(no commit message)
[ikiwiki.git] / docs / user / asuseee900 / index.mdwn
a6e97730 1# DragonFly on Asus EEE 900
ad86dd68 2
3[[!toc levels=2]]
5Following experience is from running DragonFly 2.3.2 development on laptop.
7 $ sudo lspci -nn
9## Kernel and system in general
11SMP kernel boots with IO_APIC disabled only. With IO_APIC support it panics early in boot. I haven't bothered yet to dig deeper into this issue.
13I'm not able to boot machine with SATA controller in AHCI mode while running on battery (without AC). It boots fine with AHCI AC connected as it does SATA controller in IDE mode. As some brief benchmarks showed that there is no actual difference I'm using it SATA controller in IDE mode at the moment. The issue itself is related to interrupt routing probably, it really needs some love in DragonFly.
15##### Update from 20.05.2008
17The BIOS update released in 18.05.2008 - 7NETB3WW (2.13) - solves most of problems mentioned here, but the kernel with IO_APIC is still no go.
19* A kernel without IO_APIC option boots fine with SATA controller in any mode - AHCI or compatibility, with and without AC. So there is no need to switch SATA controller to compatibility mode any more.
20* A kernel with IO_APIC boots with SATA controller in AHCI mode, but doesn't find any disks.
21* A kernel with IO_APIC boots with SATA controller in compatibility mode, finds disks, but ath(4) doesn't work.
23## ACPI and power management
25There is no issues in running with ACPI.
27 acpi0: <LENOVO TP-7N> on motherboard
28 acpi_ec0: <Embedded Controller: GPE 0x12, ECDT> port 0x66,0x62 on acpi0
29 acpi0: Power Button (fixed)
30 acpi_timer0: <24-bit timer at 3.579545MHz> port 0x1008-0x100b on acpi0
31 cpu0: <ACPI CPU (3 Cx states)> on acpi0
32 cpu1: <ACPI CPU (3 Cx states)> on acpi0
33 acpi_tz0: <Thermal Zone> on acpi0
34 acpi_tz1: <Thermal Zone> on acpi0
35 acpi_lid0: <Control Method Lid Switch> on acpi0
36 acpi_button0: <Sleep Button> on acpi0
37 atkbdc0: <Keyboard controller (i8042)> port 0x64,0x60 irq 1 on acpi0
38 acpi_cpu: throttling enabled, 8 steps (100% to 12.5%), currently 100.0%
39 acpi_cmbat0: <Control Method Battery> on acpi0
40 acpi_acad0: <AC Adapter> on acpi0
42Suspend (S3) doesn't work though - panics on resume. Haven't time to dig into it yet. Throttling CPU speed works through ACPI, but it doesn't seem to have any impact to battery life, it makes system just slower. So I disabled automatic throttling while running on battery:
44 $ sysctl hw.acpi.cpu
45 hw.acpi.cpu.throttle_max: 8
46 hw.acpi.cpu.throttle_state: 8
47 hw.acpi.cpu.performance_speed: 8
48 hw.acpi.cpu.economy_speed: 8
50[est(4)](http://leaf.dragonflybsd.org/cgi/web-man?command#est&section4) in DragonFly doesn't support Core family CPU's yet AFAIK. I have 8 cell battery and it lasts about 5 hours, which is enough for me at the moment. Therefore I haven't bothered to dig into power management issues yet.
52For tuning/controlling/monitoring ACPI extras for Thinkpad laptops I have acpi_ibm(4) and acpi_video(4) ports from FreeBSD in work. Neither of them are rady to commit yet, but I can share the current state if you contact me privately. I'll work on them after 2.0 release probably.
54 $ sysctl hw.acpi.ibm
55 hw.acpi.ibm.initialmask: 2060
56 hw.acpi.ibm.availmask: 16777215
57 hw.acpi.ibm.events: 0
58 hw.acpi.ibm.eventmask: 2060
59 hw.acpi.ibm.hotkey: 2483
60 hw.acpi.ibm.lcd_brightness: 0
61 hw.acpi.ibm.volume: 0
62 hw.acpi.ibm.mute: 0
63 hw.acpi.ibm.thinklight: 0
64 hw.acpi.ibm.bluetooth: 1
65 hw.acpi.ibm.wlan: 1
66 hw.acpi.ibm.fan_speed: 4343
67 hw.acpi.ibm.fan_level: 0
68 hw.acpi.ibm.fan: 1
69 hw.acpi.ibm.thermal: 52 48 48 49 37 -1 35 -1
71## Video controller
73The chipset in use is Intel 965GM.
75 agp0: <Intel GM965 SVGA controller> port 0x1800-0x1807 mem 0xe0000000-0xefffffff,0xf8000000-0xf80fffff irq 10 at device 2.0 on pci0
76 agp0: detected 7676k stolen memory
77 agp0: aperture size is 256M
78 drmsub0: <Intel i965GM> on agp0
80Although there are some issues with it, Intel chips are the best ones to use at the moment. I experimented a lot with Xorg and found that XAA acceleration architecture with DRI works best for me. DRM isn't in CVS yet, so you have to use info from [this mail](http://leaf.dragonflybsd.org/mailarchive/users/2008-01/msg00003.html).
