![]() |
![]() |
![]()
Post
#1
|
|
Group: Members Posts: 233 Joined: 2-December 03 From: Boston, MA Member No.: 1,029 ![]() |
Howdy,
I recently discovered that Linux Kernel 2.6.14 breaks usbnet transfers from the Zaurus. Specifically, I was using 2.6.14-1 from the Fedora Core 4 update server with a C860. The first problem I noticed was that my connection hung when trying to enter a certain directory using samba (while browsing from my PC). Some directories worked fine and some hung. I next noticed that certain file transfers would hang, regardless of the protocol (I tried scp, ftp, and samba). Some files transferred just fine, others hung. I eventually determined that any file with a size >= 1458 bytes would hang. Any file with a size smaller than this transferred just fine. Additionally, I'm assuming that the directory browsing issue is related to the number of entries in the directory and, therefore, the size of the directory contents transfer. When I rebooted into kernel 2.6.12, everything worked fine again. I later determined that kernel 2.6.13 is also fine. I tried google'ing some kernel activity to try and find the cause and found some interesting stuff but nothing conclusive. I noticed that the kernel guys are still mucking around with usbnet: they've been engaged in a massive refactoring of that module since at least 2.6.11 (which I earlier discovered completely broke connectivity with the C860). I noticed one comment regarding the implementation of some sort of "hard mtu" feature which sounded somewhat suspicious to me. I haven't researched this any further though so it may be a red herring. The latest stable kernel release is 2.6.14-3 (Fedora's latest is 2.6.14-1) so it's possible it's been fixed in -2 or -3 but I haven't been able to try them and I haven't seen any patch that addresses this issue (but I haven't looked really well yet either). The ChangeLogs for -2 and -3 don't mention anything that sounds related. So, if you're a Linux & Zaurus user, I would recommend that you avoid 2.6.14 and try later versions with caution. ~ray |
|
|
![]() |
![]()
Post
#2
|
|
Group: Members Posts: 233 Joined: 2-December 03 From: Boston, MA Member No.: 1,029 ![]() |
FYI: The above solution also seems to work if applied on the PC-side. Apparently, the fix works so long as you apply the MTU=576 to either side of the connection.
Those of us running Fedora/Red Hat or distros derived from them probably have a /etc/sysconfig/network-scripts/ifcfg-usb0 script to set-up the Z's usbnet connection. I added MTU=576 to mine (after disabling my edits, described above, from the Z) and transfers don't hang. This strikes me as a "better" solution. IMO, the problem lies in Linux 2.6.14+ (and sounds like a bug to me) and so the fix should reside there as well. ~ray |
|
|
![]()
Post
#3
|
|
![]() Group: Members Posts: 118 Joined: 12-May 05 From: Oceanside,Ca Member No.: 7,121 ![]() |
QUOTE(raybert @ Feb 20 2006, 03:08 PM) FYI: The above solution also seems to work if applied on the PC-side. Apparently, the fix works so long as you apply the MTU=576 to either side of the connection. Those of us running Fedora/Red Hat or distros derived from them probably have a /etc/sysconfig/network-scripts/ifcfg-usb0 script to set-up the Z's usbnet connection. I added MTU=576 to mine (after disabling my edits, described above, from the Z) and transfers don't hang. This strikes me as a "better" solution. IMO, the problem lies in Linux 2.6.14+ (and sounds like a bug to me) and so the fix should reside there as well. ~ray Has anyone heard of a fix for this? I am trying to NAT my Z on FC4 and it seems I also need to set the MTU on my wireless card to 576 to get it to work.... ![]() Thanks, Henry |
|
|
![]() ![]() |
![]() |
Lo-Fi Version | Time is now: 19th April 2018 - 03:37 PM |