OpenBIOS June 2007

openbios@openbios.org

4 participants
31 discussions

r441 - dev/olpc/cafenand
by svn＠openbios.org 07 Jun '07

07 Jun '07

Author: wmb Date: 2007-06-07 02:32:10 +0200 (Thu, 07 Jun 2007) New Revision: 441 Modified: dev/olpc/cafenand/selftest.fth Log: OLPC CaFe NAND selftest - fixed a stack error. Modified: dev/olpc/cafenand/selftest.fth =================================================================== --- dev/olpc/cafenand/selftest.fth 2007-06-07 00:13:32 UTC (rev 440) +++ dev/olpc/cafenand/selftest.fth 2007-06-07 00:32:10 UTC (rev 441) @@ -67,10 +67,16 @@ ." Fill " dup .d ." random blocks with h# 55, h# aa and h# ff" cr 0 ?do - i 1 > if random-page else i 0<> if usable-page-limit 1- else 0 then then - pages/eblock 1- invert and + i case + 0 of 0 endof + 1 of usable-page-limit 1- endof + ( default ) random-page swap + endcase ( block# ) + pages/eblock 1- invert and ( block#' ) - dup block-bad? not if + dup block-bad? if ( block#' ) + drop ( ) + else ( block#' ) (cr dup . sbuf over read-eblock dup record-err 0= if dup h# 55 test-eblock record-err

1 0

r440 - dev/olpc/kb3700
by svn＠openbios.org 07 Jun '07

07 Jun '07

Author: wmb Date: 2007-06-07 02:13:32 +0200 (Thu, 07 Jun 2007) New Revision: 440 Modified: dev/olpc/kb3700/ecio.fth Log: OLPC ecio.fth - In wlan-reset, added a workaround for an EC bug. Hopefully the EC bug will be fixed soon and we can remove the workaround. Modified: dev/olpc/kb3700/ecio.fth =================================================================== --- dev/olpc/kb3700/ecio.fth 2007-06-07 00:11:52 UTC (rev 439) +++ dev/olpc/kb3700/ecio.fth 2007-06-07 00:13:32 UTC (rev 440) @@ -115,6 +115,12 @@ \ the port 66 command sequence. : sci-quiet ( -- ) h# 50 h# ff03 ec! ; +.( Redefining wlan-reset to add temporary workaround) cr +: wlan-reset ( -- ) + h# fd h# fc15 ec! + wlan-reset +; + \ While accessing the SPI FLASH, we have to turn off the keyboard controller, \ because it continuously fetches from the SPI FLASH when it's on. That \ interferes with our accesses.

1 0

r439 - cpu/x86/pc/olpc
by svn＠openbios.org 07 Jun '07

07 Jun '07

Author: wmb Date: 2007-06-07 02:11:52 +0200 (Thu, 07 Jun 2007) New Revision: 439 Modified: cpu/x86/pc/olpc/resume.bth Log: OLPC resume - don't clear the wakeup status bits in PM1_STS (ACPI+0) in the assembly code, so high-level code can look at them. Modified: cpu/x86/pc/olpc/resume.bth =================================================================== --- cpu/x86/pc/olpc/resume.bth 2007-06-05 19:18:50 UTC (rev 438) +++ cpu/x86/pc/olpc/resume.bth 2007-06-07 00:11:52 UTC (rev 439) @@ -325,9 +325,6 @@ si bx cmp = until - h# 1840 # dx mov \ PM1_STS/EN I/O port number - dx ax in ax dx out \ Preserve EN bits and clear STS bits - h# 38 [bp] si lea \ Save area ax lods ax dx xchg ax lods ax dx xchg h# 4c00.0014 wmsr \ RSTPLL

