-
Notifications
You must be signed in to change notification settings - Fork 1.3k
Typically, it’s due to
- Instrumenting every instruction executed.
- Instrumenting every memory access.
Optimize your program with less instrumentation, e.g. by using UC_HOOK_BLOCK
instead of UC_HOOK_CODE
Updating PC is a very large overhead (10x slower in the worst case, see FAQ above) for emulation so the PC sync guarantee is explained below in several cases:
- A
UC_HOOK_CODE
hook is installed. In this case, the PC is sync-ed everywhere within the effective range of the hook. However, on some architectures, the PC might by sync-ed all the time if the hook is installed in any range. Note usingcount
inuc_emu_start
implies installing aUC_HOOK_CODE
hook. - A
UC_HOOK_MEM_READ
orUC_HOOK_MEM_WRITE
hook is installed. In this case, the PC is sync-ed exactly before any read/write events within the effective range of the hook. - Emulation (
uc_emu_start
) terminates without any exception. In this case, the PC will point to the next instruction. - No hook mentioned above is installed and emulation terminates with exceptions. In this case, the PC is sync-ed at the basic block boundary, in other words, the first instruction of the basic block where the exception happens.
Below is an example:
mov x0, #1 <--- the PC will be here
mov x1, #2
ldr x0, [x1] <--- exception here
If ldr x0, [x1]
fails with memory exceptions, the PC will be left at the beginning of the basic block, in this case mov x0, #1
.
However, if a UC_HOOK_MEM_READ
hook is installed, the PC will be sync-ed:
mov x0, #1
mov x1, #2
ldr x0, [x1] <--- exception here and PC sync-ed here
Unicorn is a pure CPU emulator and usually it’s due to no handler registered for instructions like syscall
and SVC
. If you expect system emulation, you probably would like qiling framework.
Currently, only a small subset of the instructions can be instrumented.
On x86, all available instructions are: in
out
syscall
sysenter
cpuid
.
- Some instructions are not enabled by default on some architectures. For example, you have to setup CSR on RISC-V or VFP on ARM before emulating floating-point instructions. Refer to the corresponding manual to check if you leave out possible switches in special registers.
- Different CPU models support different sets of instructions. This is especially observed on ARM CPUs. For example, for
THUMB2
big-endian instructions, consider setting CPU model tocortex-r5
orarm_max
. See #1725 and #1724. - If you are on ARM, please check whether you are emulating a THUMB instruction. If so, please use
UC_MODE_THUMB
and make sure the starting address is odd. - If it's not the cases above, it might be some newer instruction sets that qemu5 doesn’t support.
- Note some instruction sets are not implemented by the latest QEMU.
If you are still using Unicorn1, please upgrade to Unicorn2 for better support.
There are several possibilities, e.g.:
- The instruction might access memory multiple times like
rep stos
in x86. - The address to access is bad-aligned and thus the MMU emulation will split the access into several aligned memory access. In worst cases on some arch, it leads to byte by byte access.
This is a minor change in memory hooks behavior between Unicorn1 and Unicorn2. To gracefully recover from memory read/write error, you have to map the invalid memory before you return true.
It is due to the fact that, if users return true
without memory mapping set up correctly, we don't know what to do next. In Unicorn1, the behavior is undefined in this case but in Unicorn2 we would like to force users to set up memory mapping in the hook to continue execution.
See the sample for details.
For MIPS, you might have an address that falls in MIPS kseg
segments. In that case, MMU is bypassed and you have to make sure the corresponding physical memory is mapped. See #217, #1371, #1550.
For ARM, you might have an address that falls in some non-executable segments. For example, for m-class ARM cpu, some memory area is not executable according to the ARM document.
This is intended as python signal module states:
A long-running calculation implemented purely in C (such as regular expression matching on a large body of text) may run uninterrupted for an arbitrary amount of time, regardless of any signals received. The Python signal handlers will be called when the calculation finishes.
A workaround is to start emulation in another thread.
Unicorn is a fork of QEMU and inherits most QEMU internal mechanisms, one of which is called TB chaining. In short, every block (in most cases, a basic block
) is translated, executed and cached. Therefore, any operation on cached addresses won't immediately take effect without a call to uc_ctl_remove_cache
. Check a more detailed discussion here: #1561
Note, this doesn't mean you have to care about Self Modifying Code because the read/write happens within emulation (TB execution) and QEMU would handle such special cases. For technical details, refer to the QEMU paper.
TLDR: To ensure any modification to an address will take effect:
- Call
uc_ctl_remove_cache
on the target address. - Call
uc_reg_write
to write current PC to the PC register, if the modification happens during emulation. It restarts emulation (but doesn't quituc_emu_start
) on current address to re-translate the block.
As stated, Unicorn is a pure CPU emulator. For such emulation, you have two choices:
- Use the
timeout
parameter ofuc_emu_start
- Use the
count
parameter ofuc_emu_start
After emulation stops, you may check anything you feel interested and resume emulation accordingly.
Note that for cortex-m exec_return
, Unicorn has a magic software exception with interrupt number 8. You may register a hook to handle that.
To provide end users with simple API, Unicorn does lots of dirty hacks within qemu code which prevents it from sync painlessly.
Yes, it’s possible but that is not Unicorn’s goal and there is no simple switch in qemu to disable softmmu.
Starting from 2.0.2, Unicorn will emulate the MMU depending on the emulated architecture without further hacks. That said, Unicorn offers the full ability of the target MMU implementation. While this enables more possibilities of Uncorn, it has a few drawbacks:
- As previous question points out already, some memory regions are not writable/executable.
