- 14 Dec, 2020 2 commits
-
-
alex@thinkpad authored
--HG-- branch : qemu
-
alex@thinkpad authored
- required for any gcc 7.x and newer (tested up to 10.x) - doesn't break gcc 5.x - it also helps some recent clang versions - renamed qemu-2.5.0-gcc78.patch to qemu-2.5.0-recent-gcc.patch --HG-- branch : qemu
-
- 16 Apr, 2019 2 commits
-
-
alex@thinkpad authored
(should have been done in 2aacd2ae9274) --HG-- branch : qemu
-
alex@thinkpad authored
- counting active processes returns 1 when the system is idle (not 3) - even if the system has only one CPU core, it needs to be able to start a new job - poll the CPU load 10 times over a period of 0.5 seconds before making a decision --HG-- branch : qemu
-
- 15 Apr, 2019 2 commits
-
-
alex@thinkpad authored
(todo: update all our python scripts) --HG-- branch : qemu
-
alex@thinkpad authored
--HG-- branch : qemu
-
- 09 Apr, 2019 1 commit
-
-
alex@thinkpad authored
(i.e. go back to some vanilla version without full reinstallation) Usage: ./sure_copy_from_contrib.sh [-q] (from qemu-eos directory) Name starts with different letter to avoid typing ./copy_back_to_contrib.sh by mistake. --HG-- branch : qemu
-
- 12 Apr, 2019 1 commit
-
-
alex@thinkpad authored
(minor changes after e1c6fd64d735) --HG-- branch : qemu
-
- 11 Apr, 2019 2 commits
-
-
alex@thinkpad authored
- on all x86_64 Linux systems, user can choose between 32-bit and 64-bit toolchain - on all Intel systems, user can choose to override the existing ARM toolchain (and install a precompiled version that's known to work) - on x86 Linux systems, the script should install a 32-bit toolchain without asking - on all systems, user may keep the existing ARM compiler (arm-none-eabi-gdb) and compile gdb 8.1 from source - on non-Intel systems, the script should just check for a valid toolchain, and ask the user to install it manually (not tested) --HG-- branch : qemu
-
alex@thinkpad authored
Too much of a moving target, e.g. working on Xenial and Bionic (with different packages!) but broken on Cosmic. Keeping arm-none-eabi-gcc from package manager as an option. For gdb, we'll use either a precompiled toolchain (64-bit 7-2018-q2, or any 32-bit version), or compile gdb 8.1 from source. --HG-- branch : qemu
-
- 10 Apr, 2019 4 commits
-
-
alex@thinkpad authored
--HG-- branch : qemu
-
alex@thinkpad authored
(32-bit binaries are not going to work anyway) --HG-- branch : qemu
-
alex@thinkpad authored
QEMU install: on Ubuntu/WSL, install the precompiled toolchain even if another (valid) gcc version is present (option #2 in the Ubuntu/WSL menu) --HG-- branch : qemu
-
alex@thinkpad authored
- back to GDB 8.1 - GDB 8.2.x is no longer accepted - using precompiled toolchain 7-2018-q2-update for both Mac and WSL (with GDB 8.1.x) --HG-- branch : qemu
-
- 09 Apr, 2019 4 commits
-
-
alex@thinkpad authored
(8.1 still accepted as valid) --HG-- branch : qemu
-
alex@thinkpad authored
(use the same heuristics as with plain Ubuntu) --HG-- branch : qemu
-
alex@thinkpad authored
Quirk: if CPU is in ARM mode, the disassembler looks at the Thumb bit of the requested address. If CPU is in Thumb mode, the disassembler ignores the Thumb bit from the address, and always disassembles as Thumb. This inconsistency is also present in latest QEMU from git. Why? IDC scripts: comments containing disassembled Thumb code were updated (DIGIC 6/7/8 and 1300D). --HG-- branch : qemu
-
alex@thinkpad authored
(was breaking the tests; not required on this camera anyway) --HG-- branch : qemu
-
- 08 Apr, 2019 1 commit
-
-
alex@thinkpad authored
Instruction count is printed when returning from the function. Instructions executed by other tasks/interrupts (i.e. outside the tested function) are not included in the printed value. This feature is very useful for optimizing code. Example: -d debugmsg,calls -icount 5 --HG-- branch : qemu