1 0

r438 - dev
by svn＠openbios.org 05 Jun '07

05 Jun '07

Author: wmb Date: 2007-06-05 21:18:50 +0200 (Tue, 05 Jun 2007) New Revision: 438 Modified: dev/pckbd.fth Log: OLPC - use an abbreviated init sequence for the OLPC keyboard, to avoid startup failures with the EC wants to send us events from non-keyboard sources. Modified: dev/pckbd.fth =================================================================== --- dev/pckbd.fth 2007-06-05 19:15:23 UTC (rev 437) +++ dev/pckbd.fth 2007-06-05 19:18:50 UTC (rev 438) @@ -335,9 +335,24 @@ false ; +: ?olpc-keyboard ( -- ) + " enable-intf" $call-parent + begin get-data? while + drop + true to keyboard-present? + 5 ms + repeat + keyboard-present? if exit then + kbd-reset 0= to keyboard-present? +; + : reset ( -- ) init-data - clear-state get-initial-state + clear-state +[ifdef] olpc + ?olpc-keyboard +[else] + get-initial-state \ Test the keyboard interface clock and data lines " test-lines" $call-parent if false to keyboard-present? exit then @@ -354,6 +369,7 @@ \ 7 set-leds d# 100 ms 0 set-leds then +[then] \ Leave the keyboard in scan set 2 (its default state), but also leave \ the 8042 in the mode where it translates to scan set 1.

1 0

r437 - cpu/x86/pc/olpc/sdtools
by svn＠openbios.org 05 Jun '07

05 Jun '07

Author: wmb Date: 2007-06-05 21:15:23 +0200 (Tue, 05 Jun 2007) New Revision: 437 Modified: cpu/x86/pc/olpc/sdtools/maketar cpu/x86/pc/olpc/sdtools/sd.fth Log: OLPC sdkit - added EC access routines, put all the files in a subdirectory. Modified: cpu/x86/pc/olpc/sdtools/maketar =================================================================== --- cpu/x86/pc/olpc/sdtools/maketar 2007-06-04 23:30:10 UTC (rev 436) +++ cpu/x86/pc/olpc/sdtools/maketar 2007-06-05 19:15:23 UTC (rev 437) @@ -1,9 +1,11 @@ #!/bin/sh -ls ../../../../../cpu/x86/Linux -cp ../../../../../cpu/x86/Linux/forth . -cp ../../../../../cpu/x86/build/builder.dic . -cp ../../../../../cpu/x86/ioports.fth . -cp ../../../../../cpu/x86/pc/olpc/mfgdata.fth . -cp ../../../../../dev/olpc/dcon/dconsmb.fth . -tar cfz sdkit.tgz forth builder.dic sd.fth ioports.fth dconsmb.fth mfgdata.fth dumpgamma.fth sd -rm -f builder.dic forth ioports.fth mfgdata.fth dconsmb.fth +mkdir -p sdkit +cp ../../../../../cpu/x86/Linux/forth sdkit +cp ../../../../../cpu/x86/build/builder.dic sdkit +cp ../../../../../cpu/x86/ioports.fth sdkit +cp ../../../../../cpu/x86/pc/olpc/mfgdata.fth sdkit +cp ../../../../../dev/olpc/dcon/dconsmb.fth sdkit +cp ../../../../../dev/olpc/kb3700/ecio.fth sdkit +cp sd sd.fth dumpgamma.fth sdkit +tar cfz sdkit.tgz sdkit +# rm sdkit/* Modified: cpu/x86/pc/olpc/sdtools/sd.fth =================================================================== --- cpu/x86/pc/olpc/sdtools/sd.fth 2007-06-04 23:30:10 UTC (rev 436) +++ cpu/x86/pc/olpc/sdtools/sd.fth 2007-06-05 19:15:23 UTC (rev 437) @@ -34,7 +34,8 @@ : r ( offset -- ) sdl@ u. ; : w ( l offset -- ) sdl! ; -fff0.0000 10.0000 mmap constant flash-base +-1 value flash-base +fff0.0000 10.0000 mmap to flash-base \needs cdump : cdump ( adr len -- ) bounds ?do i c@ .x loop ; \needs .mfg-data fload mfgdata.fth @@ -60,6 +61,12 @@ \needs vp@ fload dumpgamma.fth +: unimp true abort" SPI reflashing is not implemented in this version" ; +: spicmd! unimp ; : spi-cmd-wait unimp ; : power-off unimp ; +defer spi-start defer spi@ defer spi! defer spi-out defer spi-reprogrammed +1 value spi-us +\needs ec@ fload ecio.fth + \ LICENSE_BEGIN \ Copyright (c) 2006 FirmWorks \

