Loading ...
Sorry, an error occurred while loading the content.

Status

Expand Messages
  • Jim Buzbee
    OK, I ll kick this off with some of my current hacking status. Over the weekend I experimented with a few things. I m still trying to get a USB device such as
    Message 1 of 5 , Aug 2, 2004
    • 0 Attachment
      OK, I'll kick this off with some of my current hacking status.

      Over the weekend I experimented with a few things. I'm still trying
      to get a USB device such as a USB/Serial adaptor to work. I can build
      the proper modules and insert the proper modules. I see correct
      initialization statements from the modules in the syslog. But when I plug
      my USB/Serial adaptor in, I get an oops. This also occurs with a
      USB/Ethernet adaptor. If I plug my USB Webcam in after loading the proper
      Video For Linux drivers, I get the standard messages indication that there
      was no driver for this device enven though the driver appears to have
      registered for the proper vendor/model.

      I also confirmed that the NSLU2 has three(!) USB controllers. Two 1.1
      controllers and 1 2.0 controller. Strange...

      I've been able to convince the NSLU2 to generate a crontab entry for me in
      the following -dangerous- way. I used the standard Linksys menus for
      creating a backup job and saw that it put an entry in the config partion
      for the same. At boot time it generated a cron entry for doing the
      backup. I noticed that it used the name of my backup job as a parameter
      to the backup program. This created a potential hole for me if I could
      create a name that included a subshell command e.g. :

      "myjob`/share/hdd/conf/myscript.sh`"

      I couldn't convince the web interface to accept this bizarre "name", so I
      bypassed it and wrote my name directly into the flash config partition
      using dd (real men don't need no stinkin` web interfaces or fancy
      utilities, dd is all anyone should ever need!). Now at boot time, I get a
      nice crontab entry that calls my script. Unfortunately I haven't figured
      out a way to get it called other than once a day...




      Jim



      --------------------------------------------------------------------------------
      Jim Buzbee "I was gratified to be able to
      jbuzbee@... answer promptly, and I did. I
      http://batbox.org said I didn't know." Mark Twain
    • rwhitby
      ... Thanks Jim - who better to start the ball rolling that you :-) ... build ... I plug ... I remember seeing the same thing in the Asus WL-500g forums on
      Message 2 of 5 , Aug 2, 2004
      • 0 Attachment
        --- In nslu2-linux@yahoogroups.com, Jim Buzbee <jbuzbee@n...> wrote:
        > OK, I'll kick this off with some of my current hacking status.

        Thanks Jim - who better to start the ball rolling that you :-)

        > Over the weekend I experimented with a few things. I'm still trying
        > to get a USB device such as a USB/Serial adaptor to work. I can
        build
        > the proper modules and insert the proper modules. I see correct
        > initialization statements from the modules in the syslog. But when
        I plug
        > my USB/Serial adaptor in, I get an oops. This also occurs with a
        > USB/Ethernet adaptor.

        I remember seeing the same thing in the Asus WL-500g forums on
        www.chupa.nl - I'll see if I can dig up the resolution of that
        problem from that forum.

        Can you post the syslog messages? Does it look anything like these?:
        http://www.chupa.nl/forum/showpost.php?p=1446&postcount=13 or
        http://www.chupa.nl/forum/showthread.php?t=341&highlight=kernel+oops

        We might have to buy Oleg from the WL-500g forums an NSLU2 ;-)

        > I also confirmed that the NSLU2 has three(!) USB controllers. Two
        1.1
        > controllers and 1 2.0 controller. Strange...

        The 1.1 controller is the Intel CPU chip. The two 2.0 controllers
        are on the external USB controller chip. The 1.1 controller is
        probably not pinned out (like on the Linksys WRT54G series). The 2.0
        controller chip probably runs off the external PCI bus.

        -- Rod
      • Jim Buzbee
        ... Apologies if this message shows up twice, I may have sent it from the wrong account the first time... Here s the sequence : # insmod usbserial.o Using
        Message 3 of 5 , Aug 2, 2004
        • 0 Attachment
          On Aug 2, 2004, at 2:21 PM, rwhitby wrote:

          > --- In nslu2-linux@yahoogroups.com, Jim Buzbee <jbuzbee@n...> wrote:
          >> OK, I'll kick this off with some of my current hacking status.
          >
          > Thanks Jim - who better to start the ball rolling that you :-)
          >
          >> Over the weekend I experimented with a few things. I'm still trying
          >> to get a USB device such as a USB/Serial adaptor to work. I can
          > build
          >> the proper modules and insert the proper modules. I see correct
          >> initialization statements from the modules in the syslog. But when
          > I plug
          >> my USB/Serial adaptor in, I get an oops. This also occurs with a
          >> USB/Ethernet adaptor.
          >
          > I remember seeing the same thing in the Asus WL-500g forums on
          > www.chupa.nl - I'll see if I can dig up the resolution of that
          > problem from that forum.
          >
          > Can you post the syslog messages?

          Apologies if this message shows up twice, I may have sent it from the
          wrong account the first time...


