I sent this patch back on 9/19 and it never went anywhere. I''m sending it again, against the lastest unstable. -- Steve ==== Originally sent 9/19/2006 === If someone turns on verbose and/or debug the hypervisor slams the serial pretty badly when starting a FV domain. So I pulled the printk_ratelimit from Linux and put it into the hypervisor. Right now the only user of it is the MEM_LOG in arch/x86/mm.c which can really spit out a lot. Since printk_ratelimit is very helpful in the Linux kernel, I can see it being also used in HV. I changed it slightly from Linux to match the naming convention in Xen. Since printk is defined to printf, I declared the function printf_ratelimit and made a define of printk_ratelimit to just be a clone of printf_ratelimit. -- Steve Signed-off-by: Steven Rostedt <srostedt@redhat.com> _______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com http://lists.xensource.com/xen-devel
Hi, On Mon, 2006-10-02 at 12:39 -0400, Steven Rostedt wrote:> I sent this patch back on 9/19 and it never went anywhere. I''m sending > it again, against the lastest unstable.> If someone turns on verbose and/or debug the hypervisor slams the serial > pretty badly when starting a FV domain. So I pulled the > printk_ratelimit from Linux and put it into the hypervisor. Right now > the only user of it is the MEM_LOG in arch/x86/mm.c which can really > spit out a lot.A *lot*. With serial console, starting an HVM domain can take 30 seconds for me just because of these log messages. Rate-limiting is necessary for such cases. --Stephen _______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com http://lists.xensource.com/xen-devel