1 0

r436 - cpu/x86/pc/olpc
by svn＠openbios.org 05 Jun '07

05 Jun '07

Author: wmb Date: 2007-06-05 01:30:10 +0200 (Tue, 05 Jun 2007) New Revision: 436 Modified: cpu/x86/pc/olpc/fw.bth Log: OLPC - (a) Don't freeze display by default (b) Play startup only if left-side button held (c) The only game key that invokes interactive mode is X (above power button) Modified: cpu/x86/pc/olpc/fw.bth =================================================================== --- cpu/x86/pc/olpc/fw.bth 2007-06-04 23:27:29 UTC (rev 435) +++ cpu/x86/pc/olpc/fw.bth 2007-06-04 23:30:10 UTC (rev 436) @@ -318,7 +318,10 @@ h# 1454 pl@ 2 and if ." Interacting due to botched resume" cr hex interact then - i-key-wait if ." Interacting" cr hex interact then + i-key-wait if + " patch false fast-boot? startup" eval + ." Interacting" cr hex interact + then [ifdef] lx-devel init-superio [then] @@ -434,17 +437,19 @@ \ fload ${BP}/cpu/x86/pc/report.fth fload ${BP}/ofw/core/dualconsole.fth -: user-switch? ( -- flag ) +0 value game-key-mask + +: read-game-keys ( -- f) [ifdef] lx-devel false exit [then] atest? if true exit then - game-key@ 0<> dup if + game-key@ dup to game-key-mask if ." Release the game key to continue" cr begin d# 100 ms game-key@ 0= until then ; : fast-boot? ( -- flag ) [ifdef] rom-loaded - user-switch? 0= + game-key-mask h# 100 and 0= [else] false [then] @@ -510,6 +515,9 @@ r> to playback-volume ; +: ?start-sound ( -- ) + game-key-mask h# 40 and if start-sound then +; : open-keyboard ( -- ) " keyboard" open-dev ?dup if set-stdin then ; @@ -527,25 +535,25 @@ no-page console-start + read-game-keys fast-boot? if stdout off probe-pci - start-sound + ?start-sound ['] false to interrupt-auto-boot? interpreter-init [ifndef] lx-devel enable-power-button [then] - freeze auto-boot screen-ih stdout ! open-keyboard optional-startup else banner - + probe-pci + ?start-sound open-keyboard - probe-pci optional-startup interpreter-init [ifndef] lx-devel