82+ **XAA without DRI** - Works in general, but there are random XVideo related crashes - [this bug](https://bugs.freedesktop.org/show_bug.cgi?id=13976). Crashes seem to be specific to 965GM chipset, 965Q I'm using in desktops doesn't have this problem.
83+ **EXA without DRI** - Unusably slow. Seems to be known problem (see [this bug](https://bugs.freedesktop.org/show_bug.cgi?id=13389) for example) Note, that EXA is default in new intel driver 2.2.0, so if you don't want to run DRI, switch to XAA.
84+ **XAA with DRI** - Best option so far. Running GL applications is problematic though - I'm able to crash X 100% with moving glxgears window (see [this bug](https://bugs.freedesktop.org/show_bug.cgi?id=13985) for details) and with trying to run second glxgears. The second issue is probably related to bugs in our DRM locking. XVideo crashes are rare, but they are there.
85+ **EXA with DRI** - The speed is good, I can't crash X with moving glxgears windows, but opening second glxgears still crashes X. There are also some random crashes I'm not yet sure about.
87##### Update (20080228)
89XVideo crases seems to be related Intel(R) Textured Video adaptor. Modified Xorg driver (2.2.1) which adds hardware overlay support for G965 chips works fine without crashes so far. See [this thread](http://thread.gmane.org/gmane.comp.freedesktop.xorg/26334) for details how to modify the driver.
91## Ethernet controller
93Works without any problem with em(4).
95 em0: <Intel(R) PRO/1000 Network Connection, Version - 6.2.9> port 0x1840-0x185f mem 0xf8225000-0xf8225fff,0xf8200000-0xf821ffff irq 11 at device 25.0 on pci0
97## Wireless card(s)
99AFAICS all X61(s) models have either Intel PRO/Wireless 3945ABG or Intel PRO/Wireless 4965AGN Mini-PCI Express WiFi adapters. Unfortunately neither of them is working in DragonFly at the moment. Situation isn't hopeless though, both of them are supported in OpenBSD via wpi(4) and iwn(4) drivers respectively. With some effort it would be possible to support them. There is also wpi(4) work in progress in FreeBSD which would be simpler to port than the driver from FreeBSD.
101I haven't done any porting work yet, because I can use Atheros AR5418 802.11a/b/g/n Wireless Mini-PCI Express adapter for now. I'm not very happy with it though, because I have to use ath(4) driver with binary blob. Blob in DragonFly doesn't work either and I have to use the one from [[http://people.freebsd.org/~sam/]]. This one works without problems so far and has necessary for me WPA support.
103 ath0: <Atheros 5418> mem 0xf7f00000-0xf7f0ffff irq 11 at device 0.0 on pci3
104 ath0: mac 12.10 phy 8.1 radio 12.0
106If you plan to change Mini-PCI Express card note that "unauthorised" ones don't work out of box. See [this link]( http://www.thinkwiki.org/wiki/Problem_with_unauthorized_MiniPCI_network_card) for more details.
108## Bluetooth
110Works with ubt(4) driver without problems. In fact all my bluetooth(4) work is done with this laptop.
112 ubt0: <Broadcom Corp BCM2045B, class 224/1, rev 2.00/1.00, addr 3> on uhub0
114## Audio
116Works without problems.
118 pcm0: <Intel 82801H High Definition Audio Controller> mem 0xf8220000-0xf8223fff irq 11 at device 27.0 on pci0
119 pcm0: <HDA Codec: Analog Devices AD1984>
120 pcm0: <HDA Driver Revision: 20071129_0050>
122## Firewire
124Seems to work, but I haven't done any tests so far.
126 fwohci0: vendor#1180, dev832
127 fwohci0: <1394 Open Host Controller Interface> mem 0xd7efe800-0xd7efefff irq 11 at device 0.1 on pci5
128 fwohci0: OHCI version 1.10 (ROM=0)
129 fwohci0: No. of Isochronous channel is 4.
130 fwohci0: EUI64 00:16:d3:ff:ff:c2:f7:57
131 fwohci0: Phy 1394a available S400, 1 ports.
132 fwohci0: Link S400, max_rec 2048 bytes.
133 firewire0: <IEEE1394(FireWire) bus> on fwohci0
134 fwe0: <Ethernet over FireWire> on firewire0
135 fwe0: MAC address: 02:16:d3:c2:f7:57
136 sbp0: <SBP-2/SCSI over FireWire> on firewire0
137 fwohci0: Initiate bus reset
138 fwohci0: node_id#0xc800ffc0, gen1, CYCLEMASTER mode
139 firewire0: 1 nodes, maxhop <# 0, cable IRM 0 (me)
140 firewire0: bus manager 0 (me)
142## SD card reader
144Isn't supported yet.