On Fri, 30 Sep 2011, Pitt, Jason N wrote: > Alan you rock. I've had all twelve cameras going for about half an > hour now without crashing... > > > Here's the dmesg: > [ 127.625903] ehci_hcd 0000:07:00.1: Bad frame counter change: 3559 3552 3560 > [ 148.959259] ehci_hcd 0000:07:00.3: Bad frame counter change: 3655 3648 3656 > [ 152.075198] ehci_hcd 0000:07:00.3: Bad frame counter change: 4039 4032 4040 > [ 155.758395] ehci_hcd 0000:07:00.3: Bad frame counter change: 775 768 776 > [ 195.588622] ehci_hcd 0000:09:00.1: Bad frame counter change: 183 176 184 > [ 212.207740] ehci_hcd 0000:09:00.5: Bad frame counter change: 2351 2344 2352 > [ 224.800558] ehci_hcd 0000:08:00.5: Bad frame counter change: 2495 2488 2496 > [ 285.280226] ehci_hcd 0000:07:00.7: Bad frame counter change: 2959 2952 2960 > [ 288.116527] ehci_hcd 0000:07:00.7: Bad frame counter change: 1103 1096 1104 > [ 363.278638] ehci_hcd 0000:09:00.3: Bad frame counter change: 3071 3064 3072 > [ 402.882274] ehci_hcd 0000:07:00.5: Bad frame counter change: 3599 3592 3600 > [ 933.467299] ehci_hcd 0000:07:00.7: Bad frame counter change: 1487 1480 1488 > [ 935.793586] ehci_hcd 0000:07:00.3: Bad frame counter change: 2791 2784 2792 > [ 941.374415] ehci_hcd 0000:09:00.7: Bad frame counter change: 1359 1352 1360 > [ 957.054450] ehci_hcd 0000:09:00.1: Bad frame counter change: 887 880 888 > [ 959.926293] ehci_hcd 0000:09:00.7: Bad frame counter change: 2511 2504 2512 > [ 987.516935] ehci_hcd 0000:08:00.1: Bad frame counter change: 2863 2856 2864 > [ 989.666143] ehci_hcd 0000:08:00.1: Bad frame counter change: 3695 3688 3696 > [ 992.596233] ehci_hcd 0000:09:00.1: Bad frame counter change: 2967 2960 2968 > [ 996.049852] ehci_hcd 0000:08:00.1: Bad frame counter change: 1583 1576 1584 > [ 996.439236] ehci_hcd 0000:09:00.1: Bad frame counter change: 983 976 984 > [ 1008.347011] ehci_hcd 0000:07:00.3: Bad frame counter change: 2343 2336 2344 > [ 1010.224562] ehci_hcd 0000:07:00.3: Bad frame counter change: 999 992 1000 > [ 1011.522870] ehci_hcd 0000:07:00.3: Bad frame counter change: 3207 3200 3208 > [ 1025.872075] ehci_hcd 0000:08:00.5: Bad frame counter change: 959 952 960 > [ 1025.992010] ehci_hcd 0000:07:00.3: Bad frame counter change: 327 320 328 > [ 1068.257022] ehci_hcd 0000:08:00.5: Bad frame counter change: 511 504 512 > [ 1069.914868] ehci_hcd 0000:08:00.5: Bad frame counter change: 1503 1496 1504 > [ 1070.633935] ehci_hcd 0000:08:00.5: Bad frame counter change: 3167 3160 3168 > [ 1085.826200] ehci_hcd 0000:08:00.5: Bad frame counter change: 1983 1976 1984 > [ 1170.924509] ehci_hcd 0000:08:00.7: Bad frame counter change: 599 592 600 > [ 1172.334679] ehci_hcd 0000:08:00.7: Bad frame counter change: 3703 3696 3704 > [ 1178.310917] ehci_hcd 0000:08:00.7: Bad frame counter change: 2423 2416 2424 > [ 1272.595721] ehci_hcd 0000:09:00.7: Bad frame counter change: 367 360 368 > [ 1276.286920] ehci_hcd 0000:09:00.7: Bad frame counter change: 1263 1256 1264 > [ 1278.631871] ehci_hcd 0000:09:00.7: Bad frame counter change: 3663 3656 3664 > [ 1285.614791] ehci_hcd 0000:09:00.7: Bad frame counter change: 2255 2248 2256 > [ 1328.120332] ehci_hcd 0000:08:00.7: Bad frame counter change: 2327 2320 2328 > [ 1375.211164] ehci_hcd 0000:08:00.7: Bad frame counter change: 2711 2704 2712 > [ 1381.651232] ehci_hcd 0000:07:00.1: Bad frame counter change: 1351 1344 1352 > [ 1436.468946] ehci_hcd 0000:07:00.3: Bad frame counter change: 3431 3424 3432 > [ 1450.948901] ehci_hcd 0000:07:00.1: Bad frame counter change: 3495 3488 3496 > [ 1453.130058] ehci_hcd 0000:07:00.1: Bad frame counter change: 487 480 488 > [ 1455.901514] ehci_hcd 0000:08:00.5: Bad frame counter change: 927 920 928 > [ 1456.061305] ehci_hcd 