1 0

r435 - in dev: geode olpc/kb3700 olpc/spiflash
by svn＠openbios.org 05 Jun '07

05 Jun '07

Author: wmb Date: 2007-06-05 01:27:29 +0200 (Tue, 05 Jun 2007) New Revision: 435 Modified: dev/geode/acpi.fth dev/olpc/kb3700/ecio.fth dev/olpc/spiflash/spiui.fth Log: OLPC - Fixes for new EC command protocol - see trac #1562. Modified: dev/geode/acpi.fth =================================================================== --- dev/geode/acpi.fth 2007-06-04 23:21:02 UTC (rev 434) +++ dev/geode/acpi.fth 2007-06-04 23:27:29 UTC (rev 435) @@ -19,8 +19,12 @@ \ If the keyboard controller is off (after "flash"), power off doesn't work. \ I suspect that is because the EC doesn't notice the deassertion \ of main_on and sus_on from the 5536. - kbc-on + ec-power-off + \ The rest of this will succeed in turning off the CPU, but the EC will + \ stay on. The ec-power-off above turns off both the EC and CPU, so the + \ rest of this is for historical interest only. + \ Recipe from AMD; no way I would have figured this out from manual 5 d# 10 << 1 or 8 acpi-w! \ S5 - power off Modified: dev/olpc/kb3700/ecio.fth =================================================================== --- dev/olpc/kb3700/ecio.fth 2007-06-04 23:21:02 UTC (rev 434) +++ dev/olpc/kb3700/ecio.fth 2007-06-04 23:27:29 UTC (rev 435) @@ -21,12 +21,6 @@ \ EC internal addresses -h# fc2a constant GPIO5 - -: flash-write-protect-on ( -- ) GPIO5 ec@ h# 80 invert and GPIO5 ec! ; - -: flash-write-protect-off ( -- ) GPIO5 ec@ h# 80 or GPIO5 ec! ; - : wait-ib-empty ( -- ) d# 1000 0 do h# 6c pc@ 2 and 0= if unloop exit then 1 ms loop true abort" EC port 6c input buffer timeout" @@ -70,25 +64,25 @@ : bat-gauge-id@ ( -- sn0 .. sn7 ) h# 17 ec-cmd-out 8 0 do ec-rb loop ; : bat-gauge@ ( -- b ) h# 18 ec-cmd-out h# 31 ec-wb ec-rb ; \ 31 is the EEPROM address : board-id@ ( -- b ) h# 19 ec-cmd-b@ ; -: sci-source@ ( -- b ) h# 1a ec-cmd-w@ ; +: sci-source@ ( -- b ) h# 1a ec-cmd-b@ ; : sci-mask! ( b -- ) h# 1b ec-cmd-b! ; : sci-mask@ ( -- b ) h# 1c ec-cmd-b@ ; : game-key@ ( -- w ) h# 1d ec-cmd-out ec-rw ; : ec-date! ( day month year -- ) h# 1e ec-cmd-out ec-wb ec-wb ec-wb ; -: ec-abnormal@ ( -- b ) h# 1f ec-cmd-b@ ; \ XXX is this byte or word? +: ec-abnormal@ ( -- b ) h# 1f ec-cmd-b@ ; : bat-init-lifepo4 ( -- ) h# 21 ec-cmd-out ; : bat-init-nimh ( -- ) h# 22 ec-cmd-out ; : wlan-off ( -- ) 0 h# 23 ec-cmd-b! ; : wlan-on ( -- ) 1 h# 23 ec-cmd-b! ; : wlan-wake ( -- ) h# 24 ec-cmd-out ; -: wlan-rst ( -- ) h# 25 ec-cmd-out ; +: wlan-reset ( -- ) h# 25 ec-cmd-out ; : dcon-disable ( -- ) 0 h# 26 ec-cmd-b! ; : dcon-enable ( -- ) 1 h# 26 ec-cmd-b! ; : reset-ec-warm ( -- ) h# 27 ec-cmd-out ; : reset-ec ( -- ) h# 28 ec-cmd-out ; : write-protect-fw ( -- ) h# 29 ec-cmd-out ; -: ebook-mode? ( -- ) h# 2a ec-cmd-b@ ; +: ebook-mode? ( -- b ) h# 2a ec-cmd-b@ ; \ EC mailbox access words @@ -117,8 +111,6 @@ : kb3920? ( -- flag ) h# 6c pc@ h# ff = if true exit then 9 ec-cmd 9 = ; -: snoop-board-id@ ( -- id ) h# fa20 ec@ ; - \ This makes the EC stop generating a flood of SCIs every time you do \ the port 66 command sequence. : sci-quiet ( -- ) h# 50 h# ff03 ec! ; @@ -135,14 +127,39 @@ true to kbc-off? ; +: ec-power-off ( -- ) + 1 h# ff14 ec! + 1 h# ff01 ec! \ Tell the EC to bounce us on restart + 0 h# ff14 ec! +; + \ Unfortunately, since the system reset is mediated by the keyboard \ controller, turning the keyboard controller back on resets the system. : kbc-on ( -- ) h# ff14 ec@ 1 invert and h# ff14 ec! \ Innocuous if already on + false to kbc-off? ; +: fancy-kbc-on ( -- ) + 2 h# ff01 ec! \ Prevent full system reset + h# ff14 ec@ 1 invert and h# ff14 ec! \ Innocuous if already on + d# 400 ms \ Give the EC time to start up + h# 44 h# fc80 ec! \ Re-enable scan code conversion and system flag + h# 43 h# fc81 ec! \ Re-enable fast gate and ibf/obf interrupts + h# f h# fe51 ec! \ Clear pending interrupts + h# f h# fe50 ec! \ Re-enable GPT interrupts + h# 0 h# fea7 ec! \ Clear FLASH write enable + h# 4 h# fead ec! \ fast read mode + h# fe95 ec@ h# 80 or h# fe95 ec! \ Write-protect LPC + h# 3 h# ff04 ec! \ enable IBF/OBF interrupts +\ h# 40 h# ff08 ec! \ Clear some interrupt + h# ff30 ec@ h# 10 or h# ff30 ec! \ Enable interrupt from GPIO04, PCI_RST# + + false to kbc-off? +; + \ kbc-pause temporarily halts execution of the keyboard controller microcode. \ kbc-resume makes it run again, picking up where it left off. \ This is useful for accessing the SPI FLASH in cases where you do not @@ -151,21 +168,6 @@ : kbc-pause ( -- ) h# dd ec-cmd66 ; : kbc-resume ( -- ) h# df ec-cmd66 ; - - -: kbd-led-on ( -- ) h# fc21 ec@ 1 invert and h# fc21 ec! ; -: kbd-led-off ( -- ) h# fc21 ec@ 1 or h# fc21 ec! ; - -: wlan-reset ( -- ) - \ WLAN reset is EC GPIOEE, controlled by EC registers fc15 and fc25 - h# fc15 ec@ h# fc25 ec@ ( enable data ) - dup h# 40 invert and h# fc25 ec! ( enable data ) \ WLAN_RESET data output low - over h# 40 or h# fc15 ec! ( enable data ) \ Assert output enable - 1 ms - h# 40 or h# fc25 ec! ( enable ) \ Drive data high - h# 40 invert and h# fc15 ec! ( ) \ Release output enable -; - : io-spi@ ( reg# -- b ) h# fea8 + ec@ ; : io-spi! ( b reg# -- ) h# fea8 + ec! ; @@ -179,6 +181,8 @@ : io-spi-reprogrammed ( -- ) ." Powering off..." d# 2000 ms cr power-off +\ kbc-on +\ ." Keyboard back on" cr ; : io-spi-start ( -- ) Modified: dev/olpc/spiflash/spiui.fth =================================================================== --- dev/olpc/spiflash/spiui.fth 2007-06-04 23:21:02 UTC (rev 434) +++ dev/olpc/spiflash/spiui.fth 2007-06-04 23:27:29 UTC (rev 435) @@ -236,7 +236,7 @@ $get-file ; -: flash ( ["filename"] -- ) get-file reflash power-off ; +: flash ( ["filename"] -- ) get-file reflash ; [ifdef] dev dev /flash