          Here's the sequence :

          # insmod usbserial.o
          Using usbserial.o
          # insmod keyspan.o
          Using keyspan.o
          # lsmod
          Module Size Used by Tainted: P
          keyspan 109492 0 (unused)
          usbserial 19528 0 [keyspan]
          x1226-rtc 3664 0
          pbuttons 800 0 (unused)
          rbuttons 1304 0 (unused)
          ixp425_eth 17116 1
          ixp400 608216 0 [ixp425_eth]

          So far so good. Then I plug the adaptor in, the virtual address of 0 is
          the same each time I do this and, I think, for all oops/devices :

          # dmesg
          ...
          usb.c: registered new driver serial
          usbserial.c: USB Serial support registered for Generic
          usbserial.c: USB Serial Driver core v1.4
          usbserial.c: USB Serial support registered for Keyspan - (without
          firmware)
          usbserial.c: USB Serial support registered for Keyspan 1 port adapter
          usbserial.c: USB Serial support registered for Keyspan 2 port adapter
          usbserial.c: USB Serial support registered for Keyspan 4 port adapter
          keyspan.c: v1.1.4:Keyspan USB to Serial Converter Driver
          hub.c: new USB device 00:01.1-2, assigned address 2
          Unable to handle kernel NULL pointer dereference at virtual address
          00000000
          mm = c01baa80 pgd = c0004000
          *pgd = 00000000, *pmd = 00000000
          Internal error: Oops: 0
          CPU: 0
          pc : [<c011640c>] lr : [<00000003>] Tainted: P
          sp : c02e7ea4 ip : c58e1758 fp : c02e7f4c
          r10: 00000000 r9 : 00000003 r8 : c17b7200
          r7 : c17b7200 r6 : c58c4a68 r5 : 00000000 r4 : 00000000
          r3 : 0000010b r2 : 0000010b r1 : 00000000 r0 : c17b7200
          Flags: nZCv IRQs on FIQs on Mode SVC_32 Segment kernel
          Control: 39FF Table: 01788000 DAC: 0000001D
          Process khubd (pid: 9, stack limit = 0xc02e6368)
          Stack: (0xc02e7ea4 to 0xc02e8000)
          7ea0: c58e1970 c58c325c 00000000 00000000 00000000 c58e1960
          00000000
          7ec0: 00008f2f c02e6000 c004e444 00000000 c16a0a40 c02e6000 a0000013
          c16a0a40
          7ee0: c0069b70 00000000 c0116aec c02e7eec c02e7eec 00000001 00000000
          00000000
          7f00: c02e6000 00000000 00000000 80000200 c1870ba0 00000000 c17b7200
          00000009
          7f20: 690541f1 c58c4a4c c18707c0 00000000 c17b7200 c58c4a2c 00000003
          00000000
          7f40: c01c240c c02e7f50 c01166dc c58c31fc 00000000 c17b7224 00000000
          c17b7200
          7f60: 00000000 00000001 00000018 0000000a 690541f1 c0258420 00000000
          c01167c4
          7f80: 00000000 c17b7200 00000008 00000000 c01186c8 00000008 c17b7200
          00000000
          7fa0: c0252a00 c0119ff0 00000000 00000000 c0258420 00000001 c0252a00
          c01bad14
          7fc0: 0001708c c011a1d8 00010101 c02e7fe0 c0250000 c01e9614 c01c6fe0
          c011a4ac
          7fe0: 00000000 c02e6000 00000000 00000000 00000000 c0046dc0 00000000
          010b010b
          Backtrace:
          Function entered at [<c58c31f0>] from [<c01166dc>]
          Function entered at [<c01c24cc>] from [<00000000>]
          Code: e5dc200c e5d030d2 e1520003 1affffcc (e894000c)
        • sveasoft
          ... address ... This looks like the same kernel oops reported here in the 2.3 kernel development
          Message 4 of 5 , Aug 3, 2004
          • 0 Attachment
            > hub.c: new USB device 00:01.1-2, assigned address 2
            > Unable to handle kernel NULL pointer dereference at virtual
            address
            > 00000000


            This looks like the same kernel oops reported here in the 2.3 kernel
            development
            http://www.ussg.iu.edu/hypermail/linux/kernel/0005.2/0591.html.

            The bug seems to have to do with custom USB chip code for variations
            in the internal register layout. Could Linksys have used a modified
            kernel with patches for the USB controller chip?

            I won't receive my NSLU2's for a few days so I can't reproduce or
            test this as yet.

            James
          • Jim Buzbee
            ... It is possible. There was another funny think I noticed over the weekend but didn t have time to dig into. I plugged a hub into the box, and then a
            Message 5 of 5 , Aug 3, 2004
            • 0 Attachment
              On Tue, 3 Aug 2004, sveasoft wrote:

              >
              > > hub.c: new USB device 00:01.1-2, assigned address 2
              > > Unable to handle kernel NULL pointer dereference at virtual
              > address
              > > 00000000
              >
              >
              > This looks like the same kernel oops reported here in the 2.3 kernel
              > development
              > http://www.ussg.iu.edu/hypermail/linux/kernel/0005.2/0591.html.
              >
              > The bug seems to have to do with custom USB chip code for variations
              > in the internal register layout. Could Linksys have used a modified
              > kernel with patches for the USB controller chip?


              It is possible. There was another "funny" think I noticed over the
              weekend but didn't have time to dig into. I plugged a hub into the box,
              and then a flash drive into the hub. No problem. It mounted normally.
              Then I plugged another flash drive into the hub to see if I could at least
              manually have two mounted. At this point the LED on the second flash
              drive continuously blinked instead of a steady on like normal. Then
              neither disk was available, but if I recall, the first was still shown to
              be mounted.

              I'm pretty sure that the source we have for the box is incomplete. I know
              that thttp has been modified, but the source tree provided is virgin.
              In addition, the kernel ID tag you get at boot time does not match the id
              in the kernel Makefile provided in the snapgear tree.

              >
              > I won't receive my NSLU2's for a few days so I can't reproduce or
              > test this as yet.
              >


              Looking forward to your input.

              Jim


              > James
              >
              >
              >
              >
              >
              > Yahoo! Groups Links
              >
              >
              >
              >
              >

              --------------------------------------------------------------------------------
              Jim Buzbee "I was gratified to be able to
              jbuzbee@... answer promptly, and I did. I
              http://batbox.org said I didn't know." Mark Twain
            Your message has been successfully submitted and would be delivered to recipients shortly.