Bugzilla – Bug 142831
Installation, due to ignoring matrox multihead possibily, doesn't support matrox multihead, at least in some cases.
Last modified: 2006-01-23 11:00:23 UTC
I've 3 monitors. Two 17" on Matrox AGP card (dual head) and one 15" on ATI PCI card (single head). To enable support for second monitor I had to recompile kernel & eneable some modules. W/o this the installation saw only two from free monitors. I've following hardware (taken from yast hardware detection info): 25: PCI 20a.0: 0300 VGA compatible controller (VGA) [Created at pci.277] UDI: /org/freedesktop/Hal/devices/pci_1002_5654 Unique ID: y9as.vMIeqvjWLh3 Parent ID: 6NW+.carL4jV6EX3 SysFS ID: /devices/pci0000:00/0000:00:1e.0/0000:02:0a.0 SysFS BusID: 0000:02:0a.0 Hardware Class: graphics card Model: "ATI Mach64 VT" Vendor: pci 0x1002 "ATI Technologies Inc" Device: pci 0x5654 "Mach64 VT" Revision: 0x48 Memory Range: 0xf1000000-0xf1ffffff (rw,non-prefetchable,disabled) I/O Ports: 0xd400-0xd4ff (rw,disabled) Memory Range: 0xf5ee0000-0xf5eeffff (ro,prefetchable,disabled) Module Alias: "pci:v00001002d00005654sv00000000sd00000000bc03sc00i00" Driver Info #0: XFree86 v4 Server Module: ati Config Status: cfg=yes, avail=yes, need=yes, active=unknown Attached to: #17 (PCI bridge) 23: PCI(AGP) 100.0: 0300 VGA compatible controller (VGA) [Created at pci.277] UDI: /org/freedesktop/Hal/devices/pci_102b_525 Unique ID: VCu0.mBMLZvg2uM8 Parent ID: vSkL.UFGjxxbnmrA SysFS ID: /devices/pci0000:00/0000:00:01.0/0000:01:00.0 SysFS BusID: 0000:01:00.0 Hardware Class: graphics card Model: "Matrox G450 LE" Vendor: pci 0x102b "Matrox Graphics, Inc." Device: pci 0x0525 "MGA G400 AGP" SubVendor: pci 0x102b "Matrox Graphics, Inc." SubDevice: pci 0x07c0 "G450 LE" Revision: 0x82 Driver: "matrox_w1" Memory Range: 0xf6000000-0xf7ffffff (rw,prefetchable) Memory Range: 0xf4000000-0xf4003fff (rw,non-prefetchable) Memory Range: 0xf3800000-0xf3ffffff (rw,non-prefetchable) Memory Range: 0xf5fe0000-0xf5ffffff (ro,prefetchable,disabled) IRQ: 11 (150157 events) I/O Ports: 0x3c0-0x3df (rw) Module Alias: "pci:v0000102Bd00000525sv0000102Bsd000007C0bc03sc00i00" Driver Info #0: XFree86 v4 Server Module: mga Driver Info #1: XFree86 v4 Server Module: mga 3D Support: yes Extensions: dri Config Status: cfg=yes, avail=yes, need=yes, active=unknown Attached to: #16 (PCI bridge) Unfortunately I don't remember what exactly I enabled in kernel compilation - just enabled on readaing help & all works. ;) Though I know, that SuSE 9.2 installation detects my 3 heads OK. (The only thing I disliked in SuSE multihead configurator was enabling DPMS only for monitors on my primary (AGP) card, so I had to review config by hands. %) ) Probably that beheviour should be documented in relase notes?
Not sure if this is X11 or yast, but whatever it is, it's not kernel :)
Well.. 've reread my post.. I'm speaking about second monitor on Matrox card, if not yet clear from the summary. %) In working self-recompiled kernel I've at least the following: skylab:/usr/src # grep -i matrox /usr/src/linux-2.6.13-15.7/.config CONFIG_W1_MATROX=m CONFIG_FB_MATROX=y # CONFIG_FB_MATROX_MILLENIUM is not set # CONFIG_FB_MATROX_MYSTIQUE is not set CONFIG_FB_MATROX_G=y CONFIG_FB_MATROX_I2C=m CONFIG_FB_MATROX_MAVEN=m CONFIG_FB_MATROX_MULTIHEAD=y skylab:/usr/src # The full kernel configuration for my system is _temporary_ avaliable as 'http://olli.digger.org.ru/files/pub/user.hints/linux-2.6.13-15.7-custom.config' & will be there till next web-server content upgrade. ( I donno how long %) ).
it _IS_ kernel since it require to enable the kernel module that is OFF by default. :/
Sorry, but we only support the generic framebuffer, because we've seen to much trouble with native hardware framebuffers like matrox.
OK. Though if the 'matrox dual head customisation needed' support notice 'll be documented in release notes I'm able to leave the proposed custom kernel configuration that will most probably solve the support of matrox multihead for a long time on my web. Should I?
Proposed kernel configuration changes: -# CONFIG_FB_MATROX is not set +CONFIG_FB_MATROX=m +CONFIG_FB_MATROX_MILLENIUM=y +CONFIG_FB_MATROX_MYSTIQUE=y +CONFIG_FB_MATROX_G=y +CONFIG_FB_MATROX_I2C=m +CONFIG_FB_MATROX_MAVEN=m +CONFIG_FB_MATROX_MULTIHEAD=y
Are there any reasons why we disabled build of the matrox fb module? Otherwise I suggest to do the changes mentioned in comment #10.
Well, we don't support the native framebuffer drivers in any way. Still we seem to ship most of them (with the only exception of the matrox?). So it would make sense to also ship the matrox module. Do we have to do anything to prevent them from being autodetected and used?
I agree. The native framebuffer drivers should't be autodetected and used by default, but in the past this was broken in every new release. :-(
Enabled matrox fb driver on i386. This change will be in beta2.