1 0

r434 - forth/lib
by svn＠openbios.org 05 Jun '07

05 Jun '07

Author: wmb Date: 2007-06-05 01:21:02 +0200 (Tue, 05 Jun 2007) New Revision: 434 Modified: forth/lib/breakpt.fth Log: Fixed single-stepping breakage resulting from no longer loading fixkbd.fth Modified: forth/lib/breakpt.fth =================================================================== --- forth/lib/breakpt.fth 2007-06-01 22:56:19 UTC (rev 433) +++ forth/lib/breakpt.fth 2007-06-04 23:21:02 UTC (rev 434) @@ -26,7 +26,7 @@ \ nuser restartable? restartable? off defer restart ( -- ) -defer restart-step ( -- ) +defer restart-step ( -- ) ' restart to restart-step hidden definitions

1 0

Re: [OpenBIOS] r433 - dev/usb2/device/wlan
by Segher Boessenkool 02 Jun '07

02 Jun '07

> +: reset-wlan ( -- ) " wlan-reset" evaluate ; Can't you use DEFER , or at least put a CATCH around the EVALUATE ? Segher

1 0

r433 - dev/usb2/device/wlan
by svn＠openbios.org 02 Jun '07

02 Jun '07

Author: wmb Date: 2007-06-02 00:56:19 +0200 (Sat, 02 Jun 2007) New Revision: 433 Modified: dev/usb2/device/wlan/usb8388.fth Log: OLPC WLAN driver - Don't implement EC-based WLAN reset directly, instead call out to a reset routine defined elsewhere. Modified: dev/usb2/device/wlan/usb8388.fth =================================================================== --- dev/usb2/device/wlan/usb8388.fth 2007-06-01 18:20:38 UTC (rev 432) +++ dev/usb2/device/wlan/usb8388.fth 2007-06-01 22:56:19 UTC (rev 433) @@ -510,13 +510,7 @@ \ Reset \ ========================================================================= -: reset-wlan ( -- ) - fc25 dup " ec@" evaluate 40 invert and swap " ec!" evaluate \ Assert WLAN_RESET# - fc15 dup " ec@" evaluate 40 or swap " ec!" evaluate \ Enable output - 100 ms - fc15 dup " ec@" evaluate 40 invert and swap " ec!" evaluate \ Disable output - 5 ms -; +: reset-wlan ( -- ) " wlan-reset" evaluate ; \ ========================================================================= \ MAC address

1 0