similar to: NPTL/TLS "emulation" idea

Just did a quick scan for that error message in the kernel source. Turns out it''s triggered in net/bridge/br_fdb.c:289. if (!is_valid_ether_addr(addr)) return -EADDRNOTAVAIL; Interestingly enough, it''s claiming that the mac addy isn''t any good! 8-( I''ll go add a printk line to that to see if I can determine what kind of data

I rebooted xenU from console (dom2) and exited (ctr-]). xm list does not report 2 anymore yet I can log into 2 via console and it is running just fine. Is this a bug? I''ll pull/rebuild and see what happens.. (this is xen-2.0, cset 1.1321.1.1, rh7.3 dom0/dom1) ------------------------------------------------------- This SF.net email is sponsored by: IT Product Guide on

Hi! $ xm create -n -f xmtest Using config file "xmtest". (vm (name testdomain) (memory 128) (image (linux (kernel /boot/vmlinuz-2.6.8-foo1-XenU) (ip :1.2.3.4:::tesmo:eth0:dhcp) (root ''/dev/xda1 ro'') (args 1) ) ) (device (vbd (uname phy:loop0) (dev xda1) (mode w))) (device (vif (mac

Hi folks, I''m trying to use the atropos scheduler with Xen. Couple of questions: - the scheduling param *has* to be specified at boot time right, or does Xen allow changing the scheduling scheme at run time?? - xm help atropos gives a very vague message at best, perhaps we could make it more elaborate: xm atropos DOM SLICE PERIOD LATENCY XTRATIME o is DOM the vm ID or the vm NAME (as

Hi, I have a question, If I start 2 or more guest linux, does the physical memory of code segment and readonly data segment be reused? What''s the relationship between the physical memory and the number of guest os that can create? xiaofeng ------------------------------------------------------- This SF.net email is sponsored by: IT Product Guide on ITManagersJournal Use IT

>From tools/libxc/xc_atropos.c: int xc_atropos_domain_set(int xc_handle, u32 domid, u64 period, u64 slice, u64 latency, int xtratime) which takes 6 arguments >From tools/python/xen/xm/main.py: class ProgAtropos(Prog): <snip> def main(self, args): if len(args) != 5: self.err("%s: Invalid argument(s)" % args[0])

Just to error test my own thinking here, there seems to be the following options for remote graphical access to a running domian vm. Besides the normal network service via a port and a raw console, the options are straight X, VNC, and XN. X - supports a single X window running on a local XServer, tied to a single process on the remote system. No remote desktop VNC - comes in two flavors of

> This should already be useful, at least for file-based disks. It''d be > nice to > have the ENBD script working and have some kind of iSCSI script at some > stage. For iSCSI, have a look at iscsi-init at sourceforge. http://iscsi-init.sourceforge.net. Load it as a module in an initrd and it uses some kernel parameters and automatically configures the network and then

Hi, Thanks to everyone involved in the Xen project. I think it is great and can''t wait to get my hands dirty with it. In the meantime, I cannot boot Dom0 on an ibm x335 and I am not sure what the problem is. Any insights are welcome. Here is what my grub.conf file looks like: title Fedora Core (2.6.8-1.521) root (hd0,0) kernel /vmlinuz-2.6.8-1.521 ro root=/dev/sda9

Is there a way to discovery how much free memory I still have to use in a new domain? Another question is about cpu usage. I guess that xen scheduler allocate a fixed time slice for all virtual machines and does not make any idea if the virtual machine is using or not using all this time, right? How I can identify if CPU is the bottleneck? Thanks! -- Flávio Bruno Leitner

Hi, Some questions here : 1. Is there a way to compile kernel 2.6.10 or 2.6.9-mm1 with xen 2.0? I tried both (with mkbuildtree from linux-2.6.9-xen-sparse), and both failed (on different places). My primary goal is to use reiser4 on both xen-0 and xen-U kernel, so I''m more interested in -mm kernel. 2. If I use kernel 2.4.27 (xen-U), top doesn''t show correct cpu usage

The unstable tree now includes support for SMP guests, i.e. domains which run on multiple cpus. SMP guests can use between 1 and 32 virtual cpus, even if the machine has fewer physical cpus. The code is highly experimental and performance will improve over time. To use SMP guests: - enable option CONFIG_SMP in the Linux 2.6 kernel config - dom0 will boot with upto the number of physical cpus in

Why is figlet source included in xen? Wouldn''t it make more sense to alter xen/Makefile, to have compile.h built from a .in, and the figlet view of the version info, and have the upstream developers regen the figlet part when the version changes? I''m willing to modify the build system to do this. I''m also fixing several other poor implementation of things.

>>> Hi all >>> >>> Thanks to Milan I am able to get further with xen and reiser4 >>> >>> - In swap line 6 and 7 status_flags.c >>> >>> #include <linux/bio.h> >>> #include <linux/page-flags.h> >>> >>> >>> - PACKED is redefined, but the definitions are the same, so the gcc >>>

In case it helps anyone, here''s a script I use to boot xen0 over iscsi with an initrd using linux-iscsi and iscsi-init (both on sourceforge). Some (or most) of it might be debian (sarge) specific though. The script goes in /etc/mkinitrd/scripts and puts the bits and pieces in place. It has to prepend stuff to /script in the resultant initrd so it does a few funny things. Add your Ethernet

hi all, the link on the website seems to be broken. does anyone have the doc for the hypervisor interface and the xend api. thanks. deepak ------------------------------------------------------- This SF.net email is sponsored by: IT Product Guide on ITManagersJournal Use IT products in your business? Tell us what you think of them. Give us Your Opinions, Get Free ThinkGeek Gift

FWIW, I assign my MAC addresses as "FE:FD:XX:00:YY:ZZ" Where: XX = what xen does to MAC addresses (00 in xen0, 01 in xenU) YY = vmid (unique across all physical servers) ZZ = interface number in xenU (00 = eth0, 01 = eth1) Works for me! Someone asked about trying to find a duplicate MAC address on a network - it''s pretty hard to do as the whole idea is that MAC addresses are

I would have sworn that someone on the list had a way to Xenify a stock RH kernel but I was not able to dig it up on the archives. If this does not exist I am sure it would be a nice addition to Xen to be able to patch and build some kernels that are compatable with some of the popular distributions. -- Alvin Starr || voice: (416)585-9971 Interlink Connectivity ||

Dear list, I''m trying to get Xen working on a P3-600E with 320 MB RAM. (256+64) It has suse linux 9.1 installed on it, and I tried to get xen running with both the binary download from the xen pages and the suse xen packages mentioned on this list some days ago. Both ''install packages/kernels'' lead to the same error: (XEN) Not enough memory for DOM0 memory

On my lex mainboard with a Via C3 cpu, xen 2.0(.1) reboots (nearly) immediately without any outputs. The same setup/version boots without problems on a via epia M board with a Via C3-2 cpu (Nehemia), having a rather different chipset. Starting with option noreboot, nobiostables doesn''t change anything visible. Is this a known problem? Or just a knowingly unsupported chipset? Have you