Commit Graph

3207 Commits

Author SHA1 Message Date
@Antelox
fe41e72b96 GitHub Workflow: (#2106)
- Switched from windows-2019 to windows-2022
- Switched to ubuntu-24.04-arm image for Linux aarch64 jobs
- Fixed Linux x86 job. It was failing but not reported
- Switched from Visual Studio 16 2019 generators to Visual Studio 17 2022
- Uncommented Windows MINGW32 static and shared jobs
- Generic clean-up
2025-02-15 18:23:10 +08:00
Disconnect3d
d03c0922e6 Fix #2103: qemu/target/ppc/mem_helper.c remove redundant return statements (#2104) 2025-02-15 01:04:30 +08:00
mio
7fd2aa47d4 Fix nuget workflow yaml 2025-02-13 23:42:21 +08:00
mio
3e99b859a8 Update nuget workflow to allow manual release 2025-02-13 23:39:45 +08:00
mio
a912fed662 CI(release): Update changelog 2025-02-13 22:23:08 +08:00
mio
e2915c978a CI(release): Fix 2025-02-13 22:21:25 +08:00
mio
0943d24c5f CI(release): ChangeLog 2025-02-13 22:20:58 +08:00
mio
68ec2d152f CI(release): Trigger CI 2025-02-13 21:20:38 +08:00
mio
cd1492d819 CI(release): Update docs 2025-02-13 21:17:29 +08:00
mio
1ea6b07653 Update nuget workflow events 2025-02-13 21:05:23 +08:00
mio
4caee59e8e Skip arm64 macos wheels 2025-02-13 20:42:58 +08:00
mio
2d410535f0 manually trigger release draft workflows 2025-02-13 19:43:47 +08:00
Amaan Qureshi
6b9c1c851c fix(arm): correct write to ARM coprocessor (#2099)
This code was commented out since 2021, but by default, the error
codewas initialized to `UC_REG_OK`, so there was no error returned
untila result, any write to `UC_ARM_REG_C1_C0_2` returned an error.
2025-02-13 19:25:26 +08:00
mio
967dbc4179 Allow a v-prefixed version 2025-02-13 19:20:43 +08:00
@Antelox
dea3c376d0 CI(full),CI(release): Python bindings: (#2100)
- Switched to the ubuntu-24.04-arm runner
- Bumped Windows runner to windows-2022 and Visual Studio 17 2022 GENERATORS
- Minor changes about checks in workflow jobs
- Fixed the pagefile job (even though should not be really needed)
- Refreshed the TO BE CHECKED regress tests to either update or remove the skip conditions
- Added a test to check if the created sdist archive is ok
2025-02-13 18:24:04 +08:00
mio
bf87b9a86d CI(full),CI(release): Increase pagefile to avoid OOM 2025-02-12 13:05:24 +08:00
mio
791557e404 CI(full),CI(release): Should use if instead of ifdef 2025-02-11 17:44:24 +08:00
mio
5981f10c5d CI(full),CI(release): Fix symbol clash (again) 2025-02-11 16:33:40 +08:00
mio
d437090bbc CI(full),CI(release): Fix symbol clash 2025-02-11 16:28:45 +08:00
mio
b4eb933ec8 CI(full),CI(release): Do not refer to ATOMIC128 symbols if not available 2025-02-11 16:24:49 +08:00
mio
381850356f CI(full),CI(release): More PPC64 atomic fixes 2025-02-11 10:18:01 +08:00
mio
0a02746572 CI(full),CI(release): Trigger full and bump zig 2025-02-10 22:23:53 +08:00
mio
1f82a09afc CI(full),CI(release): Trigger full 2025-02-10 22:21:40 +08:00
mio
b8e19b6eef CI(release): Bump 2.1.2 2025-02-10 22:11:12 +08:00
mio
ad33384cd2 Unsigned extension 2025-02-10 21:49:38 +08:00
mio
ca3912d9f7 Fix undefined behavior converting TCGv_i32 to TCGv_i64 2025-02-10 21:45:34 +08:00
mio
c915d13cce Fix pc issue of tcg opcode hooks 2025-02-10 21:35:17 +08:00
mio
77a841e53d Fix signature 2025-02-10 21:29:03 +08:00
mio
2619b12cf0 set pc_start on disasm 2025-02-10 21:28:23 +08:00
mio
904efc16c0 Should use tcg_ctx->pc_start as pc 2025-02-10 21:20:39 +08:00
mio
0c7017a584 more 22.04 pinning 2025-02-10 20:27:40 +08:00
mio
8082239cc3 Allow installing on Windows 2025-02-10 15:59:46 +08:00
mio
a903fa182b Avoid null ptr deref when writing to arm context pc register 2025-02-10 15:58:28 +08:00
mio
48fb28dfbc Update documents for MMU access 2025-02-10 15:36:03 +08:00
mio
a59bb4235e Use 22.04 to avoid segfault 2025-02-10 15:02:27 +08:00
dotcirill
dc1f7a88d5 Uc hook tcg improve (#2011)
* Add handling UC_TCG_OP_FLAG_CMP for ARM

