 2009-12-14 09:48:18 UTC (rev 1595) +++ dev/hdaudio/core.fth 2009-12-14 10:01:05 UTC (rev 1596) @@ -78,11 +78,11 @@ \ \\ Stream Descriptors \ Default: 48kHz 16bit stereo -0 instance value sample-base -0 instance value sample-mul -0 instance value sample-div -1 instance value sample-format -2 instance value #channels +0 value sample-base +0 value sample-mul +0 value sample-div +1 value sample-format +2 value #channels : stream-format ( -- u ) sample-base d# 14 lshift ( acc ) @@ -107,7 +107,7 @@ \ Stream descriptor register interface. \ There are multiple stream descriptors, each with their own register set. -0 instance value sd# +0 value sd# : sd+ ( offset -- adr ) sd# h# 20 * + au + ; : sdctl h# 80 sd+ ; @@ -239,8 +239,8 @@ : init-codec ( -- ) detect-codec open-codec ; : close-controller ( -- ) reset unmap-regs ; -d# 48.000 instance value sample-rate -1 instance value scale-factor +d# 48.000 value sample-rate +1 value scale-factor : low-rate? ( Hz ) dup d# 48.000 < swap d# 44.100 <> and ; @@ -404,7 +404,8 @@ upsample-channel \ left src 2+ to src dst 2+ to dst upsample-channel \ right - dst 2 - /dst ( dst dst-len ) + dst 2- to dst ( ) + dst /dst ( dst dst-len ) ; \ \\ Amplifier control @@ -453,11 +454,12 @@ stop-stream free-bdl release-sound-buffer + uninstall-playback-alarm ; : write-done ( -- ) wait-stream-done (write-done) ; : write ( adr len -- actual ) - 4 to sd# audio-out install-playback-alarm true to playing? + 4 to sd# audio-out true to playing? install-playback-alarm ; : ?end-sound ( -- ) @@ -475,11 +477,25 @@ \ Alarm handle to stop the stream when the content has been played. : playback-completed-alarm ( -- ) stop-lock if exit then - playing? if sd# ?end-sound to sd# then + playing? if + sd# ?end-sound to sd# + else + \ If playback has already stopped as a result of + \ someone else having waited for completion, we + \ just uninstall ourself. + uninstall-playback-alarm + then ; ' playback-completed-alarm is playback-alarm +: still-playing? ( -- flag ) + playing? 0= if false exit then + stop-lock if true exit then + sd# ?end-sound to sd# + playing? +; + : wait-sound ( -- ) true to stop-lock begin playing? while d# 10 ms ?end-sound repeat Modified: dev/hdaudio/test.fth =================================================================== --- dev/hdaudio/test.fth 2009-12-14 09:48:18 UTC (rev 1595) +++ dev/hdaudio/test.fth 2009-12-14 10:01:05 UTC (rev 1596) @@ -8,11 +8,9 @@ record-base record-len la1+ " dma-free" $call-parent ; -: jingle ( -- ) " play-wav rom:splash" evaluate wait-sound ; - : ?key-abort ( -- ) key? if - key h# 1b = abort" Aborting" + key h# 1b = if ." Aborting" abort then then ; : speaker-test ( -- ) @@ -21,14 +19,8 @@ ." Disconnect headphones to continue.. " begin ?key-abort pin-sense? 0= until cr then - ." Playing jingle on the left speaker.. " - true to right-mute? - jingle cr - false to right-mute? - true to left-mute? - ." Playing jingle on the right speaker.. " - jingle cr - false to left-mute? + ." Playing left to right sweep " + make-sweep 0 set-volume play cr ; : headphones-test ( -- ) @@ -37,9 +29,8 @@ ." Connect headphones to continue.. " begin ?key-abort pin-sense? until cr then - ." Press a key to play sound.. " key drop cr h# 1f to node power-off \ turn off speaker - jingle + make-sweep -9 set-volume play h# 1f to node power-on \ turn speaker back on ;

