7de130a5d7
s390x/tcg: Fix BRASL and BRCL with large negative offsets
...
This is a backport of the following upstream commits:
- commit fc3dd86a290a ("s390x/tcg: Fix BRASL with a large negative offset")
- commit 16ed5f14215b ("s390x/tcg: Fix BRCL with a large negative offset")
Signed-off-by: Ilya Leoshkevich <iii@linux.ibm.com>
2022-03-17 14:23:57 +01:00
09b15e9071
Fix exits wrongly cleared in nested uc_emu_start
2022-03-06 23:40:34 +01:00
12fa73f2b6
Generate check_exit_request for the first tb
...
Consider this case:
t:inc ecx,
inc edx,
jmp t
It would generate exactly one block and TB chain will generate an inline
jump. If we don't check exit at the entry of the block, we will never be
terminated.
This test case is recorded in the commit message since it's a bit
painful to write multithread related code that builds on all platforms.
2022-03-05 22:42:41 +01:00
b7bc13650c
Sync PC after we exit
...
We should sync PC if there is no UC_HOOK_CODE hook
2022-03-05 22:42:41 +01:00
9f62c71094
Sync PC for SYS instruction hook
2022-02-27 15:30:27 +01:00
e5207a1363
Implement UC_HOOK_INSN for aarch64 MRS/MSR/SYS/SYSL
2022-02-27 15:28:31 +01:00
8e2f70a35b
Merge pull request #1559 from yrashk/aarch64-use-map-jit
...
Problem: Could not allocate dynamic translator buffer
2022-02-25 22:46:19 +01:00
d946114dfe
Set EFLAGS correctly on startup
2022-02-25 22:44:42 +01:00
84e796494f
Format code
2022-02-25 22:24:25 +01:00
caf2fe1ddb
Clear state when exiting from uc_emu_start
...
This may cause unexpected result when we are in a nested uc_emu_start
2022-02-25 22:23:44 +01:00
ad4b56ad54
Remove armeb and aarcheb in configure
2022-02-22 15:28:16 +01:00
f652340bec
Problem: Could not allocate dynamic translator buffer
...
`Could not allocate dynamic translator buffer` error while
trying to perform any memory-related operation.
This happens on aarch64 (MacBook Pro M1)
Solution: ensure aarch64 architecture is properly checked against
Fixes #1556
2022-02-20 07:46:30 -08:00
d96083d4d1
Fix pc after ppc sc inst
2022-02-20 20:42:46 +08:00
8d9ca1ce84
Format
2022-02-17 10:37:34 +01:00
27ef63cc8d
Add UC_PPC_REG_CR
2022-02-15 22:07:53 +01:00
73e4a90d3a
Manually revert 63a445cbba
2022-02-13 10:13:01 +01:00
4266196b2d
Fix the divergence with master.
2022-02-13 10:08:58 +01:00
e382ca102a
Fix the regression bug
2022-02-13 09:52:00 +01:00
05b9a021dc
Merge pull request #1453 from scribam/reg-read-write-fpscr
...
unicorn_arm: add reg_read/write operations for FPSCR and FPSID
2022-02-13 09:02:32 +01:00
52e0963cc7
Backward compatibility to enable full 64bits address space
2022-02-12 22:54:21 +01:00
81eb7da837
Backward compatibility for c13_c0_3
2022-02-12 22:31:10 +01:00
3c4477d622
Fix another undefined shift found by sanitizer
2022-02-12 19:57:19 +01:00
93c602ead1
FIx anohter undefined shift
2022-02-12 19:47:51 +01:00
e38b1c8af3
Fix the undefined shift
2022-02-12 19:37:32 +01:00
e548cd9eb1
Fix the undefined shift
2022-02-12 18:22:56 +01:00
d0de673208
Fix the undefined lshift negative numbers
2022-02-12 17:47:57 +01:00
9ff335efdc
Fix memory leak of the empty_view
2022-02-12 15:47:25 +01:00
c2bb5c8838
Fix ns and s in cp reg encoding
2022-02-12 14:29:42 +01:00
58fc952230
Remove armeb-softmmu and aarch64eb-softmmu
2022-02-12 14:15:54 +01:00
f511d4a807
Also return error for context read/write
2022-02-11 22:19:03 +01:00
3e6665db00
Implement coprocessor register read/write for arm64
2022-02-11 22:13:01 +01:00
8bc1489210
Implement coprocessor register read/write for arm
2022-02-11 21:45:37 +01:00
141a558dd8
fix conflicts
2022-01-28 10:30:51 +08:00
ad984b9366
s390x TCG shift instruction fixes
...