Implementation is not well-tested and complete

* Hook ARM32 CMP and CMN for cmplog
2025-02-10 15:01:21 +08:00
Shivam7-1
ada8091ccc Fix heap buffer overflow in op_cksm function (#2096)
* Fix heap-buffer-overflow in op_cksm function

* Update header
2025-02-10 14:57:08 +08:00
Daniel Roethlisberger
e166cd93bb arm64 python: Avoid nested class in insn hook for 10x hook and 10% overall speedup (#2095)
* arm64 python: Avoid nested class in insn hook for 10x hook speedup

Promote CpReg to a module-level class to address unnecessary performance
reduction.  In a real-world use case tracing the emulation of real-world
machine code, this change reduces time spent in CpReg namedtuple
construction from 10% of overall time to below 1%, for a 10x speedup of
the insn hook itself, or a 10% overall speedup.  Measured using
cProfile, python 3.13.

* upgrade distro to 22.04

* revert to 22.04 for now

* also revert for wheels

---------

Co-authored-by: mio <mio@lazym.io>
2025-02-10 14:56:34 +08:00
xndcn
1ba25def8e Fix UC_HOOK_MEM on arm32 (#2091) 2025-01-26 13:28:24 +08:00
ZakDanger
d7c0497ee3 Added start in mips16 mode support, and unit test for it - on dev branch (#2089)
Co-authored-by: ZakDanger <ZakDanger@users.noreply.github.com>
2025-01-19 17:35:40 +08:00
7f6dcc7539 Test the pc is correctly sync-ed for arm64 2025-01-18 15:38:47 +08:00
ab565729e7 Format code 2025-01-18 15:18:49 +08:00
a0fa2c1249 Fix a warning 2025-01-18 15:18:37 +08:00
99c57a8ecf Add a test for #1643 2025-01-18 15:17:47 +08:00
0c512f91a1 Fix #1643
This adds an extra op to translator to allow the block hook sync pc in the very begining
2025-01-18 15:07:22 +08:00
Z. Liu
a78d690da5 configure: add <sys/timex.h> for clock_adjtime on musl (#2085)
for glibc, if _GNU_SOURCE is defined, <time.h> will include <sys/timex.h>
but not for musl, so add "#include <sys/timex.h>"

although `man clock_adjtime` said "#include <sys/timex.h>", but it won't
work for glibc w/o "#include <time.h>", I don't known why yet.

PS it seems clock_adjtime is used nowhere?

Signed-off-by: Z. Liu <zhixu.liu@gmail.com>
2025-01-14 08:47:54 +08:00
Daniel Roethlisberger
9682fc42f7 Allow uc_ctl_set_page_size() for arm64 (#2084) 2025-01-07 11:03:34 +08:00
8720632764 Add the unit test from saicao
Co-authored-by: Sai Cao <1665673333@qq.com>
2025-01-04 19:00:22 +08:00
Glenn Baker
8442eb6feb qemu/tcg: fix UC_HOOK_MEM_READ on aarch64. (#2028)
* qemu/tcg: fix UC_HOOK_MEM_READ on aarch64.

Directly jump into the slow path when there is any hookmem enabled. This
fixes #1908.

Signed-off-by: Glenn Baker <glenn.baker@gmx.com>

* qemu/tcg: fix UC_HOOK_MEM_READ on ppc64.

Directly jump into the slow path when there is any hookmem enabled.

Signed-off-by: Glenn Baker <glenn.baker@gmx.com>

* qemu/tcg: check for UC_HOOK_MEM_READ_AFTER.

Use has_hookmem() helper to determine wether "slow-path" TLB read is
needed. Add this helper to x86 architecture as well so that to check for
all hookmem.

Signed-off-by: Glenn Baker <glenn.baker@gmx.com>

* qemu/tcg: factor out has_hookmem().

It's the same implementation for all architectures, so factor out
has_hookmem() into tcg_uc_has_hookmem().

Signed-off-by: Glenn Baker <glenn.baker@gmx.com>

---------

Signed-off-by: Glenn Baker <glenn.baker@gmx.com>
2025-01-04 18:48:53 +08:00
mio
996ad57e34 Add an option to determine if building all-in-one objects archive 2025-01-04 17:35:09 +08:00