- You have to always check architecture-specific registers to confirm MMU status.
-
uc_mem_map
will always deal with physical addresses whileuc_emu_start
accepts virtual addresses.
Therefore, if you still prefer the previous paddr = vaddr
simple mapping, we have a simple experimental MMU implementation that can be switched on by: uc_ctl_tlb_mode(uc, UC_TLB_VIRTUAL)
. With this mode, you could also add a UC_HOOK_TLB_FILL
hook to manage the TLB. When a virtual address is not cached, the hook will be called. Besides, users are allowed to flush the tlb with uc_ctl_flush_tlb
.
In theory, UC_TLB_VIRTUAL
will achieve better performance as it skips all MMU details, though not benchmarked.
Unicorn uses at several places logging by the qemu implementation. This might provide a first glance what could be wrong.
The logs contains optionally the filename and the line number including additional messages to indicate what is happening. However, the qemu logs are partially commented-out and incomplete, but give it a try. You might want to dig deeper - and add your own log messages where you expect or try to find the bug.
To enable logs, you must recompile Unicorn with -DUNICORN_LOGGING=yes
to cmake.
Logs are written in different log levels, which might result into a very verbose logging if enabled. To control the log level information, two environment variables could be used.
UNICORN_LOG_LEVEL
and UNICORN_LOG_DETAIL_LEVEL
.
These environment variables are parsed into uint32_t
values once, (due to performance reasons)
so set these environment variables before you execute any line of Unicorn.
Allowed are hexa-decimal, decimal and octal values, which fits into a buffer of 10 chars. (see stroul for details).
To define how detailed and what should be logged, use the following environment variables:
-
UNICORN_LOG_LEVEL
=<32bit mask>- The qemu bit mask what should be logged.
- Use the value of
UINT32_MAX
to log everything. - If no bit is set in the mask, there will be no logging.
-
UNICORN_LOG_DETAIL_LEVEL
=<level>- The level defines how the filename and line is constructed.
- 0: no filename and no line is used.
- 1: full filename including the leading path is used with line information.
- 2: just the filename with line information. It might be a little confusing, as the file name can be used in several places.
- If unsure or unwanted, leave this variable undefined or set it to 0.
- The level defines how the filename and line is constructed.
As an example to set up the environment for python correctly, see the example below.
import os
os.environ['UNICORN_LOG_LEVEL'] = "0xFFFFFFFF" # verbose - print anything
os.environ['UNICORN_LOG_DETAIL_LEVEL'] = "1" # full filename with line info
Please note that file names are statically compiled in and can reveal the paths of the file system used during compilation.
Yes! However, Unicorn by default disables it and enabling it involves a few coding and document reading.
TLDR:
Taken from #1789.
uc_arm64_cp_reg reg;
// SCR_EL3
reg.op0 = 0b11;
reg.op1 = 0b110;
reg.crn = 0b0001;
reg.crm = 0b0001;
reg.op2 = 0b000;
err = uc_reg_read(uc, UC_ARM64_REG_CP_REG, ®);
assert(err == UC_ERR_OK);
// NS && RW && API
reg.val |= (1 | (1<<10) | (1<<17));
err = uc_reg_write(uc, UC_ARM64_REG_CP_REG, ®);
assert(err == UC_ERR_OK);
// SCTLR_EL1
reg.op0 = 0b11;
reg.op1 = 0b000;
reg.crn = 0b0001;
reg.crm = 0b0000;
reg.op2 = 0b000;
err = uc_reg_read(uc, UC_ARM64_REG_CP_REG, ®);
assert(err == UC_ERR_OK);
// EnIA && EnIB
reg.val |= (1<<31) | (1<<30);
err = uc_reg_write(uc, UC_ARM64_REG_CP_REG, ®);
assert(err == UC_ERR_OK);
// HCR_EL2
reg.op0 = 0b11;
reg.op1 = 0b100;
reg.crn = 0b0001;
reg.crm = 0b0001;
reg.op2 = 0b000;
// HCR.API
reg.val |= (1ULL<<41);
err = uc_reg_write(uc, UC_ARM64_REG_CP_REG, ®);
assert(err == UC_ERR_OK);
For further explanation, refer to related ARM documents. Here is an incomplete list:
- System register control of pointer authentication
- EnIA & EnIB
- HCR.API Note you could find the definitions of these registers at the end of corresponding documents.
This is intended for Windows. See discussion in #1841.
Please create a github issue and provide as many details as possible.
- Simplified version of your script / source
- Make sure that "no" external dependencies are needed.
- E.g. remove additional use of capstone or CTF tools.
- Used Unicorn git-hash commit
- Make sure to exclude any changes of you made in unicorn.
- Alternativily provide the repo link to your commit.
- Detailed explaination what is expected
- Try to verify if the instructions can be processed by qemu.
- Dumping the registers of unicorn and qemu helps a lot.
- Detailed explaination what is observed
- Describe what's going on (and what you might think about it).
- Output from your executed script
- You might have additional log messages which could be helpful.
- Output from the qemu-logs
- Try to gather more informations by enabling the qemu logging.
- More details
- Attach more details to help reproduce the bug.
- Like attaching a repo link to the CTF challenge containing the binary or source code.
See milestones and coding convention.
Be sure to send pull requests for our dev branch only.
Prior to 2.0.0, Unicorn is based on qemu 2.2.1. After that, Unicorn is based on qemu 5.0.1.