xserver-xorg-video-avivo: Changes to 'debian-experimental'
ChangeLog | 881 ++++++++++++++++++++++++++++++++++++++++++++++++
avivotool/avivotool.c | 192 +++++-----
configure.ac | 22 -
debian/changelog | 7
debian/control | 2
include/Makefile.am | 1
include/avivo.h | 7
include/avivo_chipset.h | 189 +++++-----
include/radeon_reg.h | 251 ++++++++++---
xorg/avivo.c | 100 ++++-
xorg/avivo_bios.c | 2
xorg/avivo_chipset.c | 562 ++++++++++++++++++------------
xorg/avivo_common.c | 1
xorg/avivo_crtc.c | 15
xorg/avivo_cursor.c | 2
xorg/avivo_output.c | 379 +++++++++-----------
xorg/avivo_state.c | 82 ++--
17 files changed, 1944 insertions(+), 751 deletions(-)
New commits:
commit f5e2b086ca84a40d909860797b3862f6c6b0b223
Author: Brice Goglin <bgoglin@debian.org>
Date: Thu Sep 27 20:39:08 2007 +0200
New upstream snapshot, built against Xserver 1.4.
diff --git a/ChangeLog b/ChangeLog
new file mode 100644
index 0000000..587ca71
--- /dev/null
+++ b/ChangeLog
@@ -0,0 +1,881 @@
+commit 7800016506da80b09a94279d17c0675b40791db0
+Author: Jerome Glisse <glisse@freedesktop.org>
+Date: Tue Jul 24 15:38:03 2007 +0200
+
+ avivo: rework output dpms handling.
+
+ This should have better behavior when dpms status change
+ is triggered.
+
+commit e03a0c86190f74962a6ff59602f274e27353d6ae
+Author: Jerome Glisse <glisse@localhost.localdomain>
+Date: Thu Jul 19 19:25:20 2007 +0200
+
+ avivo: fix previous commit on PLL (usual cut & paste bug introduction).
+
+commit 830e10b7fe41c58e2fb61c6778fe59be0febdfbd
+Author: Jerome Glisse <glisse@freedesktop.org>
+Date: Thu Jul 19 18:43:45 2007 +0200
+
+ avivo: change pll values computation algorithm.
+
+ No more infinite loop, simply goes over all values.
+ The constraint i put are based on several dump with
+ several different modes clock and barely resume to:
+ 1 < div1 < 7
+ div1 < div2 < div1 + 14
+ 1 < mul < 256
+ div1 * div2 should be as big as possible
+
+commit 559b7019a736c6fff19b63e4c4f3e646f00b5717
+Author: Jerome Glisse <glisse@freedesktop.org>
+Date: Wed Jul 18 19:07:34 2007 +0200
+
+ avivo: push missing bits from last commit.
+
+commit d779065b5ad8d3b30e4c91c8a3ad8935292e05c1
+Author: Jerome Glisse <glisse@freedesktop.org>
+Date: Wed Jul 18 18:53:51 2007 +0200
+
+ avivo: properly handle GPIO 0 (0x7E30) for i2c.
+
+commit d3564fe5f01a924018f206b9b911348ab6b0a3bb
+Author: Jerome Glisse <glisse@freedesktop.org>
+Date: Wed Jul 18 18:29:37 2007 +0200
+
+ avivotool: fix handling of gpio 0x7e30.
+
+commit 7da689181f473877d27cf9b9bb5fac0713b089b6
+Author: Jerome Glisse <glisse@freedesktop.org>
+Date: Wed Jul 18 01:05:53 2007 +0200
+
+ avivo: If EDID fails for LFP use timings from bios.
+
+ In case EDID fails we use bios timings if they are
+ present, if they are not present we should add a
+ function to extract timings from currently set mode.
+
+commit d075f407cdedc7eda490d4a24dab36aa6ef5ccc9
+Author: Jerome Glisse <glisse@freedesktop.org>
+Date: Wed Jul 18 00:57:06 2007 +0200
+
+ avivotool: add i2c bus initialization should make i2c more reliable.
+
+commit de710630316f6280663eb08a576dc87a142fce41
+Author: Jerome Glisse <glisse@freedesktop.org>
+Date: Tue Jul 17 23:11:40 2007 +0200
+
+ avivotool: add missing arg to printf.
+
+commit ecddf9cee340069e9a739b2d5570d1d689529d05
+Author: Jerome Glisse <glisse@freedesktop.org>
+Date: Tue Jul 17 23:05:20 2007 +0200
+
+ avivotool: verbose i2c transaction.
+
+commit a7e43bd704dfb3957ace1f3d94e2c323364d443b
+Author: Jerome Glisse <glisse@freedesktop.org>
+Date: Tue Jul 17 22:40:03 2007 +0200
+
+ avivotool: 0x7e30 i2c line does not seems to work like other.
+
+commit 94758f61f44cb5c385eaba4005d1d2d3e1178b39
+Author: Jerome Glisse <glisse@freedesktop.org>
+Date: Tue Jul 17 20:44:13 2007 +0200
+
+ avivotool: dump whole 128bytes of edid.
+
+commit 5edcb95bb8a92e16c781678cbde76e15bccd77b0
+Author: Nicolas Trangez <eikke@eikke.com>
+Date: Tue Jul 17 20:33:12 2007 +0200
+
+ Changed avivotool so GPIO_IN and GPIO_OUT can be given as command line options.
+
+commit c818cd9f932d73c7e19f4a8232ca4dd15e24c8f6
+Author: Jerome Glisse <glisse@freedesktop.org>
+Date: Sat Jul 14 16:13:33 2007 +0200
+
+ avivo: fix 16 & 15 bits screen color depth.
+
+commit 18462de524f7abf56763dc04767104741067a06b
+Author: Alexander Larsson <alexl@redhat.com>
+Date: Sat Jul 14 15:45:58 2007 +0200
+
+ avivo: add shadow framebuffer support.
