AMD Catalyst 9.5 Driver For Linux Released

[smitty3268]

KMS Support

The main reason I'd like KMS support is that it seems to help make suspend/resume more reliable - although maybe that wouldn't carry over to a proprietary driver anyway.

It's also required for Wayland support, although that's got a lot of work ahead of it before it becomes an issue. I hope that if Wayland does become ready for mainstream use that proprietary video drivers don't hold back adoption, though.

[mtippett]

Just wondering, do you have any plans about bringing some variant of KMS support to Catalyst?

There are no plans at this stage. Most people I have spoken to don't have an interest in seeing kernel messages appear on their screen. Last time I saw that in a production system was on old Sparc X Terminals.

Beyond that (which can already be achieved with a kernel based chvt + message to that console (aka the Linux BSOD)), it is about avoiding slow switches during boot and fast chvt's. We haven't seen any direct requirements for these. THe intel chvt takes a second or two, under fglrx takes about half a second.

The GEM/TTM/other parts of what people call KMS we have already had since the R6xx family was enabled.

Regards,

Matthew

[kfox1111]

There are no plans at this stage. Most people I have spoken to don't have an interest in seeing kernel messages appear on their screen. Last time I saw that in a production system was on old Sparc X Terminals.

I've been having instability on my system that I'm fairly sure is related to fglrx but haven't been able to prove it because:

A. Fglrx prevents the kernel from properly displaying an OOPS message
B. The box locks up so tight it has to be cold booted
C. I haven't had time to instrument up something to work around not being able to get an OOPS.

But, even if it isn't fglrx, I still can't be sure what the problem is because of A.

I think it is a very good idea to provide support for that because it lets people find out what the cause of problems. Being able to see the oops, either fglrx will get blamed less for crashing the box, or it helps you find out what is fglrx is crashing the box. Either way its a win.

Kevin

[nanonyme]

There are no plans at this stage. Most people I have spoken to don't have an interest in seeing kernel messages appear on their screen.

You'd think the most important thing KMS would give to production users is that you no longer have to run X as root with it. Can fglrx do this by its own means? It might be X servers will start being shipped without setuid root bit in Linux sooner or later (for security reasons) so it might be very important fglrx can acommodate.

[kfox1111]

You'd think the most important thing KMS would give to production users is that you no longer have to run X as root with it.

This is an aspect of the OOPS thing that I hadn't considered. It may prevent a bug in fglrx (or X) from OOPSing the box in the first place.

[gzunk]

You know, I've read all the posts in the above discussion, but none of it covered my experience - which is that 9.4 and 9.5 DO NOT WORK with Ubuntu 9.04, 64 bit, 4xxxHD X2 cards. Segfault on startup.

I realise that they're "early access" drivers, but come on, early access? The Windows 7 drivers are more functional (well, they couldn't be less functional, since mine DONT WORK)

Sorry to use capitals, but really, I spent $600 on a 4870X2 and while it works as expected in Windows (most of the time) it just really sucks under Ubuntu Jaunty at the moment.

I run the RadeonHD driver since at least it gives me a framebuffer, but that's all it gives me. No 2D acceleration. No 3D acceleration. Now you might think that that's not too bad, I've got a fast processor, I'll just to the video decoding with the CPU, and indeed that does work.

What doesn't work is XRandR for my second monitor. Which is rotated 90 degrees to fit better with my primary display. Except that rotation is disabled when you don't have 2D acceleration. So I have to disable my second monitor - I wouldn't want to risk my neck by rotating my head.

So, to summarise, I'm not happy. I spent a lot of money on a card, which in my preferred operating system runs just like an old style Cirrus Logic card. Except with a bigger framebuffer.

I think that the nvidia alternative is a space heater, and I really don't want to get one - but come on, not even 2D acceleration? Would you live with that?

[bridgman]

What problems are you encountering with 2D acceleration ? The open source "radeon" driver and drm that ship with Jaunty should give you 2D acceleration out of the box.

Can you pls pastebin xorg log ? Bonus points for xorg conf and dmesg output

[gzunk]

What problems are you encountering with 2D acceleration ? The open source "radeon" driver and drm that ship with Jaunty should give you 2D acceleration out of the box.

Can you pls pastebin xorg log ? Bonus points for xorg conf and dmesg output

I'll just post a snippet of the xorg log file for the fglrx output, since I don't know the etiquette for posting log files:

Intro:

X.Org X Server 1.6.0
Release Date: 2009-2-25
X Protocol Version 11, Revision 0
Build Operating System: Linux 2.6.24-15-server x86_64 Ubuntu
Current Operating System: Linux barney 2.6.28-11-generic #42-Ubuntu SMP Fri Apr 17 01:58:03 UTC 2009 x86_64
Build Date: 09 April 2009  02:11:54AM
xorg-server 2:1.6.0-0ubuntu14 ([email protected]) 
	Before reporting problems, check http://wiki.x.org
	to make sure that you have the latest version.
Markers: (--) probed, (**) from config file, (==) default setting,
	(++) from command line, (!!) notice, (II) informational,
	(WW) warning, (EE) error, (NI) not implemented, (??) unknown.
(==) Log file: "/var/log/Xorg.0.log", Time: Tue May 19 21:49:22 2009
(==) Using config file: "/etc/X11/xorg.conf"
(==) No Layout section.  Using the first Screen section.
(**) |-->Screen "Default Screen" (0)
(**) |   |-->Monitor "Samsung SM305T"
(**) |   |-->Device "FGLRX Video Device"
(**) Option "DontZap" "off"