1 0

r1595 - ofw/core
by svn＠openfirmware.info Dec. 14, 2009

Dec. 14, 2009

Author: wmb Date: 2009-12-14 10:48:18 +0100 (Mon, 14 Dec 2009) New Revision: 1595 Modified: ofw/core/ofwcore.fth Log: OFW core - changed the implementation of the alarm list scanner so that an alarm handler can safely remove itself from the list from inside the handler. Modified: ofw/core/ofwcore.fth =================================================================== --- ofw/core/ofwcore.fth 2009-12-13 20:22:36 UTC (rev 1594) +++ ofw/core/ofwcore.fth 2009-12-14 09:48:18 UTC (rev 1595) @@ -3898,18 +3898,29 @@ \ then check to see if the time is expired (time-remain = 0). \ If time is not expired, decrement the time-remain. -: time-expired? ( node -- flag ) +: run-alarm ( node -- ) dup >time-remain @ 1- dup 0<= if ( node time-remain ) drop dup >time-out @ over ( node time-out node ) dup >acf @ swap >ihandle @ ( node time-out acf ihandle ) ['] call-package catch if ( node time-out acf ihandle ) 2drop ( node time-out ) then ( node time-out ) - then swap >time-remain ! false ( false ) + then swap >time-remain ! ( ) ; headers -: check-alarm ( -- ) alarm-list ['] time-expired? find-node 2drop ; +\ We do this manually instead of using find-node because we need +\ to do >next-node before calling time-expired? in case the alarm +\ routine uninstalls itself, which could cause a crash if the +\ pointer to the next node were overwritten while being freed. +: check-alarm ( -- ) + alarm-list >next-node ( node ) + begin ?dup while ( node ) + dup >next-node swap ( next node ) + run-alarm ( next ) + repeat ( ) +; + : alarm ( acf n -- ) my-self -rot ( ihandle acf n ) ?dup if set-alarm-node else turn-off-alarm then

1 0

r1594 - cpu/x86/pc/olpc/via
by svn＠openfirmware.info Dec. 13, 2009

Dec. 13, 2009

Author: wmb Date: 2009-12-13 21:22:36 +0100 (Sun, 13 Dec 2009) New Revision: 1594 Added: cpu/x86/pc/olpc/via/mfgtests.txt Log: Via manufacturing tests - checked in documentation file. Added: cpu/x86/pc/olpc/via/mfgtests.txt =================================================================== --- cpu/x86/pc/olpc/via/mfgtests.txt (rev 0) +++ cpu/x86/pc/olpc/via/mfgtests.txt 2009-12-13 20:22:36 UTC (rev 1594) @@ -0,0 +1,159 @@ +=== Manufacturing Test Sequencing === + +Version as of 2009-12-11 + +The manufacturing process is sequenced by setting the BD tag, which +controls OFW's boot-device list. BD has different values at different +stages, thus controlling the way that the system boots. For a given +manufacturing phase, BD points to a list of boot scripts from local +devices or network servers that are available at that station. The +boot script performs phase-dependent activities. + +The TS tag identifies the test station. Its primary use, other than +as a human-readable indicator of the system's progress through the +manufacturing flow, is to fine-tune the selftest diagnostics for the +needs of the particular test station. + +Test results and other information about the system are sent to a +station-dependent server whose location and credentials are given by +the MS tag. That server uses a tag-exchange protocol to receive +information about the system and to return information that is used +to setup the machine for the next phase. That setup typically includes +new values for BD, TS, and MS that are suitable for the next station. + +The starting values for those tags are: + + TS=SMT + MS=cifs:\\bekins:bekind2@10.60.0.2\nb2_fvs + BD=u:\boot\olpc.fth cifs:\\bekins:bekind2@10.60.0.2\nb2_fvs\olpc.fth + +They are initialized by running a script to inject those tags into the +initially-blank manufacturing data area in the OLPC-released firmware +image. That creates a new master image that is used to preprogram (via +a gang programmer) the SPI FLASH chips before they are soldered to the +boards. + +=== SMT state === + +Entry condition: + + TS=SMT + MS=cifs:\\bekins:bekind2@10.60.0.2\nb2_fvs + BD=u:\boot\olpc.fth cifs:\\bekins:bekind2@10.60.0.2\nb2_fvs\olpc.fth + +The booted script performs selftest in test-station=1 mode. Upon success, +it performs tag exchange with the "MS" server and rewrites the tags to +the value the server specified (examples below). + +=== ASSY state === + +Entry condition: + + TS=ASSY + MS=cifs:\\Administrator:qmssdl@10.0.0.2\olpc_mono1 + BD=u:\boot\olpc.fth net + +The booted (from the net, using TFTP etc) script does the tag +exchange, runs the download (with command line given by the tag +exchange response), and rewrites the tags. The next state is RUNIN. +The "DL" state never shows up in the tags. + +The response includes one or more "Command:<Forth_commands>" lines. +Those lines are executed to perform the download steps. +An example value for such a line is "fs-update http:\\10.0.0.1\os41.zd". +The Doing it this way instead of hardcoding it in the script permits +country-specific +images without having to edit the script. The file could also be +served from a Windows server if desired; it need not be http. + +The response tags from this stage include the ones shown in entry +condition for the RUNIN state. + +The script also needs to know the URI for a directory containing +files for KA values. The filename is <ka-dir><filename> where +<filename> is what the server sends after "KA:" . For now, the +test script hardcodes the value of <ka-dir>. + + +=== RUNIN state === + + TS=RUNIN + BD=u:\boot\olpc.fth int:\runin\olpc.fth int:\boot\olpc.fth + MS=<URI for final server> + +Initially, the boot file is int:\boot\olpc.fth from the OS +distribution. The Linux-based runin tests will create +int:\runin\fail.log when finished. If the runin tests failed, +they will create int:\runin\fail.log prior to creating +int:\runin\olpc.fth. + +The script in int:\runin\olpc.fth does the following: + +# Check for the existence of int:\runin\fail.log. If it exists, + display it on the screen with a red background and stop. Later + rebooting will continue to show the red-screen error message + until the user manually deletes int:\runin\fail.log and + int:\runin\olpc.fth by manually executing the command "re-runin" + (that command is defined in \runin\olpc.fth . It first deletes + \runin\olpc.fth and then \runin\fail.log ) . + +# Otherwise (pass case) + a) Starts selftest in final mode + b) (The test are ordered so the non-interactive ones are first) + c) When the sd card insertion point happens, a big check is displayed + d) The operator moves the unit from carousel to the final test line and inserts SD card + ??? Do we need to insert a USB stick too? + ??? Do we need to insert scanner and wlan here too to give the USB test something to work on? + e) final test steps continue until pass or fail + +# If selftest passed + Tell operator to move to the download station and insert USB LAN + wait-lan + wait-scanner + tag exchange with server + barcode scan of serial number to verify that the board and case are still paired + ??? Is the filename the MB (B#) or the case serial number (SN) ? + ??? If the filename is the SN, why is the SN in the request file? + The server can't verify if it only has the SN and not the board# + delete MS and BD tags, insert tags from server + ??? what is the format of the WP:N and AK:N info from the server? Is it WP:0 / WP:1, or presence/absence ? + TS should be SHIP at this point + ??? Does the server send TS:SHIP or does the script inject it verbatim? + +# If selftest failed, red screen and stop. State of machine (BD tag, + \runin\olpc.fth) left the same as at the start of this phase so + the phase can be rerun. + +=== SHIP state === + +Entry condition + + TS=SHIP + MS (deleted) + BD (deleted) + +=== Design Features === + +The above design has the following features: + +* Once the machine leaves the factory, it does not contain any + information about factory server names or their credentials. + (The tags containing that information are deleted for CUSTOMER + state, and the Linux-based runin scripts contain no hardcoded + factory server identification.) + +* The boot scripts that perform the test sequencing do not have to + change to accomodate server changes. (The server information + is stored in mfg data tags.) + +* OFW contains very little "hardcoded" information about the + process steps - it just boots from the specified device at + each stage. This minimizes the changes to the OFW startup + sequence - the only change is to set "boot-device" from BD + if it exists. If BD doesn't exist, the normal default list + is used, which supports booting from USB, so you can recover + a system that is out of sequence. + +* At any step, the normal server-based process can be overridden by + plugging in a USB stick with a custom \boot\olpc.fth . (Each BD + list has "u:\boot\olpc.fth" at the beginning.)