+
+commit bba2303c531e65f653b652c42e08d44ec144d93c
+Author: Jerome Glisse <glisse@freedesktop.org>
+Date: Wed Jul 11 01:14:28 2007 +0200
+
+ avivo: fix framebuffer offset handling.
+
+ We were missing one register for framebuffer offset.
+ Moreover avivo engine needs that x offset be a multiple
+ of 4. There is still a bug with virtual framebuffer,
+ i don't pick properly the current screen size.
+
+commit 9b6b0acf7263627a25546c4b6452ff3abb090793
+Author: Jerome Glisse <glisse@freedesktop.org>
+Date: Tue Jul 10 20:29:01 2007 +0200
+
+ avivo: properly use avivo scaler.
+
+commit 76ac52e0d240104d121377bdc3c1c90e4d546772
+Author: Jerome Glisse <glisse@freedesktop.org>
+Date: Tue Jul 10 19:40:28 2007 +0200
+
+ avivo: remap frame buffer & mmio if necessary in screen init.
+
+commit 9b0c0412de7a28318b5d7fadd1de12f0254208b1
+Author: Jerome Glisse <glisse@freedesktop.org>
+Date: Tue Jul 10 19:39:58 2007 +0200
+
+ avivo: adjust mode to match lfp fixed mode constraint.
+
+ As far as i can tell all lfp screen support only fixed
+ timing, so make sure we use this fixed timing.
+
+commit 284beece195a90e4bd232195c5493ab075a33962
+Author: Jerome Glisse <glisse@freedesktop.org>
+Date: Tue Jul 10 17:13:26 2007 +0200
+
+ avivo: make sure LFP overwritte second DVI output if one exist.
+
+ LFP connector use second TMDS so we have to make sure we do not
+ create an output for a second DVI-I output associated to second
+ TMDS otherwise LFP mode might be overwritten by this second DVI
+ output.
+
+commit 67f264b2fc5762f4940d395a45e72fc2ac682683
+Author: Jerome Glisse <glisse@freedesktop.org>
+Date: Tue Jul 10 17:06:06 2007 +0200
+
+ avivo: initialization code cleanup.
+
+commit 7f8ac4ad6906d2ac7115d060ea0b003dd435ac29
+Author: Jerome Glisse <glisse@freedesktop.org>
+Date: Tue Jul 10 16:30:44 2007 +0200
+
+ avivo: remove fb manager, might be reintroduced latter.
+
+commit c4cac297feb5a781fd832d305b7a59dc6723d15d
+Author: Jerome Glisse <glisse@freedesktop.org>
+Date: Tue Jul 10 16:29:21 2007 +0200
+
+ avivo: add all known pci id.
+
+commit 99d51bc0095c759019cfceb0af602da3d5c60d7b
+Author: Jerome Glisse <glisse@freedesktop.org>
+Date: Mon Jul 9 18:08:32 2007 +0200
+
+ avivo: add randr 1.2 shadow framebuffer callback for rotate support.
+
+commit bd10aee45c8352a6f496a7ffd63fbb6a2dbd38b6
+Author: Jerome Glisse <glisse@freedesktop.org>
+Date: Sun Jul 8 17:42:08 2007 +0200
+
+ avivo: simplify display width alignement computation.
+
+commit 9d27771f350099a2fdc6c6160e34c187ab5833a6
+Author: Jerome Glisse <glisse@freedesktop.org>
+Date: Sun Jul 8 17:20:53 2007 +0200
+
+ avivo: properly restore vga & GPU memory mapping on VT switch.
+
+ vga now should be properly reset when closing or switching btw
+ VT. Same for GPU memory mapping. This should fix issues reported
+ on VT switch.
+
+commit 8b28e959792caa4994d44b2c08780a06abe23571
+Author: Jerome Glisse <glisse@freedesktop.org>
+Date: Mon Jul 2 20:09:29 2007 +0200
+
+ avivo: properly set version and bump version to 0.0.999.
+
+commit 8a6fb613d6f18d33aef84460fac0fe59bc00f729
+Author: Jerome Glisse <glisse@freedesktop.org>
+Date: Mon Jul 2 19:50:23 2007 +0200
+
+ avivo: code cleanup.
+
+commit 214b9f80bf2b4935df8890cb9c6bb9aeb807ba80
+Author: Julien Cristau <jcristau@debian.org>
+Date: Mon Jul 2 19:06:40 2007 +0200
+
+ avivo: compilations fix from Julien Cristau & Ludovic Rousseau.
+
+commit c3d60875cd750d54f74b0cf7e06e551e01a723bd
+Author: Jerome Glisse <glisse@freedesktop.org>
+Date: Thu Jun 28 01:16:51 2007 +0200
+
+ avivo: fix typo
+
+commit 4bae344da3ebdc1952962ecacd5a80a7dde1dd0b
+Author: Jerome Glisse <glisse@freedesktop.org>
+Date: Thu Jun 28 01:12:23 2007 +0200
+
+ avivo: change pll computation to meet new constraint.
+
+ It seems that AVIVO_PLL_POST_DIVIDER * AVIVO_PLL_DIVIDER needs
+ to be above 40 for PLL stability, changed PLL computation to
+ meet this. This seems to improve image stability and should fix
+ couple of bad behavior.
+
+commit 14e502a548b5cb3b9bc8a0ba47317cd673806364
+Author: Jerome Glisse <glisse@freedesktop.org>
+Date: Wed Jun 27 18:05:07 2007 +0200
+
+ avivo: improve connector/output creation from bios.
+
+ Now LFP represent TMDS2 if it exist as LFP as to be
+ driven by TMDS2 so we cannot create a separate output
+ for TMDS2.