Detection of card:

(II) LoadModule: "fglrx"
(II) Loading /usr/lib/xorg/modules/drivers//fglrx_drv.so
(II) Module fglrx: vendor="FireGL - ATI Technologies Inc."
	compiled for 1.4.99.906, module version = 8.61.2
	Module class: X.Org Video Driver
(II) Primary Device is: PCI 03@00:00:0
(WW) Falling back to old probe method for fglrx
(II) ATI Proprietary Linux Driver Version Identifier:8.61.2
(II) ATI Proprietary Linux Driver Release Identifier: 8.612                                
(II) ATI Proprietary Linux Driver Build Date: Apr 28 2009 21:45:06
(II) Loading PCS database from /etc/ati/amdpcsdb
(WW) This ATI Proprietary Linux Driver does not guarantee support of video driver ABI higher than 2.0
(WW) Video driver ABI version of the X server is 5.0
(--) Chipset Supported AMD Graphics Processor (0x9441) found
(WW) fglrx: No matching Device section for instance (BusID PCI:0@3:0:1) found
(**) ChipID override: 0x9441
(**) Chipset Supported AMD Graphics Processor (0x9441) found
(II) AMD Video driver is running on a device belonging to a group targeted for this release
(II) AMD Video driver is signed
(II) resource ranges after xf86ClaimFixedResources() call:
	[0] -1	0	0xffffffff - 0xffffffff (0x1) MX[B]
	[1] -1	0	0x000f0000 - 0x000fffff (0x10000) MX[B]
	[2] -1	0	0x000c0000 - 0x000effff (0x30000) MX[B]
	[3] -1	0	0x00000000 - 0x0009ffff (0xa0000) MX[B]
	[4] -1	0	0x0000ffff - 0x0000ffff (0x1) IX[B]
	[5] -1	0	0x00000000 - 0x00000000 (0x1) IX[B]
(II) fglrx(0): pEnt->device->identifier=0x1afc330
(II) pEnt->device->identifier=(nil)
(II) resource ranges after probing:
	[0] -1	0	0xffffffff - 0xffffffff (0x1) MX[B]
	[1] -1	0	0x000f0000 - 0x000fffff (0x10000) MX[B]
	[2] -1	0	0x000c0000 - 0x000effff (0x30000) MX[B]
	[3] -1	0	0x00000000 - 0x0009ffff (0xa0000) MX[B]
	[4] 0	0	0x000a0000 - 0x000affff (0x10000) MS[B]
	[5] 0	0	0x000b0000 - 0x000b7fff (0x8000) MS[B]
	[6] 0	0	0x000b8000 - 0x000bffff (0x8000) MS[B]
	[7] -1	0	0x0000ffff - 0x0000ffff (0x1) IX[B]
	[8] -1	0	0x00000000 - 0x00000000 (0x1) IX[B]
	[9] 0	0	0x000003b0 - 0x000003bb (0xc) IS[B]
	[10] 0	0	0x000003c0 - 0x000003df (0x20) IS[B]
(II) Setting vga for screen 0.

And error:

