It saves one insn to load the address of TB+4 instead of TB.
Adjust all of the indexing to match.
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Direct branch patching was disabled when using TCG_REG_TB in commit
736a1588c1 ("tcg/ppc: Fix race in goto_tb implementation").
The issue with direct branch patching with TCG_REG_TB is the lack of
synchronization between the new TCG_REG_TB being established and the
direct branch being patched in.
If each translation block is responsible for establishing its own
TCG_REG_TB then there can be no synchronization issue.
Make each translation block begin by setting up its own TCG_REG_TB.
Use the preferred 'bcl 20,31,$+4' sequence.
Signed-off-by: Jordan Niethe <jniethe5@gmail.com>
[rth: Split out tcg_out_tb_start, power9 addpcis]
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Signed-off-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Message-ID: <20231004090629.37473-6-philmd@linaro.org>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
This hook may emit code at the beginning of the TB.
Suggested-by: Jordan Niethe <jniethe5@gmail.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Pass vece to tcg_target_const_match() to allow correct interpretation of
const args of vector ops.
Signed-off-by: Jiajie Chen <c@jia.je>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Message-Id: <20230908022302.180442-4-c@jia.je>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
The SETBC family of instructions requires exactly two insns for
all comparisions, saving 0-3 insns per (neg)setcond.
Tested-by: Nicholas Piggin <npiggin@gmail.com>
Reviewed-by: Daniel Henrique Barboza <danielhb413@gmail.com>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
In the general case we simply negate. However with isel we
may load -1 instead of 1 with no extra effort.
Consolidate EQ0 and NE0 logic. Replace the NE0 zero-extension
with inversion+negation of EQ0, which is never worse and may
eliminate one insn. Provide a special case for -EQ0.
Reviewed-by: Daniel Henrique Barboza <danielhb413@gmail.com>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Commit 20b6643324 ("tcg/ppc: Reorg goto_tb implementation") modified
goto_tb to ensure only a single instruction was patched to prevent
incorrect behavior if a thread was in the middle of multiple
instructions when they were replaced. However this introduced a race
between loading the jmp target into TCG_REG_TB and patching and
executing the direct branch.
The relevant part of the goto_tb implementation:
ld TCG_REG_TB, TARGET_ADDR_LOCATION(TCG_REG_TB)
patch_location:
mtctr TCG_REG_TB
bctr
tb_target_set_jmp_target() will replace 'patch_location' with a direct
branch if the target is in range. The direct branch now relies on
TCG_REG_TB being set up correctly by the ld. Prior to this commit
multiple instructions were patched in for the direct branch case; these
instructions would initialize TCG_REG_TB to the same value as the branch
target.
Imagine the following sequence:
1) Thread A is executing the goto_tb sequence and loads the jmp
target into TCG_REG_TB.
2) Thread B updates the jmp target address and calls
tb_target_set_jmp_target(). This patches a new direct branch into the
goto_tb sequence.
3) Thread A executes the newly patched direct branch. The value in
TCG_REG_TB still contains the old jmp target.
TCG_REG_TB MUST contain the translation block's tc.ptr. Execution will
eventually crash after performing memory accesses generated from a
faulty value in TCG_REG_TB.
This presents as segfaults or illegal instruction exceptions.
Do not revert commit 20b6643324 as it did fix a different race
condition. Instead remove the direct branch optimization and always use
indirect branches.
The direct branch optimization can be re-added later with a race free
sequence.
Fixes: 20b6643324 ("tcg/ppc: Reorg goto_tb implementation")
Resolves: https://gitlab.com/qemu-project/qemu/-/issues/1726
Reported-by: Anushree Mathur <anushree.mathur@linux.vnet.ibm.com>
Tested-by: Anushree Mathur <anushree.mathur@linux.vnet.ibm.com>
Tested-by: Michael Tokarev <mjt@tls.msk.ru>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Co-developed-by: Benjamin Gray <bgray@linux.ibm.com>
Signed-off-by: Jordan Niethe <jniethe5@gmail.com>
Signed-off-by: Benjamin Gray <bgray@linux.ibm.com>
Message-Id: <20230717093001.13167-1-jniethe5@gmail.com>
Move the code from tcg/. Fix a bug in that PPC_FEATURE2_ARCH_3_10
is actually spelled PPC_FEATURE2_ARCH_3_1.