+
+commit fc0fd1494feeaa48703ca47597d865d2f93c56cb
+Author: Jerome Glisse <glisse@freedesktop.org>
+Date: Wed Jun 27 00:34:07 2007 +0200
+
+ avivo: print horizontal & vertical sync polarity flags.
+
+commit fc17fa764c7aef0232c21f5aa6ae2353d5efe7b7
+Author: Jerome Glisse <glisse@freedesktop.org>
+Date: Sun Jun 24 12:25:28 2007 +0200
+
+ avivo: fix a typo now truly assign LFP to TMDS2.
+
+commit 34b652d7f7f79e2e40ecb996a70cd51159b05965
+Author: Jerome Glisse <glisse@freedesktop.org>
+Date: Sun Jun 24 12:17:51 2007 +0200
+
+ avivo: properly set horizontal & vertical polarity.
+
+commit 9f40535f75b3dfb08efb6121294377e758070e56
+Author: Jerome Glisse <glisse@freedesktop.org>
+Date: Sat Jun 23 18:32:03 2007 +0200
+
+ avivo: increase driver verbosity.
+
+commit 9decf97be5f1aa5ae2001160fcc726236b398104
+Author: Jerome Glisse <glisse@freedesktop.org>
+Date: Fri Jun 22 16:41:04 2007 +0200
+
+ avivo: post randr 1.2 cleanup + always assignd LFP to TMDS2
+
+commit 2440ac567ce721378395c341cf0322e383bb4888
+Author: Jerome Glisse <glisse@freedesktop.org>
+Date: Fri Jun 22 15:47:39 2007 +0200
+
+ avivo: LFP should always be driven by second TMDS.
+
+commit f1de5482406ee01ace5d2c1871c5ec5e74567701
+Author: Jerome Glisse <glisse@freedesktop.org>
+Date: Thu Jun 21 10:22:27 2007 +0200
+
+ avivo: add missing extern to chipset tables declaration.
+
+commit b9771c7e26a1dbdcb7dd5cf3c525dad8f196eb6e
+Author: Jerome Glisse <glisse@freedesktop.org>
+Date: Tue Jun 19 12:49:51 2007 +0200
+
+ avivo: move chipset list in there own file.
+
+commit 3d630b3a9c878b97edf2f4aaa3e1b6a0da494427
+Author: Hans Ulrich Niedermann <hun@n-dimensional.de>
+Date: Mon Jun 18 20:15:51 2007 +0200
+
+ avivo: properly check for pciaccess & Xorg macro.
+
+commit a6d08ad97fede2cd89d1b793aad71421c8e158e3
+Author: Jerome Glisse <glisse@freedesktop.org>
+Date: Mon Jun 18 17:19:01 2007 +0200
+
+ avivo: improve pll computation & fix typo in PLL registers definitions.
+
+ pmul should not be > 255 and improved PLL values computation loop
+ break so that we break on first closest value. I still got some
+ flickering with dual screen so i guess we need to figure out more
+ of the PLL mystery as well as maybe some DAC & TMDS clock register.
+
+commit 8d01c28844dd5c434a648dfa7ec4da798224a8cb
+Author: Jerome Glisse <glisse@freedesktop.org>
+Date: Mon Jun 18 02:11:53 2007 +0200
+
+ avivo: improve PLL support (new register found & improved values computation).
+
+ Found the two PLL register there are still some mystery registers
+ with mystery values. Improved div, pdiv & mul pll computation so
+ that pll clock is above mode clock (seems to produce better result).
+
+commit ab39bf4513a99a4ecd5dd2a2eccb03432e24da51
+Author: Jerome Glisse <glisse@freedesktop.org>
+Date: Mon Jun 18 00:26:01 2007 +0200
+
+ avivo: make scan out buffer witdh * bpp 256 bytes aligned
+
+ X_LENGHT & Y_LENGHT is display buffer dimension (virtual dimension
+ if they are bigger than display mode dimension) and pitch should be
+ 256 bytes aligned.
+
+commit 052d51671716eb5a1909136bb87ab9cb1d9829cd
+Author: Jerome Glisse <glisse@freedesktop.org>
+Date: Mon Jun 18 00:03:29 2007 +0200
+
+ avivo: disable fb manager initialization
+
+ Don't know if this fb manager is usefull for something i disable
+ it now as it will get in the way of change needed to align scan
+ out buffer width to be 256 bytes aligned.
+
+commit d0d5ee700b53d376cbe2ac7f70e6d97cc076eb4a
+Author: Jerome Glisse <glisse@freedesktop.org>
+Date: Sun Jun 17 23:54:12 2007 +0200
+
+ avivo: preferably use crtc1 for feeding LFP so diffing with fglrx is easier.
+
+ This is not a hw restriction, it just easier then to compare with fglrx
+ as fglrx use crtc1 for LFP.
+
+commit 5e73799265df9e6984a241fa593902b52252c162
+Author: Jerome Glisse <glisse@freedesktop.org>
+Date: Sat Jun 16 22:01:02 2007 +0200
+
+ avivo: fix segfault on server exit (the video state are not yet properly restored)
+
+commit 7971ca48ce6d0edd4b45e685aad5711717a21e70
+Author: Jerome Glisse <glisse@freedesktop.org>
+Date: Sat Jun 16 21:50:12 2007 +0200
+
+ avivo: Change way we write to MC register.
+
+ It seems that i need to OR index with 0x00FF0000 and not
+ 0xFF7F0000 this OR value might be chipset version related
+ or might came from other register.
+
+commit 70e7f94fc8ed17c89cee8c12f8afd8eeffb77ef2
+Author: Jerome Glisse <glisse@freedesktop.org>
+Date: Sat Jun 16 21:48:20 2007 +0200
+
+ avivo: increase error message verbosity in screen init.