drmOpenDevice: node name is /dev/dri/card0
drmOpenDevice: open result is 9, (OK)
drmOpenByBusid: Searching for BusID PCI:3:0:0
drmOpenDevice: node name is /dev/dri/card0
drmOpenDevice: open result is 9, (OK)
drmOpenByBusid: drmOpenMinor returns 9
drmOpenByBusid: drmGetBusid reports PCI:3:0:0
(II) fglrx(0): AC Adapter is used
(II) Loading sub module "int10"
(II) LoadModule: "int10"
(II) Loading /usr/lib/xorg/modules//libint10.so
(II) Module int10: vendor="X.Org Foundation"
	compiled for 1.6.0, module version = 1.0.0
	ABI class: X.Org Video Driver, version 5.0
(II) fglrx(0): Primary V_BIOS segment is: 0xc000
(II) Loading sub module "vbe"
(II) LoadModule: "vbe"
(II) Loading /usr/lib/xorg/modules//libvbe.so
(II) Module vbe: vendor="X.Org Foundation"
	compiled for 1.6.0, module version = 1.1.0
	ABI class: X.Org Video Driver, version 5.0
(II) fglrx(0): VESA BIOS detected
(II) fglrx(0): VESA VBE Version 3.0
(II) fglrx(0): VESA VBE Total Mem: 16384 kB
(II) fglrx(0): VESA VBE OEM: ATI ATOMBIOS
(II) fglrx(0): VESA VBE OEM Software Rev: 11.7
(II) fglrx(0): VESA VBE OEM Vendor: (C) 1988-2005, ATI Technologies Inc. 
(II) fglrx(0): VESA VBE OEM Product: RV770
(II) fglrx(0): VESA VBE OEM Product Rev: 01.00
(II) fglrx(0): ATI Video BIOS revision 9 or later detected
(II) fglrx(0): PCI bus 4 card 0 func 0
(**) fglrx(0): Chipset: "ATI Radeon HD 4870 X2 " (Chipset = 0x9441)
(**) fglrx(0): (PciSubVendor = 0x1002, PciSubDevice = 0x2042)
(**) fglrx(0): board vendor info: original ATI graphics adapter
(EE) fglrx(0): No valid linear framebuffer address
(EE) fglrx(0): PreInit failed
(--) fglrx(0): Video RAM: 1048576 kByte, Type: GDDR5
(II) fglrx(0): PCIE card detected
(--) fglrx(0): Using per-process page tables (PPPT) as GART.
(II) fglrx(0): board/chipset is supported by this driver (original ATI board)
(II) fglrx(0): RandR 1.2 support is enabled!
(II) fglrx(0): RandR 1.2 rotation support is enabled!
(==) fglrx(0): Center Mode is disabled 
(II) Loading sub module "ddc"
(II) LoadModule: "ddc"
(II) Module "ddc" already built-in

Backtrace:
0: /usr/bin/X11/X(xorg_backtrace+0x26) [0x4f1b66]
1: /usr/bin/X11/X(xf86SigHandler+0x41) [0x485a61]
2: /lib/libc.so.6 [0x7f3f92ff4040]
3: /usr/lib/xorg/modules/drivers//fglrx_drv.so(CAIL_QueryAdditionInfo+0) [0x7f3f911ec7e0]
4: /usr/lib/xorg/modules/drivers//fglrx_drv.so(swlCailGetAsicExtRevID+0xe) [0x7f3f911c1f1e]
5: /usr/lib/xorg/modules/drivers//fglrx_drv.so(swlDalDisplayUpdateHwAsicIDForDal+0x40) [0x7f3f911dc100]
6: /usr/lib/xorg/modules/drivers//fglrx_drv.so(swlDalDisplayEnableDAL+0x184) [0x7f3f911dc384]
7: /usr/lib/xorg/modules/drivers//fglrx_drv.so [0x7f3f911d5202]
8: /usr/lib/xorg/modules/drivers//fglrx_drv.so [0x7f3f911d396f]
9: /usr/lib/xorg/modules/drivers//fglrx_drv.so(atiddxDisplayMapAddNode+0xbb) [0x7f3f911d3adb]
10: /usr/lib/xorg/modules/drivers//fglrx_drv.so(atiddxDisplayAdaptorCreate+0x9a) [0x7f3f911d459a]
11: /usr/lib/xorg/modules/drivers//fglrx_drv.so(atiddxDisplayPreInit+0x3b3) [0x7f3f911d3323]
12: /usr/lib/xorg/modules/drivers//fglrx_drv.so(atiddxPreInit+0x876) [0x7f3f91192d66]
13: /usr/bin/X11/X(InitOutput+0xdbc) [0x46f99c]
14: /usr/bin/X11/X(main+0x20e) [0x433bde]
15: /lib/libc.so.6(__libc_start_main+0xe6) [0x7f3f92fdf5a6]
16: /usr/bin/X11/X [0x433219]
Saw signal 11.  Server aborting.
 ddxSigGiveUp: Closing log

As for the RadonHD problems, that's not really appropriate for this thread (I was venting), but suffice to say that 2D acceleration is only enabled on RV770 in version 1.2.5, but on Jaunty the standard version is 1.2.4. I believe 1.2.5 requires kernel 2.6.30, and Jaunty is on 2.6.28.

I try and stick to the defaults as far as possible since that normally makes things more stable.

[Na-Fiann]

There are no plans at this stage. Most people I have spoken to don't have an interest in seeing kernel messages appear on their screen. Last time I saw that in a production system was on old Sparc X Terminals.

I'm sorry, but why is this an issue? Perhaps I'm mistaken, but shouldn't the kernel messages only appear onscreen when the driver locks-up? It seems to me that most people don't have an interest in that period.

On the other hand, kms allows for a graphical boot, less flickering during the boot process and vt switching with less chance of crashes. It will make suspend and resume functions more reliable as well. These are things that users expect from modern systems, sticking to an error-prone system that is less visually rich seems pointless.

If it's the error messages you're worried about, I'm sure it's possible to only enable them in testing systems and not production systems.

Please correct me if I'm wrong, but it seems to me the benefits far outweigh the disadvantages.

[RealNC]

KMS would probably require open-sourcing a part of fglrx as well as splitting that part of it and putting it into the kernel. I don't think this is ever gonna happen.