-
- 09 Apr, 2019 1 commit
-
-
alex@thinkpad authored
It was quite hard to pass the test suite, as there were lots of edge cases: - context switch may happen either in a timer interrupt, or in a regular task (e.g. msleep, sempahores and so on) - context switch is not atomic - it happens a few instructions (or function calls!) afterh updating the current task pointer - until the registers are swapped for the new task, we are still on the stack of the previous task - there may be dummy task switches (i.e. from one task to the same task) - an interrupt can occur in the middle of a (non-atomic) context switch (!) - VxWorks models do not even use a unique task ID (!) Result: QEMU no longer crashes during logging tests! Issue: these cameras do not appear to use an unique task ID. Workaround: assign our own task IDs. --HG-- branch : qemu
-
- 08 Apr, 2019 1 commit
-
-
alex@thinkpad authored
QEMU test suite: only use 2/3 of available CPU cores. It appears to improve the repeatability of results in some CPU- and IO-intensive tests. In other words, don't believe the Hyper-Threading hype :) --HG-- branch : qemu
-
- 06 Apr, 2019 2 commits
-
-
alex@thinkpad authored
--HG-- branch : qemu
-
alex@thinkpad authored
Before: uint32_t x; cpu_physical_memory_read(addr, &x, 4) After: uint32_t x = eos_get_mem_w(s, addr); Also, minor cleanups to eos_get_current_task_* (not expected to make a difference) --HG-- branch : qemu
-
- 27 Mar, 2019 1 commit
-
-
alex@thinkpad authored
(FindFirst/FindClose/FindNext repeated many times) --HG-- branch : qemu
-
- 25 Mar, 2019 3 commits
-
-
alex@thinkpad authored
Only FindFirst/FindNext test for now; others to be added later. Compile with e.g.: cd minimal/qemu-fio make MODEL=60D clean make MODEL=60D install_qemu CONFIG_QEMU=y FIXME: works on DIGIC 4/5, fails on DIGIC >= 6, not tested on DIGIC 2/3. --HG-- branch : qemu
-
alex@thinkpad authored
(slightly faster, no longer times out) --HG-- branch : qemu
-
alex@thinkpad authored
(a8bdfa98fe3d) --HG-- branch : qemu
-
- 05 Mar, 2019 1 commit
-
-
alex@thinkpad authored
--HG-- branch : qemu
-
- 03 Mar, 2019 1 commit
-
-
alex@thinkpad authored
Some workarounds were required to emulate the bootloader. Communication between the two cores is not yet handled. --HG-- branch : qemu
-
- 04 Mar, 2019 4 commits
-
-
alex@thinkpad authored
(slave core is still 7D2S) --HG-- branch : qemu
-
alex@thinkpad authored
(previously, it was a mix between an incomplete 5D firmware and 1000D bootloader) --HG-- branch : qemu
-
alex@thinkpad authored
(adapted commit e8bb011bcb0b) --HG-- branch : qemu
-
alex@thinkpad authored
--HG-- branch : qemu
-
- 11 Feb, 2019 1 commit
-
-
alex@thinkpad authored
This used to work in earlier versions, but target_disas changed its syntax in 2.5.0. Useful for debugging io_trace on DIGIC 6 (with CONFIG_QEMU=y). --HG-- branch : qemu
-
- 04 Mar, 2019 1 commit
-
-
alex@thinkpad authored
simplified condition for 946_PRBS size info --HG-- branch : qemu
-
- 11 Feb, 2019 2 commits
-
-
alex@thinkpad authored
- enable bit (similar to PRBS on DIGIC 5 and earlier) - consistency check for base address and size (base address must be multiple of size - Cortex R4 TRM p.178) - subregion disable bits (if used) Useful for emulating io_trace. Note: QEMU already calls tlb_flush when writing to DRSR, DRBAR or DRACR. No need to apply e4fd44e88c8f on DIGIC 6. --HG-- branch : qemu
-
alex@thinkpad authored
--HG-- branch : qemu
-
- 03 Mar, 2019 3 commits
-
-
alex@thinkpad authored
--HG-- branch : qemu
-
alex@thinkpad authored
--HG-- branch : qemu
-
alex@thinkpad authored
(19f18f0492ea from recovery branch) --HG-- branch : qemu
-