+
+commit 1723a579eb31b7244aa96159271d0ab1a774607d
+Author: Jerome Glisse <glisse@freedesktop.org>
+Date: Sat Jun 16 21:46:19 2007 +0200
+
+ avivo: wait avivo to be idle before updating PLL (safety purpose).
+
+ I don't really think this is necessary.
+
+commit 000b550bd7518227a51b071322af204297aa106a
+Author: Jerome Glisse <glisse@freedesktop.org>
+Date: Sat Jun 16 21:44:17 2007 +0200
+
+ avivo: fix MC memory map register is was given wrong location
+
+ The MC memory map register is at index 0x4 in MC register area.
+ Not index 0x1 this error is likely due to bug in radeondump which
+ reported the wrong index.
+
+commit 3708713e0510a424fb70edfd92c0e84f69377932
+Author: Jerome Glisse <glisse@freedesktop.org>
+Date: Fri Jun 15 16:29:02 2007 +0200
+
+ avivo: activate half working randr 1.2 support
+
+ Should be able to set properly mode on each output.
+ Have to decipher btw VGA & DVI using EDID digital or
+ analog link flag. MacBook Pro connector table seems
+ a bit strange and i still can't properly output mode
+ on an external monitor.
+
+commit 8f3e53376f07c5fa05eaf61626a8d498d852c667
+Author: Jerome Glisse <glisse@freedesktop.org>
+Date: Fri Jun 15 12:15:24 2007 +0200
+
+ avivo: randr1.2 fixes & PLL fixes.
+
+commit 52e78cf2d65b33ab23a6bea22caf6da68fa77768
+Author: Jerome Glisse <glisse@freedesktop.org>
+Date: Thu Jun 14 21:07:14 2007 +0200
+
+ avivo: randr1.2: initial support, not working enable with AVIVO_RR12
+
+ For some reason the server is not calling in driver crtc function
+ to set video mode thus launching with it will end up in a broken
+ display. This why this is disabled by default. Also randr 1.2 still
+ need work on rotation, allocating framebuffer, shadow buffer for
+ rotation, ...
+
+commit 61f20371eb3af031767776300587eb9dacd453ef
+Author: Jerome Glisse <glisse@freedesktop.org>
+Date: Thu Jun 14 18:19:49 2007 +0200
+
+ avivo: fix bugs introduced by randr change.
+
+commit ef115f143e6197aae9c2ed59c5759a095467eb28
+Author: Jerome Glisse <glisse@freedesktop.org>
+Date: Thu Jun 14 14:58:24 2007 +0200
+
+ avivo: crtc & output handling fixes.
+
+commit e9e9a0b35a31840a4b1bff14efd506d2c9e5ac13
+Author: Jerome Glisse <glisse@freedesktop.org>
+Date: Thu Jun 14 14:14:15 2007 +0200
+
+ avivo: randr 1.2: fix crtc & output handling (not activated yet)
+
+commit c5caf1e26baf2c90d514f9a331626c43d099e555
+Author: Jerome Glisse <glisse@freedesktop.org>
+Date: Thu Jun 14 13:16:11 2007 +0200
+
+ avivo: randr 1.2 output (not compiled yet, work in progress)
+
+commit 39ec89b5e1569b340e9b4310d646541116fddc84
+Author: Jerome Glisse <glisse@freedesktop.org>
+Date: Thu Jun 14 11:44:29 2007 +0200
+
+ avivo: more randr 1.2 work on crtc (not activated yet)
+
+commit 23730b5c0988627f0a0bb6f948a35dfb3ec1b47c
+Author: Jerome Glisse <glisse@freedesktop.org>
+Date: Wed Jun 13 23:57:25 2007 +0200
+
+ avivo: randr 1.2 support undergoing work (not compiled in yet)
+
+commit 4ce363f2a81fecf78c22cc9fa412aa877728adf8
+Author: Jerome Glisse <glisse@freedesktop.org>
+Date: Wed Jun 13 21:35:08 2007 +0200
+
+ avivo: (cleanup) kill blank line.
+
+commit 5333773aff456ef8e07d8f5fdc480f97cc051cef
+Author: Jerome Glisse <glisse@freedesktop.org>
+Date: Wed Jun 13 21:33:51 2007 +0200
+
+ avivo: move common avivo function in their own file (cleanup for randr 1.2)
+
+commit 35015f19755266ee942ec2cbd6d13d7014badc19
+Author: Jerome Glisse <glisse@freedesktop.org>
+Date: Wed Jun 13 21:27:12 2007 +0200
+
+ avivo: Move state saving in their own file (cleanup for randr 1.2)
+
+commit b9d183813ca811d9a9cbf303331f60227843116d
+Author: Jerome Glisse <glisse@freedesktop.org>
+Date: Wed Jun 13 21:18:38 2007 +0200
+
+ avivo: move cursor functions in their own file (cleanup for randr 1.2).
+
+commit 1b48bb7378d0b86894faf1973558029b2cff5a3e
+Author: Oliver McFadden <z3ro.geek@gmail.com>
+Date: Wed Jun 13 04:37:41 2007 +0000
+
+ avivo: Added 71c2 PCI ID; for Sapphire X1600 Pro HDMI.
+
+commit 36537c19a0f6fc74e79fdbf2a1bbc2185f6a0db3
+Author: Oliver McFadden <z3ro.geek@gmail.com>
+Date: Wed Jun 13 04:37:02 2007 +0000
+
+ avivotool: asm/page.h is deprecated.
+
+commit d277c9f4c0e7fb325f85a17e07bb503314958d65
+Author: Jerome Glisse <glisse@freedesktop.org>
+Date: Tue Jun 12 19:39:08 2007 +0200
+
+ avivo: Include stdint needed in avivo_i2c
+
+commit 1e9fc429a8d81f1b040611bc9d5add9b53e94310
+Author: Jerome Glisse <glisse@freedesktop.org>
+Date: Sun Jun 10 01:24:09 2007 +0200
+
+ avivo: Cleanup code a bit and use i2c bitbanging to select default output.
