diff options
Diffstat (limited to 'docs/release.html')
| -rw-r--r-- | docs/release.html | 477 |
1 files changed, 449 insertions, 28 deletions
diff --git a/docs/release.html b/docs/release.html index 15b2342..b11cf30 100644 --- a/docs/release.html +++ b/docs/release.html @@ -45,7 +45,7 @@ <h2>Releases</h2> <ul> - <li><a href="#release6">6th release</a> (2014 August 11th, <b>pre-release, 5th beta</b>)</h1> + <li><a href="#release6">6th release</a> (2014 September 3rd <b>pre-release, 6th beta</b>) (see <a href="#critical_tasks">critical</a> and <a href="#noncritical_tasks">non-critical</a> tasks)</h1> <li><a href="#release5">5th release</a> (2014 March 7th, revised 2014 June 22nd)</h1> <li><a href="#release4">4th release</a> (2014 February 21st)</h1> <li><a href="#release3">3rd release</a> (2013 December 14th)</h1> @@ -65,7 +65,7 @@ <hr/> - <h1 id="release6">6th release (pre-release, 5th beta)</h1> + <h1 id="release6">6th release (pre-release, 6th beta)</h1> <ul> <li>Released 2014 July 11th (pre-release) 1st beta</li> @@ -73,6 +73,7 @@ <li>Revised (pre-release, 3rd beta) 2014 July 20th</li> <li>Revised (pre-release, 4th beta) 2014 July 29th</li> <li>Revised (pre-release, 5th beta) 2014 August 11th (corrected 2014 August 11th)</li> + <li>Revised (pre-release, 6th beta) 2014 September 3rd</li> </ul> <div class="important"> @@ -87,21 +88,13 @@ <h2>Binaries (for flashing) (right-click save as, or use wget)</h2> <ul> - <li><a href="http://libreboot.org/release/5/prerelease_beta5/correction/libreboot_bin.tar.gz">http://libreboot.org/release/5/prerelease_beta5/correction/libreboot_bin.tar.gz</a></li> - <li><a href="http://libreboot.org/release/5/prerelease_beta5/correction/libreboot_bin.tar.gz.sig">http://libreboot.org/release/5/prerelease_beta5/correction/libreboot_bin.tar.gz.sig</a></li> - <li><a href="http://libreboot.org/release/5/prerelease_beta5/correction/libreboot_bin.tar.gz.sha512sum.txt">http://libreboot.org/release/5/prerelease_beta5/correction/libreboot_bin.tar.gz.sha512sum.txt</a></li> + <li><a href="http://libreboot.org/release/5/prerelease_beta6/libreboot_bin.tar.xz">http://libreboot.org/release/5/prerelease_beta6/libreboot_bin.tar.xz</a></li> + <li><a href="http://libreboot.org/release/5/prerelease_beta6/libreboot_bin.tar.xz.sig">http://libreboot.org/release/5/prerelease_beta6/libreboot_bin.tar.xz.sig</a></li> </ul> <h2>Source code (for hacking) (right-click save as, or use wget)</h2> <ul> - <li><a href="http://libreboot.org/release/5/prerelease_beta5/correction/libreboot_src.tar.gz">http://libreboot.org/release/5/prerelease_beta5/correction/libreboot_src.tar.gz</a></li> - <li><a href="http://libreboot.org/release/5/prerelease_beta5/correction/libreboot_src.tar.gz.sig">http://libreboot.org/release/5/prerelease_beta5/correction/libreboot_src.tar.gz.sig</a></li> - <li><a href="http://libreboot.org/release/5/prerelease_beta5/correction/libreboot_src.tar.gz.sha512sum.txt">http://libreboot.org/release/5/prerelease_beta5/correction/libreboot_src.tar.gz.sha512sum.txt</a></li> - </ul> - <h2>Metadata (for re-creating the source archive) (right-click save as, or use wget)</h2> - <ul> - <li><a href="http://libreboot.org/release/5/prerelease_beta5/correction/libreboot_meta.tar.gz">http://libreboot.org/release/5/prerelease_beta5/correction/libreboot_meta.tar.gz</a></li> - <li><a href="http://libreboot.org/release/5/prerelease_beta5/correction/libreboot_meta.tar.gz.sig">http://libreboot.org/release/5/prerelease_beta5/correction/libreboot_meta.tar.gz.sig</a></li> - <li><a href="http://libreboot.org/release/5/prerelease_beta5/correction/libreboot_meta.tar.gz.sha512sum.txt">http://libreboot.org/release/5/prerelease_beta5/correction/libreboot_meta.tar.gz.sha512sum.txt</a></li> + <li><a href="http://libreboot.org/release/5/prerelease_beta6/libreboot_src.tar.xz">http://libreboot.org/release/5/prerelease_beta6/libreboot_src.tar.xz</a></li> + <li><a href="http://libreboot.org/release/5/prerelease_beta6/libreboot_src.tar.xz.sig">http://libreboot.org/release/5/prerelease_beta6/libreboot_src.tar.xz.sig</a></li> </ul> <h2>Machines still supported (compared to previous release):</h2> @@ -130,6 +123,13 @@ <li>See notes below for exceptions, and <a href="../docs/index.html#supported_t60_list">../docs/index.html#supported_t60_list</a> for known working LCD panels.