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Reviewed-by: Daniel Henrique Barboza <danielhb413@gmail.com>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Restructure the ifdef ladder, separating 64-bit from 32-bit,
and ensure _CALL_AIX is set for ELF v1. Fixes the build for
ppc64 big-endian host with clang.
Reviewed-by: Daniel Henrique Barboza <danielhb413@gmail.com>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Disconnect the layout of ArchCPU from TCG compilation.
Pass the relative offset of 'env' and 'neg.tlb.f' as a parameter.
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
This makes CPUTLBEntry agnostic to the address size of the guest.
When 32-bit addresses are in effect, we can simply read the low
32 bits of the 64-bit field. Similarly when we need to update
the field for setting TLB_NOTDIRTY.
For TCG backends that could in theory be big-endian, but in
practice are not (arm, loongarch, riscv), use QEMU_BUILD_BUG_ON
to document and ensure this is not accidentally missed.
For s390x, which is always big-endian, use HOST_BIG_ENDIAN anyway,
to document the reason for the adjustment.
For sparc64 and ppc64, always perform a 64-bit load, and rely on
the following 32-bit comparison to ignore the high bits.
Rearrange mips and ppc if ladders for clarity.
Reviewed-by: Anton Johansson <anjo@rev.ng>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
All uses replaced with TCGContext.addr_type.
Reviewed-by: Anton Johansson <anjo@rev.ng>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Use LQ/STQ with ISA v2.07, and 16-byte atomicity is required.
Note that these instructions do not require 16-byte alignment.
Reviewed-by: Daniel Henrique Barboza <danielhb413@gmail.com>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Disconnect guest page size from TCG compilation.
While this could be done via exec/target_page.h, we want to cache
the value across multiple memory access operations, so we might
as well initialize this early.
The changes within tcg/ are entirely mechanical:
sed -i s/TARGET_PAGE_BITS/s->page_bits/g
sed -i s/TARGET_PAGE_MASK/s->page_mask/g
Reviewed-by: Anton Johansson <anjo@rev.ng>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
For 32-bit hosts, we cannot simply rely on TCGContext.addr_bits,
as we need one or two host registers to represent the guest address.
Create the new opcodes and update all users. Since we have not
yet eliminated TARGET_LONG_BITS, only one of the two opcodes will
ever be used, so we can get away with treating them the same in
the backends.
Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Replace the unparameterized TCG_TARGET_HAS_MEMORY_BSWAP macro
with a function with a memop argument.
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Instead of using helper_unaligned_{ld,st}, use the full load/store helpers.
This will allow the fast path to increase alignment to implement atomicity
while not immediately raising an alignment exception.
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
With the current structure of cputlb.c, there is no difference
between the little-endian and big-endian entry points, aside
from the assert. Unify the pairs of functions.
Hoist the qemu_{ld,st}_helpers arrays to tcg.c.
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Never used since its introduction.
Fixes: 3d582c6179 ("tcg-ppc64: Rearrange integer constant constraints")
Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
The softmmu tlb uses TCG_REG_{TMP1,TMP2,R0}, not any of the normally
available registers. Now that we handle overlap betwen inputs and
helper arguments, we can allow any allocatable reg.
Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Daniel Henrique Barboza <danielhb413@gmail.com>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Allocate TCG_REG_TMP2. Use R0, TMP1, TMP2 instead of any of
the normally allocated registers for the tlb load.
Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Daniel Henrique Barboza <danielhb413@gmail.com>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Use tcg_out_ld_helper_args, tcg_out_ld_helper_ret,
and tcg_out_st_helper_args.
Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Daniel Henrique Barboza <danielhb413@gmail.com>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Merge tcg_out_tlb_load, add_qemu_ldst_label, tcg_out_test_alignment,
and some code that lived in both tcg_out_qemu_ld and tcg_out_qemu_st
into one function that returns HostAddress and TCGLabelQemuLdst structures.
Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Collect the parts of the host address into a struct.
Reorg tcg_out_qemu_{ld,st} to use it.
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Interpret the variable argument placement in the caller. Pass data_type
instead of is64 -- there are several places where we already convert back
from bool to type. Clean things up by using type throughout.
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Reviewed-by: Daniel Henrique Barboza <danielhb413@gmail.com>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
We will want a backend interface for register swapping.
This is only properly defined for x86; all others get a
stub version that always indicates failure.
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
This is common code in most qemu_{ld,st} slow paths, extending the
input value for the store helper data argument or extending the
return value from the load helper.
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
We will need a backend interface for type truncation. For those backends
that did not enable TCG_TARGET_HAS_extrl_i64_i32, use tcg_out_mov.
Use it in tcg_reg_alloc_op in the meantime.
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
We will need a backend interface for type extension with zero.
Use it in tcg_reg_alloc_op in the meantime.
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
We will need a backend interface for type extension with sign.
Use it in tcg_reg_alloc_op in the meantime.
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
We will need a backend interface for performing 32-bit zero-extend.
Use it in tcg_reg_alloc_op in the meantime.
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
We will need a backend interface for performing 32-bit sign-extend.
Use it in tcg_reg_alloc_op in the meantime.
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
We will need a backend interface for performing 16-bit zero-extend.
Use it in tcg_reg_alloc_op in the meantime.
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
We will need a backend interface for performing 16-bit sign-extend.
Use it in tcg_reg_alloc_op in the meantime.
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
We will need a backend interface for performing 8-bit zero-extend.
Use it in tcg_reg_alloc_op in the meantime.
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
We will need a backend interface for performing 8-bit sign-extend.
Use it in tcg_reg_alloc_op in the meantime.
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
For both _CALL_SYSV and _CALL_DARWIN, return is by reference,
not in 4 integer registers. For _CALL_SYSV, argument is also
by reference.
This error resulted in
$ ./qemu-system-i386 -nographic
qemu-system-i386: tcg/ppc/tcg-target.c.inc:185: \
tcg_target_call_oarg_reg: Assertion `slot >= 0 && slot <= 1' failed.
Fixes: 5427a9a760 ("tcg: Add TCG_TARGET_CALL_{RET,ARG}_I128")
Tested-by: Mark Cave-Ayland <mark.cave-ayland@ilande.co.uk>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Fill in the parameters for the host ABI for Int128 for
those backends which require no extra modification.
Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Daniel Henrique Barboza <danielhb413@gmail.com>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Replace the flat array tcg_target_call_oarg_regs[] with
a function call including the TCGCallReturnKind.
Extend the set of registers for ARM to r0-r3 to match the ABI:
https://github.com/ARM-software/abi-aa/blob/main/aapcs32/aapcs32.rst#result-return
Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Daniel Henrique Barboza <danielhb413@gmail.com>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Implement the function for arm, i386, and s390x, which will use it.
Add stubs for all other backends.
Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Daniel Henrique Barboza <danielhb413@gmail.com>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
The old ppc64 implementation replaces 2 or 4 insns, which leaves a race
condition in which a thread could be stopped at a PC in the middle of
the sequence, and when restarted does not see the complete address
computation and branches to nowhere.
The new implemetation replaces only one insn, swapping between
b <dest>
and
mtctr r31
falling through to a general-case indirect branch.
Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Replace 'tc_ptr' and 'addr' with 'tb' and 'n'.
Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
The INDEX_op_goto_tb opcode needs no register allocation.
Split out a dedicated helper function for it.
Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Similar to the existing set_jmp_reset_offset. Move any assert for
TCG_TARGET_HAS_direct_jump into the new function (which now cannot
be build-time). Will be unused if TCG_TARGET_HAS_direct_jump is
constant 0, but we can't test for constant in the preprocessor,
so just mark it G_GNUC_UNUSED.
Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