1 0

r1593 - cpu/x86/pc/olpc/via
by svn＠openfirmware.info Dec. 13, 2009

Dec. 13, 2009

Author: rsmith Date: 2009-12-13 15:27:27 +0100 (Sun, 13 Dec 2009) New Revision: 1593 Modified: cpu/x86/pc/olpc/via/versions.fth Log: OLPC 1.5: Firmware release version q3a23 Modified: cpu/x86/pc/olpc/via/versions.fth =================================================================== --- cpu/x86/pc/olpc/via/versions.fth 2009-12-13 14:12:28 UTC (rev 1592) +++ cpu/x86/pc/olpc/via/versions.fth 2009-12-13 14:27:27 UTC (rev 1593) @@ -2,7 +2,7 @@ \ The overall firmware revision macro: FW_MAJOR A -macro: FW_MINOR 22 +macro: FW_MINOR 23 \ The EC microcode macro: EC_VERSION 1_9_18

1 0

r1592 - cpu/x86/pc/olpc/via
by svn＠openfirmware.info Dec. 13, 2009

Dec. 13, 2009

Author: rsmith Date: 2009-12-13 15:12:28 +0100 (Sun, 13 Dec 2009) New Revision: 1592 Modified: cpu/x86/pc/olpc/via/versions.fth Log: OLPC 1.5: Update EC code version Modified: cpu/x86/pc/olpc/via/versions.fth =================================================================== --- cpu/x86/pc/olpc/via/versions.fth 2009-12-13 10:30:04 UTC (rev 1591) +++ cpu/x86/pc/olpc/via/versions.fth 2009-12-13 14:12:28 UTC (rev 1592) @@ -5,7 +5,7 @@ macro: FW_MINOR 22 \ The EC microcode -macro: EC_VERSION 1_9_16 +macro: EC_VERSION 1_9_18 \ Alternate command for getting EC microcode, for testing new versions. \ Temporarily uncomment the line and modify the path as necessary

1 0