</li> <li>It is unknown whether a T61 can have it's mainboard replaced with a T60 motherboard.</li> <li>See <a href="../docs/future/index.html#t60_cpu_microcode">../docs/future/index.html#t60_cpu_microcode</a>.</li> + <li>T60p (and T60's with ATI GPU) will likely never be supported: <a href="../docs/index.html#t60_ati_intel">../docs/index.html#t60_ati_intel</a></li> + </ul> + </li> + <li> + <b>Apple MacBook1,1</b> (MA255LL/A, MA254LL/A, MA472LL/A) + <ul> + <li>See <a href="../docs/index.html#macbook11">../docs/index.html#macbook11</a>.</li> </ul> </li> <li> @@ -468,6 +468,7 @@ </li> <li>Re-added .git files to bucts</li> <li>Fixed the oversight where macbook21_firstflash wasn't included in binary archives</li> + <li>Release archives are now compressed using .tar.xz for better compression</li> </ul> <h2> @@ -478,25 +479,311 @@ <li>Fixed incorrect grub.cfg that was actually placed in resources/grub/config/x60/grub_usqwerty.cfg which broke the default GRUB menu entry on X60</li> </ul> + <h2> + Revisions for 6th beta (2014 September 3rd) + </h2> + <ul> + <li>Added modified builddeb* scripts for Parabola GNU/Linux-libre: buildpac, buildpac-flashrom, buildpac-bucts (courtesy of Noah Vesely)</li> + <li>Documentation: updated all relevant areas to mention use of buildpac* scripts for Parabola users.</li> + <li>Documentation: added information showing how to enable or disable bluetooth on the X60</li> + <li>MacBook1,1 tested! See <a href="../docs/index.html#macbook11">../docs/index.html#macbook11"</a></li> + <li>Documentation: fixed typo in ../docs/index.html#get_edid_panelname (get-edit changed to get-edid)</li> + <li>Documentation: added ../docs/howtos/x60_lcd_change/ (pics only for now)</li> + <li>Added gcry_serpent and gcry_whirlpool to the GRUB module list in the 'build' script (for luks users)</li> + <li> + <b>Libreboot is now based on a new coreboot version from August 23rd, 2014:<br/> + Merged commits (relates to boards that were already supported in libreboot):</b> + <ul> + <li><a href="http://review.coreboot.org/#/c/6697/">http://review.coreboot.org/#/c/6697/</a></li> + <li><a href="http://review.coreboot.org/#/c/6698/">http://review.coreboot.org/#/c/6698/</a> (merged already)</li> + <li><a href="http://review.coreboot.org/#/c/6699/">http://review.coreboot.org/#/c/6699/</a> (merged already)</li> + <li><a href="http://review.coreboot.org/#/c/6696/">http://review.coreboot.org/#/c/6696/</a> (merged already)</li> + <li><a href="http://review.coreboot.org/#/c/6695/">http://review.coreboot.org/#/c/6695/</a> (merged already)</li> + <li><b><a href="http://review.coreboot.org/#/c/5927/">http://review.coreboot.org/#/c/5927/</a> (merged already)</b></li> + <li><a href="http://review.coreboot.org/#/c/6717/">http://review.coreboot.org/#/c/6717/</a> (merged already)</li> + <li><a href="http://review.coreboot.org/#/c/6718/">http://review.coreboot.org/#/c/6718/</a> (merged already)</li> + <li> + <a href="http://review.coreboot.org/#/c/6723/">http://review.coreboot.org/#/c/6723/</a> (merged already) + (text-mode patch, might enable memtest. macbook21) + </li> + <li> + <a href="http://review.coreboot.org/#/c/6732/">http://review.coreboot.org/#/c/6732/</a> (MERGED) + (remove useless ps/2 keyboard delay from macbook21. already merged) + </li> + </ul> + </li> + <li> + These were also merged in coreboot (relates to boards that libreboot already supported): + <ul> + <li><a href="http://review.coreboot.org/#/c/5320/">http://review.coreboot.org/#/c/5320/</a> (merged)</li> + <li><a href="http://review.coreboot.org/#/c/5321/">http://review.coreboot.org/#/c/5321/</a> (merged)</li> + <li><a href="http://review.coreboot.org/#/c/5323/">http://review.coreboot.org/#/c/5323/</a> (merged)</li> + <li><a href="http://review.coreboot.org/#/c/6693/">http://review.coreboot.org/#/c/6693/</a> (merged)</li> + <li><a href="http://review.coreboot.org/#/c/6694/">http://review.coreboot.org/#/c/6694/</a> (merged)</li> + <li><a href="http://review.coreboot.org/#/c/5324/">http://review.coreboot.org/#/c/5324/</a> (merged)</li> + </ul> + </li> + <li>Documentation: removed the section about tft_brightness on X60 (new code makes it obsolete)</li> + <li>Removed all patches from resources/libreboot/patch/ and added new patch: 0000_t60_textmode.git.diff</li> + <li>Updated getcb script and DEBLOB script.</li> + <li>Updated configuration files under resources/libreboot/config/ to accomodate new coreboot version.