Re: juju do not create fw* with vt6306l
Brought to you by:
aeb,
bencollins
Menu
▾
▴
Re: juju do not create fw* with vt6306l
|
From: Stefan R. <st...@s5...> - 2017-03-10 20:41:01
|
On Mar 10 lagunax wrote: > root@producer-center:/home/producer# lspci > 00:00.0 Host bridge: Intel Corporation Xeon E3-1200 v2/3rd Gen Core > processor DRAM Controller (rev 09) > ... > 01:00.0 PCI bridge: ASMedia Technology Inc. ASM1083/1085 PCIe to PCI > Bridge (rev 01) > 02:00.0 FireWire (IEEE 1394): VIA Technologies, Inc. VT6306/7/8 [Fire > II(M)] IEEE 1394 OHCI Controller (rev 80) [...] > 02:00.0 FireWire (IEEE 1394) [0c00]: VIA Technologies, Inc. VT6306/7/8 > [Fire II(M)] IEEE 1394 OHCI Controller [1106:3044] (rev 80) (prog-if 10 > [OHCI]) [...] > root@producer-center:/home/producer# dmesg | grep fire > [ 0.535278] firewire_ohci 0000:02:00.0: enabling device (0080 -> 0083) > [ 0.598118] firewire_ohci 0000:02:00.0: added OHCI v1.10 device as > card 0, 4 IR + 8 IT contexts, quirks 0x11 > root@producer-center:/home/producer# dmesg | grep 1394 > root@producer-center:/home/producer# As far as I know, the PCI identifiers [1106:3044] (rev 80) belong to VIA VT6308. I have VT6306 and VT6315, and had a VT6307 for a while, but no VT6308; but I am quite sure that the VT6308 is very similar internally to VT6307, and that other Linux users have working VT6308s. Your dmesg starts like it should, but then at least one important line does not follow. Here is a log for a VIA VT6306 CardBus Card right after insertion, though with a little extra verbosity added: # cat /etc/modprobe.d/firewire.conf options firewire-ohci debug=2 # dmesg | tail [4831778.128422] pcmcia_socket pcmcia_socket0: pccard: CardBus card inserted into slot 0 [4831778.128451] pci 0000:13:00.0: [1106:3044] type 00 class 0x0c0010 [4831778.128487] pci 0000:13:00.0: reg 0x10: [mem 0x00000000-0x000007ff] [4831778.128504] pci 0000:13:00.0: reg 0x14: [io 0x0000-0x007f] [4831778.128521] pci 0000:13:00.0: reg 0x18: [mem 0x00000000-0x000000ff] [4831778.128664] pci 0000:13:00.0: supports D2 [4831778.128665] pci 0000:13:00.0: PME# supported from D2 D3hot D3cold [4831778.128743] pci 0000:13:00.0: BAR 0: assigned [mem 0xf7080000-0xf70807ff] [4831778.128751] pci 0000:13:00.0: BAR 2: assigned [mem 0xf7080800-0xf70808ff] [4831778.128760] pci 0000:13:00.0: BAR 1: assigned [io 0xa000-0xa07f] [4831778.128836] firewire_ohci 0000:13:00.0: enabling device (0080 -> 0083) [4831778.181476] firewire_ohci 0000:13:00.0: added OHCI v1.0 device as card 9, 4 IR + 8 IT contexts, quirks 0x41 [4831778.181539] firewire_ohci 0000:13:00.0: 1 selfIDs, generation 1, local node ID ffc0 [4831778.181542] firewire_ohci 0000:13:00.0: selfID 0: 807f8956, phy 0 [---] S400 gc=63 +15W Lci [4831778.687512] firewire_core 0000:13:00.0: created device fw11: GUID 00110600000041cc, S400 If you don't have a CardBus card but an onboard conroller or PCI card instead, then you won't get the pcmcia related lines of course. The two "selfID" lines near the end are due to the debug=2 module parameter which I added. Could you try this too? First add the config file like I did (or enter the firewire-ohci line in an already existing modprobe.d file), reboot, check that the line took effect: # cat /sys/module/firewire_ohci/parameters/debug 2 then inspect the kernel log again (dmesg | grep fire). To continue: The last line in my log above, the one with "created device", is the one which I am missing from your log. The reason for its missing can only be that the selfID DMA or/and selfID interrupt did not work. And in turn, reasons for this could be: - defective controller, - bug in the ASMedia PCIe-to-PCI bridge, - whatever else. Actually I strongly suspect a bug of the bridge. Nowadays, by far most commercially available mainboards that still feature PCI contain buggy PCIe-to-PCI bridges that aren't worth the cents that they must have cost the mainboard vendor. The following bridges are good: PCI bridge [0604]: PLX Technology, Inc. PEX8112 x1 Lane PCI Express-to-PCI Bridge [10b5:8112] (rev aa) (this is on an add-in card of mine) PCI bridge [0604]: Tundra Semiconductor Corp. Device [10e3:8113] (rev 01) (this is on a Supermicro mainboard) Do you see any messages in your kernel log that could be related to the ASMedia bridge? Anyway. FYI, here are log lines (still with debug=2) from the moment when I plug in/ switch on a Focusrite Saffire audio interface: [4832701.959552] firewire_ohci 0000:13:00.0: isochronous cycle inconsistent [4832701.959576] firewire_ohci 0000:13:00.0: 2 selfIDs, generation 2, local node ID ffc0 [4832701.959579] firewire_ohci 0000:13:00.0: selfID 0: 807f8964, phy 0 [-p-] S400 gc=63 +15W Lc [4832701.959581] firewire_ohci 0000:13:00.0: selfID 0: 817f8cd2, phy 1 [c-.] S400 gc=63 -3W Lci [4832704.017887] firewire_core 0000:13:00.0: BM lock failed (timeout), making local node (ffc0) root [4832704.017890] firewire_core 0000:13:00.0: phy config: new root=ffc0, gap_count=5 [4832704.017974] firewire_ohci 0000:13:00.0: 2 selfIDs, generation 3, local node ID ffc1 [4832704.017976] firewire_ohci 0000:13:00.0: selfID 0: 80458c90, phy 0 [p-.] S400 gc=5 -3W Lc [4832704.017978] firewire_ohci 0000:13:00.0: selfID 0: 81458976, phy 1 [-c-] S400 gc=5 +15W Lci [4832704.911382] firewire_ohci 0000:13:00.0: 2 selfIDs, generation 4, local node ID ffc1 [4832704.911385] firewire_ohci 0000:13:00.0: selfID 0: 807f8c92, phy 0 [p-.] S400 gc=63 -3W Lci [4832704.911387] firewire_ohci 0000:13:00.0: selfID 0: 817f8974, phy 1 [-c-] S400 gc=63 +15W Lc [4832704.911404] firewire_core 0000:13:00.0: phy config: new root=ffc1, gap_count=5 [4832704.911478] firewire_ohci 0000:13:00.0: 2 selfIDs, generation 5, local node ID ffc1 [4832704.911479] firewire_ohci 0000:13:00.0: selfID 0: 80458c90, phy 0 [p-.] S400 gc=5 -3W Lc [4832704.911481] firewire_ohci 0000:13:00.0: selfID 0: 81458976, phy 1 [-c-] S400 gc=5 +15W Lci [4832707.611096] firewire_core 0000:13:00.0: created device fw12: GUID 00130e040140dd28, S400 So there are more SelfID lines, and finally a "created device" line again. Also, if "grep fire /proc/interrupts" is inspected before and after these events, it should show that they increase firewire-ohci's interrupt count. -- Stefan Richter -======----= --== -=-=- http://arcgraph.de/sr/ |
