On Sat, Feb 10, 2007 at 02:34:43PM +0100, Waba wrote: > It took me a while, but I figured it out... nearly!
Great. We get closer to the real problem. > Further investigating, I tricked ls(1) into using the optimised libc > through LD_LIBRARY_PATH and obtained a core. mdb(1) told me that the > culprit was hiding at libc`memset+0x74. And finally, dis(1) revealed > that the faulty instruction is "movups (%esp), %xmm0", a SSE feature. > The %xmm0 register is apparently for storage purposes only, as the only > instructions used to access it are movups, movntps and movaps. This differs a bit from my investigations. I got the #UD in SVM always on the same RIP (which is unlikely when it is triggered in usermode). I assume the error comes from the lazy FPU switching code inside the kernel, triggered by the SSE instruction. But it is weird that this #UD in the kernel results in a SIGILL to the userspace process, but maybe Solaris does such things. It is possible that the Kernel and the Userspace on Solaris have different assumptions about the CPU capabilities? > At this point I hope that it makes a lot of sense to you, because I > have no idea why it works fine on Avi's Opteron, etc. Yes, thats another open question... Joerg -- Joerg Roedel Operating System Research Center AMD Saxony LLC & Co. KG ------------------------------------------------------------------------- Using Tomcat but need to do more? Need to support web services, security? Get stuff done quickly with pre-integrated technology to make your job easier. Download IBM WebSphere Application Server v.1.0.1 based on Apache Geronimo http://sel.as-us.falkag.net/sel?cmd=lnk&kid=120709&bid=263057&dat=121642 _______________________________________________ kvm-devel mailing list kvm-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/kvm-devel
