On Thursday 08 December 2005 01:08 am, Artemiev Igor wrote: > On Wed, 7 Dec 2005 08:16:44 -0500 > > John Baldwin <jhb_at_freebsd.org> wrote: > > On Tuesday 06 December 2005 01:30 am, Artemiev Igor wrote: > > > I add pseudo-device amdpmsub because nForce2 have two SMBus > > > interfaces. Patch here: > > > > > > http://bmc.brk.ru/~ai/patches/amdpm.nforce2_support.diff > > > > > > dmesg output: > > > > > > amdpm0: <nForce2 MCP-T SMBus Controller> port > > > 0xdc00-0xdc1f,0x5000-0x501f,0x5500 -0x551f irq 5 at device 1.1 on > > > pci0 smbus0: <System Management Bus> on amdpm0 > > > smb0: <SMBus generic I/O> on smbus0 > > > amdpmsub0: <nForce2 MCP-T Slave SMBus Controller> on amdpm0 > > > smbus1: <System Management Bus> on amdpmsub0 > > > smb1: <SMBus generic I/O> on smbus1 > > > > > > I`ve tested it with xmbmon and all works fine. > > > > Is there any reason you can't make smbus1 a childof amdpm0 directly > > and skip the amdpmsub0 device? > > I simply do not see any way to do it with current implementation of > ampdm(and also viapm, etc) & smbus, without modifying the smb->smbus->smbus > driver interface. I may be wrong, but as far as I know, currently it's one > smb for one driver (smbus_* limitation) Hmm, it doesn't specify the child device, just the parent. *sigh* That's lame. You don't have to call it amdpmsub0 btw, you could just call it amdpm1 if you wanted and have amdpm1 a child of amdpm0. All that would need to change for that is the NF2_SUBDEV string and the DRIVER_MODULE line (it would be DRIVER_MODULE(amdpm, amdpm, ...)). This has the added advantage that you don't have to patch smbus.c. -- John Baldwin <jhb_at_FreeBSD.org> <>< http://www.FreeBSD.org/~jhb/ "Power Users Use the Power to Serve" = http://www.FreeBSD.orgReceived on Thu Dec 08 2005 - 13:57:02 UTC
This archive was generated by hypermail 2.4.0 : Wed May 19 2021 - 11:38:48 UTC