+
+ Do a ddc on each connector found in bios table and select
+ one with a monitor attached to it, if there is an lvds
+ output than it's the default one.
+
+commit 255fad8d3e1d232404942e96bffbc16cff26add8
+Author: Jerome Glisse <glisse@freedesktop.org>
+Date: Thu May 10 23:57:40 2007 +0200
+
+ avivo: Minor cleanup avoiding to print useless information.
+
+commit 245dc3b76b5bc3b1213a62fa5b208d783aaccd58
+Author: Jerome Glisse <glisse@freedesktop.org>
+Date: Thu May 10 23:57:02 2007 +0200
+
+ avivo: Fix gpu vram mapping.
+
+ Off course i did obviously setup memory mapping in
+ the wrong place, shame on me. Now this should work
+ smoothly.
+
+commit 2a773916624aba3e8be07123a1f4d8450f731577
+Author: Jerome Glisse <glisse@freedesktop.org>
+Date: Thu May 10 23:48:42 2007 +0200
+
+ avivo: Add gpu memory mapping setup.
+
+ You can setup where gpu sees vram, i basicly set it
+ to pci address (like a post from the bios should do).
+ This however doesn't seems to work yet properly after
+ newer fglrx run, even thought all involved regs are
+ know as i can fix the situation by setting register
+ myself.
+
+commit 04b6bd487efa8baff3fe7242f6621dce8c29a42a
+Author: Jerome Glisse <glisse@freedesktop.org>
+Date: Wed May 9 22:25:01 2007 +0200
+
+ avivo: Fix comment about pll i did give wrong computation instruction.
+
+commit 299844522b118bc80f5efd2f23b6e4bcfe1d5e73
+Author: Jerome Glisse <glisse@freedesktop.org>
+Date: Thu May 3 23:29:59 2007 +0200
+
+ avivo: Improve i2c handling.
+
+ We stop waiting after 3 NACK because this likely mean
+ that there is no one to speak with on this connector
+ or bus :).
+
+ Also added reset flags to i2c control register. So we
+ properly close and shutdown i2c bus.
+
+commit a6a3eea833fa3c6b9b344df3b1ddda44e87f36bb
+Author: Jerome Glisse <glisse@freedesktop.org>
+Date: Thu May 3 21:16:40 2007 +0200
+
+ avivo: Add some regs mostly mirror crtc1 regs for crtc2.
+
+ Small cleanup and add same regs for crtc2 as for crtc1.
+
+commit 4840ddc3f6369f338b1cc85c9e8315cc04fb8f3c
+Author: Jerome Glisse <glisse@freedesktop.org>
+Date: Thu May 3 20:46:32 2007 +0200
+
+ avivo: Avoid restoring crtc sync width register.
+
+ crtc sync width register shouldn't be zero when going
+ back to text mode otherwise we i don't get any things
+ on screen here with rv515.
+
+commit 8f4d756a3bb3d70183ac1978e14ef4723963e630
+Author: Jerome Glisse <glisse@freedesktop.org>
+Date: Thu May 3 20:45:13 2007 +0200
+
+ avivo: registers cleanup.
+
+commit 1899f2556d99d6b81e1443f87318ad64b3ac85fc
+Author: Jerome Glisse <glisse@freedesktop.org>
+Date: Thu May 3 19:35:04 2007 +0200
+
+ avivo: remove PLL mystery regs as they are useless for cold card init.
+
+ This 2 mysterys regs doesn't seems to be important on rv515
+ at least.
+
+commit 01c4ce02936f4a81d450f80fe084af49d629f2db
+Author: Jerome Glisse <glisse@freedesktop.org>
+Date: Thu May 3 19:21:52 2007 +0200
+
+ avivo: Forgot the divider cntl in previous commit.
+
+commit e421892c764ad2737c7b27513351006d95a40539
+Author: Jerome Glisse <glisse@freedesktop.org>
+Date: Thu May 3 17:53:19 2007 +0200
+
+ avivo: Add pll register which need to be initialized on rv515.
+
+ The two pll register need to be set on rv515 as they
+ aren't initialized by the bios. The PLL_CNTL reg is a
+ guess.
+
+commit c37016f008d68cb21a849fc198e2a7cb6f7c6772
+Author: Jerome Glisse <glisse@freedesktop.org>
+Date: Thu May 3 11:17:41 2007 +0200
+
+ avivo: Rework output/connector association.
+
+ So now we have a list of each available physical connector
+ reported by the bios (which might be innaccurate on some bios)
+ and with each connector a list of output that can feed this
+ connector.
+
+commit 1e22c44a60931ea44c98cc4f97c973d2bd3f8565
+Author: Jerome Glisse <glisse@freedesktop.org>
+Date: Thu May 3 11:15:17 2007 +0200
+
+ avivotool: Add i2c bit banging support using xf86 code.
+
+commit 5517670c4bb093ec89dea7ec3ec2e63c1d461dcc
+Author: Jerome Glisse <glisse@freedesktop.org>
+Date: Thu May 3 11:13:51 2007 +0200
+
+ avivo: Add new registers definitions and do some cleanup.
+
+commit b9c30302a137e8a634aa6a2eb868865ac43e4c0a
+Author: Jerome Glisse <glisse@freedesktop.org>
+Date: Mon Apr 30 18:09:33 2007 +0200
+
+ avivotool: Improve i2c handling.
+
+ I think i was a bit wrong on i2c start and acknowledge.
+
+commit 2b4ba60608d771056caa36ed0aed3fc9eba6337c
+Author: Jerome Glisse <glisse@freedesktop.org>
+Date: Wed Apr 18 18:46:06 2007 +0200
+
+ avivo: Change output structure comment and remove dead code.
+
+ Change comment on output driver and output structure to
+ somethings more accurate :) and remove dead code.
