[Qemu-devel] PATCH: Ensure RTL 8139 MMIO regions are 4k apart

2007-07-16 Thread Daniel P. Berrange
The current QEMU code for the RTL-8193 network device has some issues if there is more than one device activated in a guest. Specifically, even if you specify difference MAC addresses, inside the guest all NICs end up seeing the same MAC - the MAC of the last NIC. Full details are recorded in

Re: [Qemu-devel] PATCH: Ensure RTL 8139 MMIO regions are 4k apart

2007-07-16 Thread Paul Brook
The hw/rtl8139.c code is passing 0x100 in as the size parameter for the cpu_register_physical_memory call, despite the fact that the API contract says size has to be a multiple of page size. Have you tried recent CVS? This should not be necessary. Paul

Re: [Qemu-devel] PATCH: Ensure RTL 8139 MMIO regions are 4k apart

2007-07-16 Thread Daniel P. Berrange
On Mon, Jul 16, 2007 at 06:07:18PM +0100, Paul Brook wrote: The hw/rtl8139.c code is passing 0x100 in as the size parameter for the cpu_register_physical_memory call, despite the fact that the API contract says size has to be a multiple of page size. Have you tried recent CVS? This

Re: [Qemu-devel] Solaris 10 dies during install

2007-07-16 Thread Ben Taylor
Alexey Eremenko [EMAIL PROTECTED] wrote: Some news about Solaris on Qemu: My Solaris is v2005, x86-64 + x86 on the same DVD. I installed the x86-64 version. Ugh. don't bother with that version. Download S10 U3 (11/06) or S10 U4 (07/07) [if it's been released yet] Host: Fedora 7,

[Qemu-devel] Question about a particular trap generation scenario

2007-07-16 Thread John
Hi, I am trying to run a program (Windows guest) that I protected with a software protection product. The program is segfaulting when run and it appears to me that the problem lies in how Qemu is handling an anti debug measure that involves traps. The anti-debug sequence is