0000:08:00.5: Bad frame counter change: 2207 2200 2208 > [ 1460.503536] ehci_hcd 0000:08:00.5: Bad frame counter change: 927 920 928 > [ 1476.015388] ehci_hcd 0000:08:00.5: Bad frame counter change: 2303 2296 2304 > [ 1500.958988] ehci_hcd 0000:08:00.5: Bad frame counter change: 1407 1400 1408 > [ 1529.302618] ehci_hcd 0000:07:00.7: Bad frame counter change: 2543 2536 2544 > [ 1533.933048] ehci_hcd 0000:09:00.5: Bad frame counter change: 1743 1736 1744 > [ 1539.472964] ehci_hcd 0000:08:00.5: Bad frame counter change: 2719 2712 2720 > [ 1540.340230] ehci_hcd 0000:07:00.7: Bad frame counter change: 847 840 848 > [ 1626.699779] ehci_hcd 0000:09:00.3: Bad frame counter change: 2367 2360 2368 > [ 1667.247788] ehci_hcd 0000:08:00.3: Bad frame counter change: 1391 1384 1392 This pretty much confirms my guess. Notice that each time the bug occurs, the second number is 7 less than the first, and the third number is one larger than the first and is divisible by 8. > here's lspci: ... > 07:00.0 USB Controller: NetMos Technology MCS9990 PCIe to 4‐Port USB 2.0 Host Controller > 07:00.1 USB Controller: NetMos Technology MCS9990 PCIe to 4‐Port USB 2.0 Host Controller > 07:00.2 USB Controller: NetMos Technology MCS9990 PCIe to 4‐Port USB 2.0 Host Controller > 07:00.3 USB Controller: NetMos Technology MCS9990 PCIe to 4‐Port USB 2.0 Host Controller > 07:00.4 USB Controller: NetMos Technology MCS9990 PCIe to 4‐Port USB 2.0 Host Controller > 07:00.5 USB Controller: NetMos Technology MCS9990 PCIe to 4‐Port USB 2.0 Host Controller > 07:00.6 USB Controller: NetMos Technology MCS9990 PCIe to 4‐Port USB 2.0 Host Controller > 07:00.7 USB Controller: NetMos Technology MCS9990 PCIe to 4‐Port USB 2.0 Host Controller > 08:00.0 USB Controller: NetMos Technology MCS9990 PCIe to 4‐Port USB 2.0 Host Controller > 08:00.1 USB Controller: NetMos Technology MCS9990 PCIe to 4‐Port USB 2.0 Host Controller > 08:00.2 USB Controller: NetMos Technology MCS9990 PCIe to 4‐Port USB 2.0 Host Controller > 08:00.3 USB Controller: NetMos Technology MCS9990 PCIe to 4‐Port USB 2.0 Host Controller > 08:00.4 USB Controller: NetMos Technology MCS9990 PCIe to 4‐Port USB 2.0 Host Controller > 08:00.5 USB Controller: NetMos Technology MCS9990 PCIe to 4‐Port USB 2.0 Host Controller > 08:00.6 USB Controller: NetMos Technology MCS9990 PCIe to 4‐Port USB 2.0 Host Controller > 08:00.7 USB Controller: NetMos Technology MCS9990 PCIe to 4‐Port USB 2.0 Host Controller > 09:00.0 USB Controller: NetMos Technology MCS9990 PCIe to 4‐Port USB 2.0 Host Controller > 09:00.1 USB Controller: NetMos Technology MCS9990 PCIe to 4‐Port USB 2.0 Host Controller > 09:00.2 USB Controller: NetMos Technology MCS9990 PCIe to 4‐Port USB 2.0 Host Controller > 09:00.3 USB Controller: NetMos Technology MCS9990 PCIe to 4‐Port USB 2.0 Host Controller > 09:00.4 USB Controller: NetMos Technology MCS9990 PCIe to 4‐Port USB 2.0 Host Controller > 09:00.5 USB Controller: NetMos Technology MCS9990 PCIe to 4‐Port USB 2.0 Host Controller > 09:00.6 USB Controller: NetMos Technology MCS9990 PCIe to 4‐Port USB 2.0 Host Controller > 09:00.7 USB Controller: NetMos Technology MCS9990 PCIe to 4‐Port USB 2.0 Host Controller ... Evidently this is a silicon bug in the MosChip MCS9990. I couldn't find any errata documents on their web site; I wonder if they are aware of it? Do you want to try reporting it to them? Please try running again with that same patch, but this time erase the line that says: udelay(1); It would be great to know if the wrong values are still present when there's no delay between the two reads. Alan Stern -- To unsubscribe from this list: send the line "unsubscribe linux-usb" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