Cherry-pick the following upstream commits:
521130f267 target/s390x: Fix SLDA sign bit index
57556b28af target/s390x: Fix SRDA CC calculation
df103c09bc target/s390x: Fix cc_calc_sla_64() missing overflows
6da170beda target/s390x: Fix shifting 32-bit values for more than 31 bits
2022-01-27 13:15:54 +01:00
5f89981a3c
Free info->name on demand
2022-01-20 21:34:21 +01:00
55b4865945
Fix memory leak of custom helpers
2022-01-20 19:50:57 +01:00
7095605607
Merge branch 'dev' into systemz
2022-01-18 21:10:55 +01:00
218bddc0e0
Only use MAP_JIT on Apple Silicon
...
MAP_JIT causes performance regression for fork()
See https://github.com/desktop/desktop/issues/12978
2022-01-18 21:01:49 +01:00
28e791a37f
Add debug tracing feature
...
It's disabled by default, use -DUNICORN_TRACER=on to enable it
2022-01-18 19:35:43 +01:00
ea9c7425b0
Fix the wrong PC when arm translation fectches unmapped memory
...
This behavior keeps the same with Unicorn1, though, different from arm doc
2022-01-16 16:42:38 +01:00
a5ceca6d51
Remove the static variable in flatviews_init
...
Or we may get an invalid old (and free-ed) uc instance reference
2022-01-15 22:11:14 +01:00
459a595a98
Merge branch 'dev' into s390x
...
Mostly for bindings update.
2022-01-15 20:56:39 +01:00
6ed2214399
Rebuilt hflags when swithing modes
...
Or we may get the wrong mode during translation
2022-01-14 19:37:48 +01:00
33afdcf872
Save CC at the end of emulation
2022-01-10 21:48:03 +01:00
9ac796531a
Don't cache S390SkeyState and S390SkeysClass
2022-01-10 19:18:52 +01:00
980eae7f44
Sync PC at the end of emulation
2022-01-10 15:45:56 +01:00
441afe17e6
Add psw.mask register
2022-01-10 15:34:04 +01:00
71f044ca50
Merge branch 'dev' into s390x
2022-01-10 15:17:42 +01:00
36afa1022c
More PPC registers
...
Add FPR0-31, CR0-7, LR, CTR, MSR, XER, FPSCR for PPC
Add a test for ppc32 float point
2022-01-10 15:16:10 +01:00
be7fbf1306
Handle CPU fault when invalidating TB cache
2022-01-08 22:10:17 +01:00
8ad9f8ecb1
This reverts Hack 05ba21160619724033ec83469bbb66bda9e3f5fb and applies the correct fix
...
And enable experimental v8 support for arm max cpu
2022-01-05 21:58:40 +01:00
c3a49766d8
Fix #1522
2022-01-05 20:02:41 +01:00
7a886f59df
Fix #1525
2022-01-05 19:38:22 +01:00
6fabf30537
Fix a invalid memory access
...