+
+commit 29476d9ed094317c03dd2fe12e656d9283eee677
+Author: Jerome Glisse <glisse@freedesktop.org>
+Date: Wed Apr 18 18:25:19 2007 +0200
+
+ avivo: Rework of output structure.
+
+ Initial rework of output structure, we can have several
+ output driver (DAC, TMDS, ...) for each physical output
+ (VGA, DVI, ...).
+
+commit 6d9d9bb7786483d1f14e39036c81bdcc621f2fb4
+Author: Jerome Glisse <glisse@freedesktop.org>
+Date: Wed Apr 18 16:00:14 2007 +0200
+
+ avivo: Rename OUTPUT to CONNECTOR, move crtc & output initialization
+
+ When using i2c you have to select over which connector you want
+ to communicate, rename reg to mirror that.
+
+ Move crtc & output initialization out of avivo_get_info, prior work
+ before redoing the crtc,output,connector code.
+
+commit ad99626b2056de1e99490dc9cb2ce919befb4e78
+Author: Daniel Stone <daniel@fooishbar.org>
+Date: Tue Apr 17 02:16:06 2007 +0300
+
+ Xorg: Add less bad mode validation
+
+ Pull in modes from DDC, and validate them using xf86ValidateModes. Obviously
+ not ideal, but it's a start, and means we don't need any hardcoded mode
+ lists.
+
+commit 379b2738fb58dd8250633387cbc7c0745a5364f4
+Author: Jerome Glisse <j.glisse@gmail.com>
+Date: Tue Apr 17 01:49:40 2007 +0300
+
+ New PLL calculation formula
+
+ Use a new formula to calculate the values of our PLL registers, and give
+ them more sensible names.
+
+commit 2004f60d5e5a8341a69514648c1ca25c8f92f94e
+Author: Daniel Stone <daniel@fooishbar.org>
+Date: Tue Apr 17 01:21:55 2007 +0300
+
+ Xorg: Revert accidental hunk
+
+ The offset wasn't supposed to be there (artefact of how new fglrx sets up
+ the memory map).
+
+commit 33981c0f3518665afcb000380810b1217ee9ac8c
+Author: Daniel Stone <daniel@fooishbar.org>
+Date: Tue Apr 17 01:19:27 2007 +0300
+
+ Xorg: Generalise CRTC registers, cleanup
+
+ Set all CRTC registers on all chipsets, not just RV515.
+ Coding style cleanup.
+
+commit 1c5477a53e7bec0b7f26fc93e64e5cde7e6e15de
+Author: Jerome Glisse <j.glisse@gmail.com>
+Date: Tue Apr 17 01:05:35 2007 +0300
+
+ Add support for RV515
+
+ Add proper support for the RV515, with some more mysterious CRTC registers.
+ Make authors relatively harmonious.
+
+commit a9c7c8e5269fc9211dadbe9d69222a2676a60740
+Author: Jerome Glisse <j.glisse@gmail.com>
+Date: Tue Apr 17 01:02:11 2007 +0300
+
+ Delete generated file
+
+commit cc846ff0734492c563cfeb3454386a6ec22d35b5
+Author: Jerome Glisse <j.glisse@gmail.com>
+Date: Tue Apr 17 00:55:58 2007 +0300
+
+ I2C: Fix implementation (working DDC over I2C)
+
+ Numerous fixes and changes to the I2C implementation so that it actually
+ works correctly. The reading and writing is still a little bit sketchy,
+ as we can't push more than 12 bytes at a time, and might need to flag
+ an ACK or so. Also, abort situations might need a better exit.
+
+commit 873470e5555a2cdb9fd45ae1622ff7385f14e45e
+Author: Daniel Stone <daniel@fooishbar.org>
+Date: Thu Apr 12 21:05:12 2007 +0300
+
+ Xorg: Initial (broken) I2C implementation
+
+ At the moment, it doesn't actually work. Looks like a great deal more work
+ needs to be done in between read/write (or just after either), rather than
+ before the initial operation, judging by the BIOS dumps.
+
+commit 41b661573ccd00f3dd4f476435a6916d4447359d
+Author: Daniel Stone <daniel@fooishbar.org>
+Date: Fri Mar 23 17:54:59 2007 +0200
+
+ Xorg: Protect against lockups
+ Don't loop forever waiting for vblank.
+
+commit 1fc3e5ccbd8ca719e55830d83372127271bf69ab
+Author: Daniel Stone <daniel@fooishbar.org>
+Date: Fri Mar 23 01:14:52 2007 +0200
+
+ Registers: Clarify comment
+ To write something, you've got to write the slave address to _DATA.
+
+commit d3a55007dfdc2b07e8903b9424c7617d9e4fbf4f
+Author: Daniel Stone <daniel@fooishbar.org>
+Date: Thu Mar 22 23:52:05 2007 +0200
+
+ Registers: Add I2C
+ I haven't got a working I2C implementation yet, but this will at least
+ document the registers a bit, so people can see what's going on.
+
+commit 54adc7ffc5ea59d731c621ef4bb0518946eec596
+Author: Daniel Stone <daniel@fooishbar.org>
+Date: Thu Mar 22 23:48:07 2007 +0200
+
+ Xorg: Minor location 'fixes'
+ For some reason, with the new fglrx, I have to change my base address
+ from 0xb0000000 to 0xc0000000. Who knows why. Sigh.
+ Make the probing for our various areas a little more robust while I'm
+ at it.