</li> + <li> + Removed grub_serial*.cfg and libreboot_serial*.rom, all configs/rom's are now unified (containing same configuration as serial rom's from before). + <ul> + <li>Documentation: updated ../docs/index.html#rom to reflect the above.</li> + </ul> + </li> + <li>Updated GRUB to new version from August 14th, 2014.</li> + <li>Unified all grub configurations for all machines to a single grub.cfg under resources/grub/config/</li> + <li>Updated flashrom to new version from August 20th, 2014</li> + <li> + Added getseabios and builddeps-seabios (builddeps and getall were also updated) + <ul> + <li> + Added instructions to 'buildrom-withgrub' to include bios.bin.elf and vgaroms/vgabios.bin from SeaBIOS inside the ROM. + </li> + </ul> + </li> + <li>Added seabios (and sgavgabios) to grub as payload option in menu</li> + <li> + Disabled serial output in Memtest86+ (no longer needed) to speed up tests. + <ul> + <li>MemTest86+ now works properly, it can output on the laptop screen (no serial port needed anymore).</li> + </ul> + </li> + <li> + Added getgrubinvaders, builddeps-grubinvaders scripts. Added these to getall and builddeps. + <ul> + <li> + Added <a href="http://www.coreboot.org/GRUB_invaders">GRUB Invaders</a> menu entry in resources/grub/config/grub.cfg + </li> + </ul> + </li> + <li>Added rules to builddeps-coreboot to build libpayload with TinyCurses. (added appropriate instructions to cleandeps script).</li> + <li>Commented out lines in resources/grub/config/grub.cfg for loading font/background (not useful anymore, now that GRUB is in text-mode).</li> + <li>Commented out lines in buildrom-withgrub that included backgrounds/fonts (not useful anymore, now that GRUB is in text-mode).</li> + <li> + Added resources/utilities/i945-pwm/ (from git://git.mtjm.eu/i945-pwm), for debugging acpi brightness on i945 machines. + <ul> + <li>Added instructions for it in builddeps, builddeps-i945pwm, builddeb and cleandeps</li> + </ul> + </li> + <li>'build' script: removed the parts that generated sha512sum manifests (not needed, since release tarballs are GPG-signed)</li> + <li> + 'build' script: removed the parts that generated libreboot_meta directory (not needed anymore, since _meta will be hosted in git) + <ul> + <li>Updated ../docs/index.html#build_meta (and other parts of documentation) to accomodate this change.</li> + </ul> + </li> + <li>Documentation: simplified (refactored) the notes in ../docs/index.html#rom</li> + <li> + 'build' script: removed the parts that generated libreboot_bin and added them to a new script: 'build-release' + <ul> + <li>Documentation: ../docs/index.html#build updated to reflect the above.</li> + </ul> + </li> + <li>Removed 'sudo' from builddeb, builddeb-flashrom, powertop.trisquel6 and builddeb-bucts scripts (assuming that the user has it is a really bad idea).</li> + <li><s>Added all gcry_* modules to grub (luks/cryptomount): gcry_arcfour gcry_camellia gcry_crc gcry_dsa gcry_md4 gcry_rfc2268 gcry_rmd160 gcry_seed gcry_sha1 gcry_sha512 gcry_twofish gcry_blowfish gcry_cast5 gcry_des gcry_idea gcry_md5 gcry_rijndael gcry_rsa gcry_serpent gcry_sha256 gcry_tiger gcry_whirlpool</s></li> + <li>Added GNUtoo's list of GRUB modules (includes all of the gcry_* modules above), cryptomount should be working now.</li> + <li>Removed builddeb-bucts and builddeb-flashrom, merged them with builddeb (../docs/index.html updated accordingly)</li> + <li>Removed buildpac-bucts and buildpac-flashrom, merged them with buildpac (../docs/index.html updated accordingly)</li> + <li>Renamed builddeb to deps-trisquel (../docs/index.html updated accordingly)</li> + <li>Renamed buildpac to deps-parabola (../docs/index.html updated accordingly)</li> + <li>Documentation: removed all parts talking about build dependencies, replaced them with links to ../docs/index.html#build_dependencies</li> + <li>Documentation: emphasized more strongly on the documentation, the need to re-build bucts and/or flashrom before flashing a ROM image.</li> + <li>build-release: flashrom, nvramtool, cbfstool and bucts are no longer provided pre-compiled in binary archives, and are now in source form only. (to maximize distro compatibility).</li> + <li>Documentation: added ../docs/howtos/encrypted_trisquel.html showing how to setup a fully encrypted Trisquel installation (including /boot) and boot it from the GRUB payload.</li> + <li>'build' script: replaced grub.elf assembly instructons, it's now handled by a utility added under resources/utilities/grub-assemble</li> + <li>Moved resources/grub/keymap to resources/utilities/grub-assemble/keymap, and updated that utility to use it</li> + <li>Documentation: removed useless links to pictures of keyboard layouts and unmodified layouts.</li> + <li>Removed all unused fonts from dejavu-fonts-ttf-2.34/ directory</li> + <li>'buildrom-withgrub' script: updated it to create 2 sets of ROM's for each machine: one with text-mode, one with coreboot framebuffer.</li> + <li>Documentation: updated ../docs/index.html#rom to reflect the above</li> + <li>Deleted unused README and COPYING file from main directory</li> + <li>Removed some rm -rf .git* instructions from the get* scripts and moved them to build-release script</li> + <li> + Split up default grub.cfg into 6 parts: extra/{common.cfg,txtmode.cfg,vesafb.cfg} and menuentries/{common.cfg,txtmode.cfg,vesafb.cfg} + <ul> + <li>buildrom-withgrub script uses these to generate the correct grub.cfg for each type of configuration.</li> + </ul> + </li> + <li>grub_memdisk.cfg (used inside grub.elf) now only loads grub.cfg from cbfs. It no longer enables serial output or sets prefix. (menuentries/common.cfg does instead)</li> + <li> + resources/grub/config/extra/common.cfg, added: + <ul> + <li>insmod instructions to load those modules: nativedisk, ehci, ohci, uhci, usb, usbserial_pl2303, usbserial_ftdi, usbserial_usbdebug</li> + <li>set prefix=(memdisk)/boot/grub</li> + <li> + For native graphics (recommended by coreboot wiki):<br/> + gfxpayload=keep<br/> + terminal_output --append gfxterm + </li> + <li> + Play a beep on startup:<br/> + play 480 440 1 + </li> + </ul> + </li> + <li> + Documentation: added note about 'fb=false' workaround for text-mode debian-installer (Trisquel net install) to + ../docs/howtos/grub_boot_installer.html + </li> + <li> + Documentation: updated ../docs/howtos/grub_cbfs.html to make it safer (and easier) to follow. + </li> + </ul> + <div class="important"> - <h2> - Other tasks (to be completed before declaring the initial stable release) + + <h2 id="critical_tasks"> + Critical tasks (to be completed before declacing the initial stable release) </h2> <ul> <li> - <b><u><i>TODO:</i></u></b> Fix remaining incompatible LCD panels in native graphics on T60. - (see <a href="../docs/future/index.html#lcd_i945_incompatibility">../docs/future/index.html#lcd_i945_incompatibility</a>) and <b>submit patches upstream</b>. + <b>GIT! GIT! GIT! Setup a git repository for libreboot_meta. (on dev.libreboot.org): git init, git add, git commit</b> </li> <li> - <b><u><i>TODO:</i></u></b> Implement VBT on X60/T60/X60T (and macbook21?) - (see <a href="../docs/future/index.html#i945_vbt">../docs/future/index.html#i945_vbt</a> and - <a href="../docs/future/index.html#intelvbttool_results">../docs/future/index.html#intelvbttool_results</a>) and <b>submit changes upstream</b>. + <b>Look at notes and TODO's about untested changes from changelog in libreboot 6th release beta 6, and test them. include test results in next change log. + (and for those commits that become merged in coreboot, mention them)</b> </li> <li> - <b><u><i>TODO:</i></u></b> Submit 3D fix (for X60/T60/macbook21/X60T on kernel 3.12+) upstream for 5320 changeset. + Keep an eye on these commits for next rebase: + <ul> + <li><a href="http://review.coreboot.org/#/c/6789/">http://review.coreboot.org/#/c/6789/</a> <b>(MERGED!!!)</b></li> + <li><a href="http://review.coreboot.org/#/c/6790/">http://review.coreboot.org/#/c/6790/</a> <b>(MERGED!!!)</b></li> + <li><a href="http://review.coreboot.org/#/c/6792/">http://review.coreboot.org/#/c/6792/</a></li> + <li><a href="http://review.coreboot.org/#/c/6793/">http://review.coreboot.org/#/c/6793/</a></li> + <li><a href="http://review.coreboot.org/#/c/6788/">http://review.coreboot.org/#/c/6788/</a> <b>(MERGED!!!)</b></li> + <li><a href="http://review.coreboot.org/#/c/6813/">http://review.coreboot.org/#/c/6813/</a></li> + <li><a href="http://review.coreboot.org/#/c/6804/">http://review.coreboot.org/#/c/6804/</a></li> + <li><a href="http://review.coreboot.org/#/c/6817/">http://review.coreboot.org/#/c/6817/</a> <b>(MERGED!!!)</b></li> + <li><a href="http://review.coreboot.org/#/c/6820/">http://review.coreboot.org/#/c/6820/</a></li> + </ul> + </li> + <li> + <b>Also keep an eye on these; they are not yet merged in coreboot (need to test) but included in libreboot:</b> + <ul> + <li> + <a href="http://review.coreboot.org/#/c/6725">http://review.coreboot.org/#/c/6725</a> + (text-mode patch, might enable memtest. x60. TEST IT! <b>(Disable 'Keep VESA Framebuffer' in menuconfig)</b>) + <ul> + <li>a patch for T60 should be created based on the above. See <b>resources/libreboot/patch/0000_t60_textmode.git.diff</b> (submit it upstream!!)</li> + </ul> + </li> + <li> + <s><a href="http://review.coreboot.org/#/c/6720">http://review.coreboot.org/#/c/6720/</a> + (ACPI backlight patch for X60. TEST IT!)</s> + <ul> + <li><s>a patch for T60 should be created based on the above.</s></li> + <li>see below</li> + </ul> + </li> + <li> + ACPI backlight controls for i945 <b>(TEST IT!)</b>: + <ul> + <li><a href="http://review.coreboot.org/#/c/6730">http://review.coreboot.org/#/c/6730</a> (MERGED (was just removing unused code))</li> + <li><a href="http://review.coreboot.org/#/c/6731/">http://review.coreboot.org/#/c/6731/</a> (has uneven backlight. fixed in libreboot. phcoder says he will re-write it)</li> + </ul> + </li> + <li> + <a href="http://review.coreboot.org/#/c/5345/">http://review.coreboot.org/#/c/5345/</a> + (already in libreboot. it still needs to be cherry picked because it's not merged yet, there are bugs that need fixing) + </li> + <li> + TODO: figure out how to add VGA output (external monitor) support in i945 native graphics. + <ul> + <li>Gnutoo says that the seabios page on coreboot wiki might give some hints</li> + </ul> + </li> + </ul> + </li> + <li> + PTE errors still exist with 6718 (which is merged). <a href="http://review.coreboot.org/#/c/6718/">http://review.coreboot.org/#/c/6718/</a> + <ul> + <li><a href="future/dumps/pte_x60_6718/dmesg">dmesg</a></li> + <li><a href="future/dumps/pte_x60_6718/kern.log">kern.log</a></li> + </ul> + </li> + <li> + ACPI brightness (tested on X60/T60): left half of screen is darker than the right side, when turning down brightness. Investigate.<br/> + mtjm: I'll send you my very hacky program for changing brightness and PWM frequency, it makes testing easier<br/> + mtjm: git://git.mtjm.eu/i945-pwm.git, exponent = 4, divisor = 0 is what 6731 would do<br/> + Setting exponent 9 and divisor 0 is what libreboot is doing currently (as a modification to 6731) + <ul> + <li>mtjm says: ACPI spec appendix B video extensions documents what IDs do, igd.asl doesn't use them correctly</li> + <li>phcoder says: igd.asl is so broken. Mismatching IDs, listing methods returning garbage, phantom devices. Any one of those is enough to break ACPI spec and windows. We have all 3. On all chipsets.</li> + <li>6731 (currently cherry picked in libreboot) is -2'd at this time of writing 20140824 while phcoder works on improving it. (worked in GNU/Linux, though).</li> + <li><b>A temporary fix (for uneven backlight) is provided in libreboot. Work in coreboot to fix it properly (phcoder says src/northbridge/intel/i945/acpi/igd.asl needs re-written)</b></li> + </ul> + </li> + <li> + There are issues with i945 text mode graphics initialization:<br/> + <b>This might not be a problem at all, but instead because there is no framebuffer available when using text-mode. A workaround for + the debian-installer issue below, detailed in docs/howtos/grub_boot_installer.html</b> + <a href="http://www.coreboot.org/pipermail/coreboot/2014-August/078468.html">http://www.coreboot.org/pipermail/coreboot/2014-August/078468.html</a> + look into it, follow up on that post and try to fix it. + <ul> + <li>Trisquel isolinux menu doesn't show up in seabios (seavgabios) with text-mode, it just says 'Error setting up gfxboot', + but it works in text-mode with the extract vbios. Looking at gma.c, it looks like this isn't setup in text-mode by native init, + but obviously when using the extracted vbios, it is setting everything up properly?</li> + <li> + Debian/Trisquel net-install shows graphical corruption (see mailing list link) when booting directly from GRUB in text-mode, but + works just fine when using the extracted vbios instead of seavgabios. Are these using text-mode or trying + to use graphics? <b>(graphical installers work just fine in native init or with + extracted vbios)</b> + </li> + <li> + Debian net-installer (graphical one) fails (trisquel graphical installer is ok) in native graphics and text-mode whether using + native graphics or extracted vbios (works fine in vesa/cbfb): + <ul> + <li>Scrolling/flickering text in a loop (segmentation fault):</li> + <li>Xorg (xorg_backtrace+0x49) [0xb7*******] <i>(numbers change)</i></li> + <li>Xorg (0xb75******) [0xb7*******] <i>(numbers change</i></li> + <li>(vdso) (__kernel_rt_sigreturn+0x0) [0xb7******] <i>(numbers change)</i></li> + <li>/lib/libc.so.6 (cfree+0x49) [0xb7******] <i>(numbers change)</i></li> + <li>Xorg (xf86DeleteMode+0x51) [0xb7******] <i>(numbers change)</i></li> + <li>Segmentation fault at address 0xb7200000</li> + <li>Fatal server error:</li> + <li>Caught signal 11 (Segmentation fault). Server aborting</li> + </ul> + </li> + <li> + An already-installed (on the HDD) Trisquel or Debian system boots and works just fine with native graphics + text mode. + </li> + <li> + phcoder says that there are limitations in native graphic: for example, he says native init doesn't provide int10h at all, and that it lacks VBT + (sometimes referred to as VBE). + Are there other issues? He says that there are also lots of ACPI issues in general. + </li> + <li><b>Right now it is recommended to use the vesa/cbfb ROM's (these do not use text-mode, they use coreboot's vesa framebuffer instead). MemTest86+ + does work in text-mode, though.</b></li> + <li> + <b>Most (maybe all?) of the above are not a problem, I think.</b><br/> + mtjm: text mode is for payloads/OSes that use the 80x25 VGA text mode and do not change mode, or run KMS<br/> + mtjm: a more featureful VGA ROM is needed for payloads/OSes that use int10h interfaces to change modes (e.g. VESA drivers) (also, VBT and other fixes are needed)<br/> + Systems that have int10h (interrupt service) code available for reference:<br/> + Geode boards and QEMU (look in coreboot or seabios source code). + </li> + <li> + phcoder: paging will not help with mtrr granularity. This was already investigated<br/> + phcoder: interesting: coreboot exposed mistakes in fwts (firmware test suite). Like its mtrr check assumes that default=uncachable.<br/> + (might be related to those debian-installer issues in text-mode) + </li> + </ul> + </li> + <li> + About fixing remaining LCD panels on 5345:<br/> + 'polarity' is mentioned in coreboot log (cbmem -c). compare output (with working and non-working panel). (and see the other notes in docs/future/index.html)<br/> + phcoder says: hint for T60: it might be that failing panels are 8bpc<br/> + fchmmr: what does 8bpc mean? And what do you think the other (non-failing) panel are?<br/> + phcoder: 6bpc. bits per colour. May also be reffered as 18-bit vs 24-bit panels<br/> + phcoder: just collect EDIDs from failing and working panels<br/> + <b>phcoder gave me this for collecting EDID data: <a href="http://www.o2genum.com/2013/08/lp156wh2-tlaa-lcd-panel-edid.html">http://www.o2genum.com/2013/08/lp156wh2-tlaa-lcd-panel-edid.html</a></b> + </li> + <li> + <b><u><i>TODO:</i></u></b> Fix remaining incompatible LCD panels in native graphics on T60. + (see <a href="../docs/future/index.html#lcd_i945_incompatibility">../docs/future/index.html#lcd_i945_incompatibility</a>) and <b>submit patches upstream</b>. </li> - <li><b><u><i>TODO:</i></u></b> Re-tooled linux-libre deblob scripts for use with coreboot. <b>Deblobbing is already complete; completion of this additional task will simply make future deblobbing work easier.</b></li> - <li><b><u><i>TODO:</i></u></b> Remove mention of blob/3rdparty repository in Kconfig</li> <li> <b><u><i>TODO:</i></u></b> Modified coreboot to make wifi and trackpoint be enabled by default on the ThinkPad X60 (so no need for nvramtool) <ul> @@ -516,14 +803,135 @@ </li> </ul> </li> + </ul> + <p> + Apparently, only 8MB VRAM is available on i945 GPU's (though it could do 64MB):<br/> + phcoder: No. Hardware default is 8 MiB. When I wanted to make it configurable, I saw that docs mention only one other alternative: 1MiB. Later isn't event enough for 1024x768 at 24bpp without any acceleration or double buffering. It's possible there are undocumented values. Which options do you have in vendor BIOS? + How to find out how much vram you have:<br/> + phcoder: TOM - BSM<br/> + phcoder: check what vendor BIOS offers as options<br/> + fchmmr: I thought it could do 64MB usually<br/> + phcoder: not accorging to doc.<br/> + phcoder: see mobile-945-express-chipset-datasheet page 93<br/> + phcoder: see also src/northbridge/intel/i945/{early_init,northbridge,gma}.c<br/> + fchmmr: "011 = DVMT (UMA) mode, 8 MB of memory pre-allocated for<br/> + fchmmr: frame buffer."<br/> + fchmmr: "Others - reserved"<br/> + phcoder: the easiest way is a loop at this position which tries different values and reads (and prints) BSM with them<br/> + stefanct: fchmmr: he suggest that you change the value and look how BSM reacts to that<br/> + stefanct: as he pointed out earlier vram size = TOM - BSM<br/> + stefanct: different values of GMS<br/> + stefanct: phcoder: hm... this could be a hint. look at the text description of TOLUD at page 103<br/> + stefanct: it mentions 64 MB in the text about BSM as well<br/> + stefanct: table 18...<br/> + phcoder: stefanct: I have a guess which value make is 64 but I will not tell to avoid skewing test results<br/> + stefanct: phcoder: sure... i assumed you were not sure if it supports it at all. testing it properly is of course a good idea :)<br/> + stefanct: test the various possible (but reserved) values of GMS and see what the resulting VRAM size is<br/> + fchmmr: so, TOM - BSM + </p> + </div> + + <div class="important"> + <h2 id="noncritical_tasks"> + Other tasks (non-critical, but still important) + </h2> + <ul> <li> - <b><u><i>TODO:</i></u></b> make writing the cmos.layout to the rom default + Look into those machines: + <ul> + <li> + Acer CB5 (Chromebook 5) - I'm told there is still microcode in the GPU that needs RE'd, also the wifi + is soldered and requires non-free firmware (have to disable it in nvramtool and use usb wifi dongle) + </li> + <li> + ThinkPad X200 - I'm told the Roda RK9 (same chipset as X200) doesn't come with ME, so maybe X200 can have ME removed/disabled. + Also, we don't know if it can run without microcode updates yet. + </li> + <li> + Roda rk9? (see above) + </li> + <li> + ThinkPad R400 (if X200 experiment works). I'm told it's very similar to the X200 and could easily be ported. + </li> + <li> + + TODO: f2a85m and e350m1 rom's (libreboot_*_headless.rom) for headless operation. find out which PCI-E graphics cards can run without + the video bios (have to patch seabios, or patch coreboot (if using grub payload), in either case to load but not execute video bios: + then find what the driver for that card still needs, and copy/disassemble(understand) the data and write code that places that in memory) + --> maybe this could fuel a side-project (libre graphics project, for many different cards) + </li> + <li> + ASUS F2A85-M and ASrock E350M1 (desktop motherboards) + <ul> + <li> + The onboard graphics on both of these boards requires non-free video bios. Find a PCI-E graphics card that can + be used without running it's VBIOS (might still require loading data from it, which could be extracted + and studied) + </li> + <li> + <a href="future/f2a85m_dmidecode">dmidecode output for f2a85m</a> (still waiting for e350m1 dmidecode output). + </li> + </ul> + </li> + </ul> + </li> + <li> + Add 'set -e' to the start of all relevant scripts, to make them quit if an error occurs (that a command exits with non zero status). + See 'help set' for info. + </li> + <li> + Re-write deps-parabola script (again based on deps-trisquel scripts), really make it install all correct dependencies for Parabola + (on a fresh install, with only the 'base' repository installed). + <ul> + <li>Find what other scripts need fixed for this.</li> + <li>Make libreboot build in parabola</li> + </ul> + </li> + <li> + Add those to seabios: + <ul> + add ipxe (executable from seabios) + </ul> + </li> + <li> + Re-write the "Parse USB" options based on the "Search local HDD for grub.cfg" menu entry (make it more intelligent). </li> + <li>update info on x60/t60/macbook21 pages on coreboot wiki (not really a libreboot todo, but something that I should do anyway)</li> + <li>get all useful info from http://www.coreboot.org/GRUB2 and https://wiki.parabolagnulinux.org/User:GNUtoo/laptop and make it into a tutorial</li> + <li> + Test that commit and include it (maybe get it merged in coreboot): + <a href="http://review.coreboot.org/#/c/5242/17">http://review.coreboot.org/#/c/5242/17</a> (test this!) + </li> + <li> + Test that commit and include it (maybe get it merged in coreboot): + <a href="http://review.coreboot.org/#/c/5243/">http://review.coreboot.org/#/c/5243/</a> (test this!) + </li> + <li> + <s>Todo: (after rebase): include 6694 from review.coreboot.org (still some hanging issues). This is needed for reliable booting (for now) on macbook21</s> merged in master!<br/> + <b>work with phcoder to try and re-implement it properly (for example, ask him how to found those entries to add)</b><br/> + He says that the cst_arrays array was based on _CST object from vendor BIOS. Use acpidump and 'iasl -d' to find them. + </li> + <li> + Todo: test gnutoo's fallback patches: <a href="http://www.coreboot.org/Fallback_mechanism">http://www.coreboot.org/Fallback_mechanism</a> + (some parts use systemd. adapt for upstart, which is what trisquel uses)<br/> + see <a href="future/gnutoo_fallback_patch">this IRC log</a> (warning, not cleaned up) + </li> + <li> + are there known differences in /sys/class/power_supply between libreboot and vendor BIOS? or does the GNOME bug occur with vendor BIOS too? + (speaking about when battery is detected as "low" in Trisquel 6 / gnome when disconnecting AC/adapter) + </li> + <li> + <b><u><i>TODO:</i></u></b> Implement VBT on X60/T60/X60T (and macbook21?) + (see <a href="../docs/future/index.html#i945_vbt">../docs/future/index.html#i945_vbt</a> and + <a href="../docs/future/index.html#intelvbttool_results">../docs/future/index.html#intelvbttool_results</a>) and <b>submit changes upstream</b>. + </li> + <li><b><u><i>TODO:</i></u></b> Re-tooled linux-libre deblob scripts for use with coreboot. <b>Deblobbing is already complete; completion of this additional task will simply make future deblobbing work easier.</b></li> + <li><b>Alternatively, host actual libreboot deblobbed git tree. pehjota writes: If you want to permanently rewrite the Git history, see the "git filter-branch" command. Of course all commit IDs change with that, so it may not be ideal for you. But it can give you the full Git history, with no blobs anywhere (not even recoverable with "git reset" or "git checkout").</b></li> + <li><b><u><i>TODO:</i></u></b> Remove mention of blob/3rdparty repository in Kconfig</li> <li> <b><u><i>TODO:</i></u></b> T60: find (for rare buggy CPU's that are unstable without microcode updates) if there is a workaround: see <a href="../docs/future/index.html#t60_cpu_microcode">../docs/future/index.html#t60_cpu_microcode</a> </li> - <li><b><u><i>TODO:</i></u></b> Modify coreboot so that tft_brightness is 0xff by default (for fixing the looping issue automatically)</li> <li><b><u><i>TODO:</i></u></b> Modify the Kconfig options in coreboot so that the (non-existent, deleted) microcode is never loaded when building.</li> <li><b><u><i>TODO:</i></u></b> Ask Mono what license he wants to use for his pages (see notes in docs/index.html#macbook21) and then include a copy in the relevant part of the documentation. (note: the backups are not included for now, for this exact reason)</li> <li><b><u><i>TODO:</i></u></b> Retroactively GPG-sign the final revisions of each previous release. (SHA512 sums aswell).</li> @@ -569,6 +977,12 @@ <b><u><i>TODO:</i></u></b><a href="https://wiki.archlinux.org/index.php/TLP">https://wiki.archlinux.org/index.php/TLP</a> (example article) a user mentioned that this also removes the high-pitched noise (like powertop does). Must look into this. </li> + <li> + <b><u><i>TODO:</i></u></b> For images that use foreign layouts, create translated grub menus aswell. (i18n!). Modify the build scripts to do this automatically. + </li> + <li> + Add tint payload to GRUB: <a href="http://www.coreboot.org/Tint">http://www.coreboot.org/Tint</a> + </li> </ul> </div> @@ -876,8 +1290,15 @@ <hr/> <p> - Copyright © 2014 Francis Rowe, All Rights Reserved.<br/> - See <a href="license.html">license.html</a> for license conditions. + Copyright © 2014 Francis Rowe <info@gluglug.org.uk><br/> + This document is released under the Creative Commons Attribution-ShareAlike 4.0 International Public License and all future versions. + A copy of the license can be found at <a href="license.txt">license.txt</a>. + </p> + + <p> + This document is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See <a href="license.txt">license.txt</a> for more information. </p> </body> |