2
I Use This!
Activity Not Available

News

Posted over 14 years ago
Ladies and gentlemen, we’re happy to release SerialICE 1.5 with many new features today.Download SerialICE 1.5 at http://www.serialice.com/downloads/SerialICE-1.5.tar.bz2Among the new features: - Rework memory and IO filters to provide more control - ... [More] Improved PCI, PCIe and memory access logging - Windows (MINGW and Cygwin) support - New mainboard supported: ASUS P2B - SerialICE connection now survives target resets - CPUID now honors ECX values - RDMSR/WRMSR now honor EDI unlock keys - Add LUA patch to correctly operate on 32bit hosts - Drop SerialICE specific machine type in QemuBest regards,Stefan Reinauer [Less]
Posted over 14 years ago
SerialICE has a new mailing list. Go to http://serialice.com/mailman/listinfo/serialice to subscribe to the list.
Posted over 14 years ago
It’s been a while, and we’ve been working on a new version of SerialICE. After some intermediate steps that were shared among developers, we’re glad to release SerialICE 1.4 today.Download SerialICE 1.4 at ... [More] http://www.serialice.com/downloads/SerialICE-1.4.tar.bz2Among the new features: - Completely scriptable filtering and logging with LUA and bitlib. - Kconfig/Kbuild support (make menuconfig) - New mainboards supported: Intel D945GCLF, ASUS M2V-MX SE, MSI MS6178, Thomson IP1000, Dell PowerEdge s1850, RCA RM4100 - Patch against latest Qemu version 0.11.0. - Support for compiling with XMMSTACK and ROMCC. - Many bugs fixed: serial communication, data corruption by signed/unsigned casts, ...Best regards,Stefan Reinauer [Less]
Posted almost 15 years ago
coresystems GmbH is proud to release the first version of our "Integrated Circuit Emulator over Serial", short SerialICE.This piece of software consists of two parts: - a serial console "rom shell" compiled with romcc, with minimal footprint. (Due to ... [More] romcc the image is still 128k because it did not fit in 64k but this can be optimized later) - a patch to Qemu 0.10.4, which adds a new "SerialICE" machine.Short description:SerialICE is a BIOS/Firmware debugging tool. It allows you to run and observe BIOS images (such as coreboot®: http://www.coreboot.org/) written for real hardware in Qemu (http://www.qemu.org/) for debugging purposes. Thanks to Qemu's compelling feature set, it's also possible to debug this BIOS code with GNU GDB.SerialICE can be downloaded from http://www.serialice.com/downloads/SerialICE-1.0.tar.bz2With "qemu -m serialice -serialice /dev/ttyS0 -L path-to-your-bios.bin-dir -hda /dev/zero" you can run an arbitrary BIOSbinary written for your target hardware in Qemu, thus logging all IO and memory accesses. Those operations will additionally be transmitted to the target system's shell and are executed there, while their results are submitted back to Qemu.Operations sent to the target: - memory reads/writes (some of them) - IO reads/writes - MSR reads/writes - CPUID calls (the bios code path might rely on this)Note: The code is very experimental and still buggy, but it was already useful in some debugging scenarios we had and was able to reveil information that would normally only be available with a hardware debugger of the price of a new car. Don't expect SerialICE to completely replace a ICE/JTAG/ITP device, but it might just work for your case, as it did for us.The code needs minimal board/chipset specific setup in order to have serial console operational for communication with Qemu. See mainboard/* for an example. This release contains demo code for two mainboards with Intel CPUs. Also, some hardware accesses have to be caught in the Qemu code (hw/serialice.c) in order to prevent the system from locking up(ie. by disabling the serial console).Known issues: - The code is ugly, and the Qemu part is light years from a state where integration would be possible. - infrastructure for compiling with gcc + xmmstack is there, but it still fails with some odd assembler errors. This should push the SerialICE rom shell clearly below 64k again. - microcode updates from within emulated ROM code will fail. - some rarely used calls of cpuid will not give the correct information (those using two registers for input)Special thanks go to • Alex Graf for listening to my odd ideas while embedded world and supporting this project from early on. • Paul Brook for helping me find the last bug that prevented surviving all of RAM initialization on one board. • Patrick Georgi for Development and Testing. • Ron Minnich for advice and encouragement. • Eric Biederman for romccComments and patches are of course very welcome!Best regards,Stefan Reinauer [Less]