Environment: VMware (latest Portage stable) client running under Gentoo (kernel 2.4.20-gentoo-r2). The guest operating system was Windows 2000 Professional, but the problem is persistent after termination of the VMware client (in fact, this is when most of the trouble begins), therefore I do not believe the guest operating system in particular is at fault. Problem: I attempt to transfer a large file (~1GB) from the client to host filesystem over a Bridged (vmnet) connection over a Samba fileshare installed on the host. After about two and a half minutes, the transfer stops. An error is reported in the client OS indicating that the network path has been lost. Networking throughout the system from that point forward on any port or protocol is severely slowed to the point of intermittent connection to messaging services (via GAIM) and timeouts on incoming HTTP requests. A wget, for bandwidth testing purposes, from gentoo.oregonstate.edu reports a maximum connection speed of ~1.15K/s, where I can normally reach as high as ~350K/s on a decent day. This also seems to affect connections over the loopback interface and to LAN addresses, since I run an Apache webserver and seem achieve slower connections when connecting to myself as well. The same situation can be reproduced when a large number of smaller files are transferred. When the init.d script attempts to bring down the vmnet1 and vmnet8 (host-only and bridged) interfaces, it fails, and system processor usage increases to 100% in the process "vmnet-netifup". A repeating notification is displayed in standard output, which reads: "unregister_netdevice: waiting for vmnet1 to become free. Usage count = xxx" where "xxx" is a number of two to three digits. It becomes impossible to bring the net.eth0 interface down or kill the "vmnet-netifup" process. The system fails to perform an "init 0" or "init 6" operation, because these processes and services do not terminate. This means filesystems cannot be unmounted properly, and the system must be manually reset, which (obviously) potentially damages filesystems. This problem does not manifest itself in any way, shape, or form under a kernel compiled from vanilla-sources. The problem continued to be reproducible after several recompiles, with fresh sources. Reproducible: Always Steps to Reproduce: 1. Emerge gentoo-sources [2.4.20-r2] 2. Emerge and configure VMware for Linux. Use Host-Only and Bridged networking. Use existing Samba configuration (by binding it to the vmnet1 interface). 3. Attempt to transfer files between the host and guest filesystem via Samba (or any protocol, for that matter, as long as it is over the vmnet1 host-only interface). 4. When the file transfer fails to complete or bandwidth seems to sharply drop off, power off the guest machine and execute "/etc/init.d/vmware stop" 5. Observe. Actual Results: As described in the "details" section of this report, all networking on the affected system becomes dramatically slowed, exhibiting symptoms similar to one under exceptionally high-traffic conditions. Connection to most remote systems fails. Refer to the "details" section. Expected Results: File transfer between the guest and host operating systems should have gone unimpeded. VMware and networking services should have been stoppable. Processes should have been killable. In short, things should have worked normally. Note that the running system is of SMP architecture with dual CPUs. Portage 2.0.47-r10 (default-x86-1.4, gcc-3.2.2, glibc-2.3.1-r4) ================================================================= System uname: 2.4.20-gentoo-r2 i686 AMD Athlon(tm) Processor GENTOO_MIRRORS="http://gentoo.oregonstate.edu/ http://distro.ibiblio.org/pub/Linux/distributions/gentoo" CONFIG_PROTECT="/etc /var/qmail/control /usr/kde/2/share/config /usr/kde/3/share/config /usr/X11R6/lib/X11/xkb /usr/kde/3.1/share/config /usr/share/config" CONFIG_PROTECT_MASK="/etc/gconf /etc/env.d" PORTDIR="/usr/portage" DISTDIR="/usr/portage/distfiles" PKGDIR="/usr/portage/packages" PORTAGE_TMPDIR="/var/tmp" PORTDIR_OVERLAY="" USE="x86 oss 3dnow apm avi crypt cups encode gif jpeg libg++ mikmod mmx mpeg ncurses nls pdflib png quicktime spell truetype xml2 xmms xv zlib gdbm berkdb slang readline arts svga tcltk mysql X sdl gpm tcpd pam libwww ssl perl esd imlib oggvorbis qt kde motif opengl cdr dga python -gnome -gtk -java mozilla sse usb" COMPILER="gcc3" CHOST="i686-pc-linux-gnu" CFLAGS="-march=athlon-mp -O2 -pipe -fomit-frame-pointer -finline-functions -falign-jumps=5 -falign-loops=5 -falign-functions=64 -funroll-loops -mfpmath=sse -mmmx -msse -m3dnow" CXXFLAGS="-march=athlon-mp -O2 -pipe -fomit-frame-pointer -finline-functions -falign-jumps=5 -falign-loops=5 -falign-functions=64 -funroll-loops -mfpmath=sse -mmmx -msse -m3dnow" ACCEPT_KEYWORDS="x86" MAKEOPTS="-j3" AUTOCLEAN="yes" SYNC="rsync://rsync.us.gentoo.org/gentoo-portage" FEATURES="sandbox ccache"
could you please see if you can reproduce on pfeifer-sources-2.4.20_pre9? Thanks, Jay
Using sys-kernel/pfeifer-sources-2.4.20.1_pre9 I attempted to reproduce the problem under this kernel, and was unsuccessful. None of the symptoms described in this report occurred.
thanks for the info. these sources will turn into the next release of gentoo-sources... til then, use these and you should be fine. closing Jay