Note: This probably addresses the ramdom failed CI on mingw64
2022-01-05 19:12:36 +01:00
d854e22301
Add x87 FPU registers #1524
2022-01-04 21:12:12 +01:00
47097b55b7
Fix #1520
2022-01-04 21:01:20 +01:00
039cd50187
unicorn_arm: add reg_read/write operations for FPSCR and FPSID
2022-01-04 11:36:06 +01:00
e55b76f057
s390x: cleanup & re-enable some skey code
2021-12-31 10:05:05 +08:00
1a0f0d0768
s390x: remove some unused fields in S390CPU
2021-12-31 09:48:16 +08:00
a06563ecdd
Fix memory leak
2021-12-31 00:24:18 +01:00
3b667338cf
Fix s390x warnings
2021-12-31 00:10:50 +01:00
fa3fb82c9c
s390x: fix warning on commented code
2021-12-30 17:17:49 +08:00
085ee07c73
No more hard-coded cpu models
2021-12-30 01:05:10 +01:00
fdbd743c21
Remove hard-coded cpu model
2021-12-30 00:54:55 +01:00
a72cbda6de
Initialize empty structs explictly to build on MSVC
2021-12-30 00:51:07 +01:00
03f9dd8b61
Expand case ranges to build on MSVC
2021-12-30 00:42:13 +01:00
dc402d78ec
Ignore QEMU_BUILD_BUG_MSG on MSVC
2021-12-30 00:28:24 +01:00
ab4ef2e1de
Fix MSVC build and remove warning about unused functions
2021-12-30 00:26:25 +01:00
298795a9f8
Fix build on MSVC
2021-12-29 23:18:49 +01:00
034a1aa5f2
Make s390x stopping mechanism work
2021-12-27 23:48:20 +01:00
a38151bf77
Make s390x skey work
2021-12-27 23:19:17 +01:00
e977f81813
Make s390x build
2021-12-26 23:09:25 +01:00
faa689c0f0
Merge systemz to the latest uc2 codebase
2021-12-26 22:58:32 +01:00
cddc9cf2ed
Fix arm post init
2021-12-25 00:16:51 +01:00
4f73d75ea8
Fix #1500
2021-12-23 21:46:27 +01:00
ef6f8a2427
Fix x86 CPUID
2021-12-22 23:39:41 +01:00
7bb756249a
Better design of cpuid instruction hook
2021-12-22 20:36:56 +01:00
63a445cbba
fxsave / fxsave64 should store the floating point instruction pointer (fpip) ( #1467 )
...
* fxsave / fxsave64 should store the floating point instruction pointer (fpip)
- fxsave / fxsave64 happen to be used as GetPC code in exploits
* unit tests for the storage of FPIP in fxsave (x86) and fxsave64 (x64)
2021-12-13 08:40:32 +08:00
09b0c66f11
move all static vars in translate.c to tcg.h
2021-12-07 04:53:32 +08:00
4059906e78
Bug fix for LUI instruction (MIPS)
2021-12-06 19:15:00 +03:00
b042a6a01d
add missing files
2021-12-06 04:28:13 +08:00
97b92d8861
initial systemz support
2021-12-06 04:19:37 +08:00
8a0ca8715e
Fix SR read/write and a test
2021-12-04 23:22:28 +01:00
d204dc6374
Added SR to M68K reg_read and reg_write ( #1507 )
2021-12-02 14:12:49 +08:00
221cde18df
Write CPSR as it is initiated from instructions to allow regs switch
2021-11-24 17:10:51 +01:00
78e0ddbc4d
Fix mmio unmap
2021-11-24 00:18:19 +01:00
c733bbada3
Fix wrong offset used in split_region
2021-11-23 23:22:53 +01:00
c1c5f72918
Fix the sizemask for inline hooking
2021-11-23 21:18:21 +01:00
7a1de17f37
Fix UC_HOOK_EDGE_GENERATED to work with indirect jump
...
For an indirect jump (lookup_tb_ptr), last_tb would be NULL
2021-11-23 00:25:55 +01:00
083ccf160b
Use fprintf
2021-11-22 21:22:21 +01:00
87a391d549
Inline uc_tracecode when there is only exactly one hook
2021-11-21 16:44:39 +01:00
c1106b811b
Fix a memory leak in mmio
2021-11-16 22:44:03 +01:00
fc467edbc6
Fix 32bit target getting wrong offset for mmio
2021-11-16 22:40:57 +01:00
247ffbe0e8
Support nested uc_emu_start calls
2021-11-16 21:07:03 +01:00
43c643d4af
Fix #1488
2021-11-16 09:41:21 +01:00
7e244f87b4
Fix UC_HOOK_EDGE_GENERATED implementation
2021-11-11 22:15:15 +01:00
23ef5da491
Merge pull request #1481 from bet4it/cp15
...
Restore cp15 registers
2021-11-09 16:50:31 +01:00
acaed986b5
Restore cp15 registers
2021-11-09 13:13:08 +08:00
640251e1aa
Leave out size parameter in callback
2021-11-09 00:21:34 +01:00
2f61592ff9
Fix uc_mem_protect
2021-11-07 20:37:58 +01:00
Ilya Leoshkevich