+
+commit 327b8b8762ee32fda43062f8d5d7917b273e67f3
+Author: Daniel Stone <daniel@fooishbar.org>
+Date: Sat Mar 10 13:20:44 2007 +0200
+
+ Add .gitignore
+
+commit a90a303b9bd2e81f5bfd2439cef5c9e92a8f1cbf
+Author: Daniel Stone <daniel@fooishbar.org>
+Date: Sat Mar 10 13:11:57 2007 +0200
+
+ Update STATUS
+
+commit a43ab986217df238085898ab596eccf432a987e1
+Author: Daniel Stone <daniel@fooishbar.org>
+Date: Sat Mar 10 03:32:45 2007 +0200
+
+ More magical mystery registers
+ Now we get a picture on the screen on a fresh boot, even if the clock is
+ hopelessly garbled.
+
+commit c480ac2c8ac4657363b3d69bcb9f7cc75953234f
+Author: Daniel Stone <daniel@fooishbar.org>
+Date: Fri Mar 9 19:18:26 2007 +0200
+
+ Add PLL post-divider registers
+ Add post-divider registers. Thanks to Adam Jackson for a PLL tutorial.
+
+commit 626af6f98cf3741062bff400f7b44e1360949e15
+Author: Daniel Stone <daniel@fooishbar.org>
+Date: Tue Mar 6 23:55:58 2007 +0200
+
+ Add PLL divider
+ 0x43c is definitely a PLL divider, with a constant value. Still not
+ entirely sure as to what 0x404 and 0x430 are ...
+
+commit 1a268a08818223ed3758f22a1160a8a538a02e61
+Author: Daniel Stone <daniel@fooishbar.org>
+Date: Tue Mar 6 23:55:29 2007 +0200
+
+ avivotool: Fix paths in check-reglist.sh
+
+commit 47b3c15bcf240bd6e7f31505ad362987a261aeb6
+Author: Daniel Stone <daniel@fooishbar.org>
+Date: Tue Mar 6 23:18:34 2007 +0200
+
+ Add CRTC1 offset, implement AdjustFrame
+
+commit fcee8dbabf3c1e3b903f3c030a87210d840cda27
+Author: Matthew Garrett <mjg59@srcf.ucam.org>
+Date: Tue Mar 6 03:31:33 2007 +0200
+
+ xorg: Fix typo in avivo_crtc_enable
+
+commit 24e7dbc49f14bc894916ae1c0b2c953b1f2ded38
+Author: Daniel Stone <daniel@fooishbar.org>
+Date: Tue Mar 6 03:15:45 2007 +0200
+
+ xorg: Fix mode restoring on exit
+ It turns out we have to wrap CloseScreen to be able to restore text mode.
+
+ Why?
+
+ No, seriously, why?
+
+commit 1fa0e0b3cc75a9a71b83102d3a348fab7f5ce0cb
+Author: Daniel Stone <daniel@fooishbar.org>
+Date: Tue Mar 6 02:49:21 2007 +0200
+
+ Update STATUS for text mode
+ VT switching works. Exit is still broken for some reason.
+
+commit 67e71055074db378af86afdbe0a4458038195592
+Author: Daniel Stone <daniel@fooishbar.org>
+Date: Tue Mar 6 00:04:13 2007 +0200
+
+ cosmetic: Clarify comments, update STATUS re: LVDS
+ LVDS works on the MBP now.
+
+commit 6e85f066535c1661fdcec9e36d3203d55d5f2cc0
+Author: Matthew Garrett <mjg59@srcf.ucam.org>
+Date: Mon Mar 5 23:59:28 2007 +0200
+
+ xorg: Preserve another TMDS2 mystery bit
+ 1 << 29 breaks LVDS but is necessary for TMDS; only carry it over if it was
+ already set.
+
+commit 25c1ce9d0f41e9fcc8e423e4c59f2a11ea3549a2
+Author: Daniel Stone <daniel@fooishbar.org>
+Date: Mon Mar 5 23:29:35 2007 +0200
+
+ xorg: Add RV530, add support for multiple chipsets
+ Add the RV530 from the MacBook Pro, and detect which chipset we're running
+ in PreInit.
+
+commit 66419cbadebea4607bd90c141efa0d0228d89730
+Author: Matthew Garrett <mjg59@srcf.ucam.org>
+Date: Mon Mar 5 23:16:19 2007 +0200
+
+ Tweak magical mystery TMDS2 registers
+ Updates from Matthew Garrett that _almost_ get LVDS on MacBook Pros
+ going. Setting (1 << 24) in TMDS2_CLOCK_CNTL breaks my actual TMDS,
+ though.
+
+commit a3f9afe919d83a771a80eefe2989b1f59d8f5ba4
+Author: Daniel Stone <daniel@fooishbar.org>
+Date: Mon Mar 5 22:26:49 2007 +0200
+
+ avivotool: fix path in script
+
+commit fd9dbd6f91e83a2c695472d5926e966232da4993
+Author: Matthew Garrett <mjg59@srcf.ucam.org>
+Date: Sun Mar 4 20:26:51 2007 +0200
+
+ xorg: Add text mode save/restore support
+ Add support for saving and restoring text mode.
+
+commit fb17e4cfc4cd42abac8e558649f0a2bfac17e13a
+Author: Daniel Stone <daniel@fooishbar.org>
+Date: Sun Mar 4 20:16:45 2007 +0200
+
+ initial import
+ Initial import of the Avivo driver; see STATUS for more information.
diff --git a/debian/changelog b/debian/changelog
index 53c9f5a..7615da7 100644
--- a/debian/changelog
+++ b/debian/changelog
@@ -1,3 +1,10 @@
+xserver-xorg-video-avivo (0.0.999+git20070927-1) experimental; urgency=low
+
+ * New upstream snapshot.
+ * Build against Xserver 1.4, closes: #444263.
+
+ -- Brice Goglin <bgoglin@debian.org> Thu, 27 Sep 2007 20:37:23 +0200
+
xserver-xorg-video-avivo (0.0.999+git20070805-1) experimental; urgency=low
* New upstream snapshot (closes: #435222).
diff --git a/debian/control b/debian/control
index d022059..5154929 100644
--- a/debian/control
+++ b/debian/control
@@ -3,7 +3,7 @@ Section: x11
Priority: extra
Maintainer: Debian X Strike Force <debian-x@lists.debian.org>
Uploaders: Julien Cristau <jcristau@debian.org>, Brice Goglin <bgoglin@debian.org>
-Build-Depends: debhelper (>= 5.0.0), pkg-config, xserver-xorg-dev (>= 2:1.2.99.902), x11proto-video-dev, x11proto-core-dev, x11proto-fonts-dev, x11proto-randr-dev, x11proto-render-dev, x11proto-xext-dev, libpciaccess-dev, autotools-dev, quilt
+Build-Depends: debhelper (>= 5.0.0), pkg-config, xserver-xorg-dev (>= 2:1.4), x11proto-video-dev, x11proto-core-dev, x11proto-fonts-dev, x11proto-randr-dev, x11proto-render-dev, x11proto-xext-dev, libpciaccess-dev, autotools-dev, quilt
Standards-Version: 3.7.2
Package: xserver-xorg-video-avivo
commit be7d380aba8d1ff2272bf2c931b2c9f78a549b46
Author: Matthew Garrett <mjg59@srcf.ucam.org>
Date: Sat Sep 15 00:46:47 2007 +0100
Minor include file cleanup
diff --git a/include/radeon_reg.h b/include/radeon_reg.h
index 5363fd9..c34fb92 100644
--- a/include/radeon_reg.h
+++ b/include/radeon_reg.h
@@ -3408,7 +3408,7 @@
#define AVIVO_TMDSA_DCBALANCER_CONTROL 0x78d0
# define AVIVO_TMDSA_DCBALANCER_CONTROL_EN (1 << 0)
# define AVIVO_TMDSA_DCBALANCER_CONTROL_TEST_EN (1 << 8)
-# define AVIVO_TMDSA_DCBALANCER_CONTROL_TEST_IN (16)
+# define AVIVO_TMDSA_DCBALANCER_CONTROL_TEST_IN_SHIFT (16)
# define AVIVO_TMDSA_DCBALANCER_CONTROL_FORCE (1 << 24)
#define AVIVO_TMDSA_DATA_SYNCHRONIZATION 0x78d8
# define AVIVO_TMDSA_DATA_SYNCHRONIZATION_DSYNSEL (1 << 0)
@@ -3437,7 +3437,7 @@
#define AVIVO_TMDSA_TRANSMITTER_CONTROL 0x7910
# define AVIVO_TMDSA_TRANSMITTER_CONTROL_PLL_ENABLE (1 << 0)
# define AVIVO_TMDSA_TRANSMITTER_CONTROL_PLL_RESET (1 << 1)
-# define AVIVO_TMDSA_TRANSMITTER_CONTROL_PLL_HPD_MASK (2)
+# define AVIVO_TMDSA_TRANSMITTER_CONTROL_PLL_HPD_MASK_SHIFT (2)
# define AVIVO_TMDSA_TRANSMITTER_CONTROL_IDSCKSEL (1 << 4)
# define AVIVO_TMDSA_TRANSMITTER_CONTROL_BGSLEEP (1 << 5)
# define AVIVO_TMDSA_TRANSMITTER_CONTROL_PLL_PWRUP_SEQ_EN (1 << 6)
@@ -3445,13 +3445,12 @@
# define AVIVO_TMDSA_TRANSMITTER_CONTROL_TMCLK_FROM_PADS (1 << 13)
# define AVIVO_TMDSA_TRANSMITTER_CONTROL_TDCLK (1 << 14)
# define AVIVO_TMDSA_TRANSMITTER_CONTROL_TDCLK_FROM_PADS (1 << 15)
-# define AVIVO_TMDSA_TRANSMITTER_CONTROL_CLK_PATTERN (16)
+# define AVIVO_TMDSA_TRANSMITTER_CONTROL_CLK_PATTERN_SHIFT (16)
# define AVIVO_TMDSA_TRANSMITTER_CONTROL_BYPASS_PLL (1 << 28)
# define AVIVO_TMDSA_TRANSMITTER_CONTROL_USE_CLK_DATA (1 << 29)
# define AVIVO_TMDSA_TRANSMITTER_CONTROL_INPUT_TEST_CLK_SEL (1 << 31)
#define AVIVO_LVTMA_CNTL 0x7a80
-#define AVIVO_TMDSA_CNTL 0x7880
# define AVIVO_LVTMA_CNTL_ENABLE (1 << 0)
# define AVIVO_LVTMA_CNTL_HPD_MASK (1 << 4)
# define AVIVO_LVTMA_CNTL_HPD_SELECT (1 << 8)
@@ -3473,7 +3472,7 @@
#define AVIVO_LVTMA_DCBALANCER_CONTROL 0x7ad0
# define AVIVO_LVTMA_DCBALANCER_CONTROL_EN (1 << 0)
# define AVIVO_LVTMA_DCBALANCER_CONTROL_TEST_EN (1 << 8)
-# define AVIVO_LVTMA_DCBALANCER_CONTROL_TEST_IN (16)
+# define AVIVO_LVTMA_DCBALANCER_CONTROL_TEST_IN_SHIFT (16)
# define AVIVO_LVTMA_DCBALANCER_CONTROL_FORCE (1 << 24)
#define AVIVO_LVTMA_DATA_SYNCHRONIZATION 0x78d8
@@ -3495,20 +3494,20 @@
# define AVIVO_LVTMA_TRANSMITTER_ENABLE_LNKDEN_HPD_MASK (1 << 18)
#define AVIVO_LVTMA_TRANSMITTER_CONTROL 0x7b10
-# define AVIVO_LVTMA_TRANSMITTER_CONTROL_PLL_ENABLE (1 << 0)
-# define AVIVO_LVTMA_TRANSMITTER_CONTROL_PLL_RESET (1 << 1)
Reply to: