Once we form a combined qemu_st_i32 opcode, we won't be able to
have separate constraints based on size. This one is fairly easy
to work around, since eax is available as a scratch register.
When storing variable data, this tends to merely exchange one mov
for another. E.g.
-: mov %esi,%ecx
...
-: mov %cl,(%edx)
+: mov %esi,%eax
+: mov %al,(%edx)
Where we do have a regression is when storing constant data, in which
we may load the constant into edi, when only ecx/ebx ought to be used.
The proper way to recover this regression is to allow constants as
arguments to qemu_st_i32, so that we never load the constant data into
a register at all, must less the wrong register. TBD.
Signed-off-by: Richard Henderson <rth@twiddle.net>
Pass two TCGReg to tcg_out_tlb_load, rather than idx+args.
Move ldst_optimization routines just below tcg_out_tlb_load to avoid
the need for forward declarations.
Use TCGReg enum in preference to int where apprpriate.
Signed-off-by: Richard Henderson <rth@twiddle.net>
Step three in the transition: helpers not tied to the target
"default" endianness. To be used when the guest uses a memory
operation with non-default endianness.
Signed-off-by: Richard Henderson <rth@twiddle.net>
Step two in the transition, adding the new ldst opcodes. Keep the old
opcodes around until all backends support the new opcodes.
Signed-off-by: Richard Henderson <rth@twiddle.net>
This is a no-op backend data implementation, for those targets that
are not currently using the load/store optimization path.
This is prepatory to always requiring these functions in all backends.
Signed-off-by: Richard Henderson <rth@twiddle.net>
A minimal update to use the new helpers with the return address argument.
Tested-by: Claudio Fontana <claudio.fontana@linaro.org>
Reviewed-by: Claudio Fontana <claudio.fontana@linaro.org>
Signed-off-by: Richard Henderson <rth@twiddle.net>
For the few targets that actually use these, we'd not report
them symbolicly in the tcg opcode logs.
Signed-off-by: Richard Henderson <rth@twiddle.net>
One call inside of a loop to tcg_register_helper instead of hundreds
of sequential calls.
Presumably more icache and branch prediction friendly; resulting binary
size mostly unchanged on x86_64, as we're trading 32-bit rip-relative
references in .text for full 64-bit pointers in .rodata.
Signed-off-by: Richard Henderson <rth@twiddle.net>
Slightly changes the interface, in that we now return name
instead of a TCGHelperInfo structure, which goes away.
Reviewed-by: Stefan Weil <sw@weilnetz.de>
Signed-off-by: Richard Henderson <rth@twiddle.net>
# By Richard Henderson
# Via Richard Henderson
* rth/tcg-arm-pull:
tcg-arm: Move the tlb addend load earlier
tcg-arm: Remove restriction on qemu_ld output register
tcg-arm: Return register containing tlb addend
tcg-arm: Move load of tlb addend into tcg_out_tlb_read
tcg-arm: Use QEMU_BUILD_BUG_ON to verify constraints on tlb
tcg-arm: Use strd for tcg_out_arg_reg64
tcg-arm: Rearrange slow-path qemu_ld/st
tcg-arm: Use ldrd/strd for appropriate qemu_ld/st64
Message-id: 1380663109-14434-1-git-send-email-rth@twiddle.net
Signed-off-by: Anthony Liguori <anthony@codemonkey.ws>
# By Stefan Weil
# Via Stefan Weil
* sweil/tci:
misc: Use new rotate functions
bitops: Add rotate functions (rol8, ror8, ...)
tci: Add implementation of rotl_i64, rotr_i64
Message-id: 1380137693-3729-1-git-send-email-sw@weilnetz.de
Signed-off-by: Anthony Liguori <anthony@codemonkey.ws>
There are free scheduling slots between the sequence of
comparison instructions. This requires changing the
register in use to avoid conflict with those compares.
Signed-off-by: Richard Henderson <rth@twiddle.net>
The main intent of the patch is to allow the tlb addend register
to be changed, without tying that change to the constraint. But
the most common side-effect seems to be to enable usage of ldrd
with the r0,r1 pair.
Signed-off-by: Richard Henderson <rth@twiddle.net>
This allows us to make more intelligent decisions about the relative
offsets of the tlb comparator and the addend, avoiding any need of
writeback addressing.
Signed-off-by: Richard Henderson <rth@twiddle.net>
One of the two constraints we already checked via #if, but
the tlb offset distance was only checked at runtime.
Signed-off-by: Richard Henderson <rth@twiddle.net>
Use the new helper_ret_*_mmu routines. Use a conditional call
to arrange for a tail-call from the store path, and to load the
return address for the helper for the load path.
Signed-off-by: Richard Henderson <rth@twiddle.net>
It is used by qemu-ppc64 when running Debian's busybox-static.
Cc: qemu-stable <qemu-stable@nongnu.org>
Signed-off-by: Stefan Weil <sw@weilnetz.de>
Reviewed-by: Richard Henderson <rth@twiddle.net>
Less conditional compilation. Merge an add insn with the indexed
memory load insn. Load the tlb addend earlier. Avoid the address
update memory form.
Fix a bug in not allowing large enough tlb offsets for some guests.
Signed-off-by: Richard Henderson <rth@twiddle.net>
Previously we'd only handle 16-bit offsets from memory operand without falling
back to indexed, but it's easy to use ADDIS to handle full 32-bit offsets.
This also lets us unify code that existed inline in tcg_out_op for handling
addition of large constants.
The new R2 temporary was marked reserved for the AIX calling convention, but
the register really is call-clobbered and since tcg generated code has no use
for a TOC, it's available for use.
Signed-off-by: Richard Henderson <rth@twiddle.net>
Remove conditionalization from tcg_target_reg_alloc_order, relying on
reserved_regs to prevent register allocation that shouldn't happen.
So R11 is now present in reg_alloc_order for __APPLE__, but also now
reserved.
Sort reg_alloc_order into call-saved, call-clobbered, and parameters.
This reduces the effect of values getting spilled and reloaded before
function calls.
Whether or not it is reserved, R2 (TOC) is always call-clobbered.
Signed-off-by: Richard Henderson <rth@twiddle.net>
While these are rare from code that's been through the optimizer,
it's not uncommon within the tcg backend.
Signed-off-by: Richard Henderson <rth@twiddle.net>
Instead of bare N, for clarity. The only (intentional) exception made
is for insns that encode R|0, i.e. when R0 encoded into the insn is
interpreted as zero not the contents of the register.
Signed-off-by: Richard Henderson <rth@twiddle.net>
The fix is that sparc has so many mmu modes that the last one overflowed
the 16-bit signed offset we assumed would fit. Handle this, and check
the new assumption at compile time.
Load the tlb addend earlier for the fast path.
Remove the explicit address + addend and make use of index addressing.
Adjust constraints for qemu_ld64 such that we don't clobber the address
register or tlb addend before loading both values.
Signed-off-by: Richard Henderson <rth@twiddle.net>
Coding style fixes. Use TCGReg enumeration values instead of raw
numbers. Don't needlessly pull the whole TCGLabelQemuLdst struct
into local variables. Less conditional compilation.
No functional changes.
Signed-off-by: Richard Henderson <rth@twiddle.net>
While these are rare from code that's been through the optimizer,
it's not uncommon within the tcg backend.
Signed-off-by: Richard Henderson <rth@twiddle.net>
These use a 32-bit load-of-immediate to save a mflr+addi+mtlr sequence.
Tested with a Windows 98 guest (pretty much the most recent thing I
could run on my PPC machine) and kvm-unit-tests's sieve.flat. The
speed up for sieve.flat is as high as 10% for qemu-system-i386, 25%
(no kidding) for qemu-system-x86_64 on my PowerBook G4.
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Richard Henderson <rth@twiddle.net>
For the AIX ABI, the function pointer and small area pointer need
to be loaded in the trampoline. The trampoline instead is called
with a normal BL instruction.
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Richard Henderson <rth@twiddle.net>
error: suggest parentheses around comparison in operand of ‘&’ [-Werror=parentheses]
Signed-off-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Michael Tokarev <mjt@tls.msk.ru>
# By Stefan Weil (6) and others
# Via Michael Tokarev
* mjt/trivial-patches:
aio / timers: use g_usleep() not sleep()
adlib: sort offsets in portio registration
qmp: fix integer usage in examples
tci: Remove function tcg_out64 (fix broken build)
target-arm: Report unimplemented opcodes (LOG_UNIMP)
pflash_cfi02.c: fix debug macro
configure: Remove unneeded redirections of stderr (pkg-config --exists)
configure: Remove unneeded redirections of stderr (pkg-config --cflags, --libs)
configure: Don't write .pyc files by default (python -B)
curl: qemu_bh_new() can never return NULL
slirp/arp_table.c: Avoid shifting into sign bit of signed integers
configure: disable clang -Wstring-plus-int warning
rdma: silly ipv6 bugfix
misc: Fix some typos in names and comments
slirp: Port redirection option behave differently on Linux and Windows
Message-id: 1378119695-14568-1-git-send-email-mjt@msgid.tls.msk.ru
Signed-off-by: Anthony Liguori <anthony@codemonkey.ws>
* 'tcg-next' of git://github.com/rth7680/qemu: (29 commits)
tcg-i386: Make use of zero-extended memory helper routines
tcg: Introduce zero and sign-extended versions of load helpers
exec: Split softmmu_defs.h
target: Include softmmu_exec.h where forgotten
exec: Rename USUFFIX to LSUFFIX
tcg-i386: Don't perform GETPC adjustment in TCG code
exec: Reorganize the GETRA/GETPC macros
configure: Allow x32 as a host
tcg-i386: Adjust tcg_out_tlb_load for x32
tcg-i386: Use intptr_t appropriately
tcg: Fix jit debug for x32
tcg: Use appropriate types in tcg_reg_alloc_call
tcg: Change tcg_out_ld/st offset to intptr_t
tcg: Change tcg_gen_exit_tb argument to uintptr_t
tcg: Use uintptr_t in TCGHelperInfo
tcg: Change relocation offsets to intptr_t
tcg: Change memory offsets to intptr_t
tcg: Change frame pointer offsets to intptr_t
tcg: Define TCG_ptr properly
tcg: Define TCG_TYPE_PTR properly
...
On MIPS ext8s and ext16s ops are implemented with a dedicated
instruction only on MIPS32R2, otherwise the same kind of implementation
than at TCG level (shift left followed by shift right) is used.
Change that by only implementing the ext8s and ext16s ops on MIPS32R2 so
that optimizations can be done by the optimizer. Use an inline version to
avoid having to test again for MIPS32R2 instructions. Keep the shift
implementation for the ld/st routines.
Reviewed-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
Use an inline version for the bswap16 and bswap32 ops to avoid
testing for MIPS32R2 instructions availability, as these ops are
only available in that case.
Reviewed-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
Now that TCG supports enabling and disabling ops at runtime, it's
possible to detect the available host instructions at runtime, and
enable the corresponding ops accordingly.
Unfortunately it's not easy to probe for available instructions on
MIPS, the information is partially available in /proc/cpuinfo, and
not available in AUXV. This patch therefore probes for the instructions
by trying to execute them and by catching a possible SIGILL signal.
Reviewed-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
For 8 and 16-bit unsigned loads, rely on the zero-extension
from the helper and use a smaller 32-bit move insn.
Reviewed-by: Aurelien Jarno <aurelien@aurel32.net>
Signed-off-by: Richard Henderson <rth@twiddle.net>
The _cmmu helpers can be moved to exec-all.h. The helpers that are
used from TCG will shortly need access to tcg_target_long so move
their declarations into tcg.h.
This requires minor include adjustments to all TCG backends.
Reviewed-by: Aurelien Jarno <aurelien@aurel32.net>
Signed-off-by: Richard Henderson <rth@twiddle.net>
Since we now perform it inside the helper, no need to do it here.
This also lets us perform a tail-call from the store slow path to
the helper.
Reviewed-by: Aurelien Jarno <aurelien@aurel32.net>
Signed-off-by: Richard Henderson <rth@twiddle.net>
There are several hosts for which it would be useful to use the
available 64-bit registers in a 32-bit pointer environment.
Reviewed-by: Aurelien Jarno <aurelien@aurel32.net>
Signed-off-by: Richard Henderson <rth@twiddle.net>
Using these instead of mulu2 and muls2 lets us avoid having to argument
overlap analysis in the backend. Normal register allocation will DTRT.
Reviewed-by: Aurelien Jarno <aurelien@aurel32.net>
Signed-off-by: Richard Henderson <rth@twiddle.net>
With the optimization in tcg_liveness_analysis,
we can avoid the MFLO when it is unused.
Reviewed-by: Aurelien Jarno <aurelien@aurel32.net>
Signed-off-by: Richard Henderson <rth@twiddle.net>
Use them in places where mulu2 and muls2 are used.
Optimize mulx2 with dead low part to mulxh.
Reviewed-by: Aurelien Jarno <aurelien@aurel32.net>
Signed-off-by: Richard Henderson <rth@twiddle.net>
Commit ac26eb69a3 added tcg_out64 to tcg/tcg.c.
tcg/tci/tcg-target.c already had a nearly identical implementation which is
now removed to fix a compiler error.
Signed-off-by: Stefan Weil <sw@weilnetz.de>
Reviewed-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Michael Tokarev <mjt@tls.msk.ru>
Discontinue the jump-around-jump-to-jump scheme, trading it for a single
immediate move instruction. The two extra jumps always consume 7 bytes,
whereas the immediate move is either 5 or 7 bytes depending on where the
code_gen_buffer gets located.
Signed-off-by: Richard Henderson <rth@twiddle.net>
No point in splitting the write into 32-bit pieces.
Reviewed-by: Aurelien Jarno <aurelien@aurel32.net>
Signed-off-by: Richard Henderson <rth@twiddle.net>
Aliasing was forcing s->code_ptr to be re-read after the store.
Keep the pointer in a local variable to help the compiler.
Signed-off-by: Richard Henderson <rth@twiddle.net>
tcg/mips/tcg-target.h defines various operations conditionally depending
upon the isa revision, however these operations are included in
mips_op_defs[] unconditionally resulting in the following runtime errors
if CONFIG_DEBUG_TCG is defined:
Invalid op definition for movcond_i32
Invalid op definition for rotl_i32
Invalid op definition for rotr_i32
Invalid op definition for deposit_i32
Invalid op definition for bswap16_i32
Invalid op definition for bswap32_i32
tcg/tcg.c:1196: tcg fatal error
Fix with ifdefs like the i386 backend does for movcond_i32.
Signed-off-by: James Hogan <james.hogan@imgtec.com>
Cc: Aurelien Jarno <aurelien@aurel32.net>
Cc: Richard Henderson <rth@twiddle.net>
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
The definition of macro BIT in tci/tcg-target.c now conflicts with the
definition of the same macro in includes qemu/bitops.h.
This conflict was triggered by a recent change in the include chain of
tcg.c (probably commit 949fc82314).
Signed-off-by: Stefan Weil <sw@weilnetz.de>
Message-id: 1375216883-23969-1-git-send-email-sw@weilnetz.de
Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
We can check the condition at compile time, rather than run time.
Reviewed-by: Andreas Färber <afaerber@suse.de>
Signed-off-by: Richard Henderson <rth@twiddle.net>
These will necessarily be the same layout for all hosts. This limits
the amount of boilerplate required to implement jit debug for a host.
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Richard Henderson <rth@twiddle.net>
I don't think the debugger actually looks at this for anything,
using the correct .debug_frame contents, but might as well get
it all correct.
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Richard Henderson <rth@twiddle.net>
With this we can generate armv7 insns even when the OS compiles for a
lower common denominator. The macros are arranged so that when we do
compile for a given ISA, all of the runtime checks for that ISA are
optimized away.
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Richard Henderson <rth@twiddle.net>
GCC 4.8 defines a handy __ARM_ARCH symbol that we can use, which
will make us nicely forward compatible with ARMv8 AArch32.
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Richard Henderson <rth@twiddle.net>
As it really controls the availability of a thumb interworking
instruction on armv5t.
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Richard Henderson <rth@twiddle.net>
We can now detect and use divide instructions at runtime, rather than
having to restrict their availability to compile-time.
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Richard Henderson <rth@twiddle.net>
Expand the definition of "not present" to include "should not be present".
This means we can simplify the logic surrounding the generic tcg opcodes
for which the host backend ought not be providing definitions.
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Richard Henderson <rth@twiddle.net>
This allows TCG_TARGET_HAS_* to be a variable rather than a constant,
which allows easier support for differing ISA levels for the host.
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Richard Henderson <rth@twiddle.net>
There are several hosts with only a "div" insn. Remainder is computed
manually from the quotient and inputs. We can do this generically.
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Richard Henderson <rth@twiddle.net>
implement the 12bit scaled unsigned immediate offset
variant of LDR/STR. This improves code size by avoiding
the movi + ldst_r for naturally aligned offsets in range.
Signed-off-by: Claudio Fontana <claudio.fontana@huawei.com>
Reviewed-by: Richard Henderson <rth@twiddle.net>
rotr_i32 calculates the amount to left shift and puts it into a
temporary, but then doesn't use it when doing the shift.
Cc: qemu-stable@nongnu.org
Signed-off-by: Anton Blanchard <anton@samba.org>
Signed-off-by: Richard Henderson <rth@twiddle.net>
add2_i64 was adding the lower double word to the upper double word
of each input. Fix this so we add the lower double words, then the
upper double words with carry propagation.
Cc: qemu-stable@nongnu.org
Signed-off-by: Anton Blanchard <anton@samba.org>
Signed-off-by: Richard Henderson <rth@twiddle.net>
If our input and output is in the same register, bswap64 tries to
undo a rotate of the input. This just ends up rotating the output.
Cc: qemu-stable@nongnu.org
Signed-off-by: Anton Blanchard <anton@samba.org>
Signed-off-by: Richard Henderson <rth@twiddle.net>
The rldcl instruction doesn't have an sh field, so the minor opcode
is shifted 1 bit. We were using the XO30 macro which shifted the
minor opcode 2 bits.
Remove XO30 and add MD30 and MDS30 macros which match the
Power ISA categories.
Cc: qemu-stable@nongnu.org
Signed-off-by: Anton Blanchard <anton@samba.org>
Signed-off-by: Richard Henderson <rth@twiddle.net>
# By Claudio Fontana (9) and others
# Via Peter Maydell
* pmaydell/tcg-aarch64.next:
MAINTAINERS: add tcg/aarch64 maintainer
configure: permit compilation on arm aarch64
tcg/aarch64: implement user mode qemu ld/st
user-exec.c: aarch64 initial implementation of cpu_signal_handler
tcg/aarch64: implement sign/zero extend operations
tcg/aarch64: implement byte swap operations
tcg/aarch64: implement AND/TEST immediate pattern
tcg/aarch64: improve arith shifted regs operations
tcg/aarch64: implement new TCG target for aarch64
include/elf.h: add aarch64 ELF machine and relocs
configure: Drop CONFIG_ATFILE test
linux-user: Drop direct use of openat etc syscalls
linux-user: Allow getdents to be provided by getdents64
Message-id: 1371052645-9006-1-git-send-email-peter.maydell@linaro.org
Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
also put aarch64 in the list of archs that do not need an ldscript.
Signed-off-by: Jani Kokkoken <jani.kokkonen@huawei.com>
Signed-off-by: Claudio Fontana <claudio.fontana@huawei.com>
Reviewed-by: Richard Henderson <rth@twiddle.net>
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Message-id: 51AF40EE.1000104@huawei.com
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
implement the optional sign/zero extend operations with the dedicated
aarch64 instructions.
Signed-off-by: Claudio Fontana <claudio.fontana@huawei.com>
Reviewed-by: Richard Henderson <rth@twiddle.net>
Message-id: 51AC9A58.40502@huawei.com
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
implement the optional byte swap operations with the dedicated
aarch64 instructions.
Signed-off-by: Claudio Fontana <claudio.fontana@huawei.com>
Reviewed-by: Richard Henderson <rth@twiddle.net>
Message-id: 51AC9A33.9050003@huawei.com
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
for arith operations, add SUBS, ANDS, ADDS and add a shift parameter
so that all arith instructions can make use of shifted registers.
Signed-off-by: Claudio Fontana <claudio.fontana@huawei.com>
Reviewed-by: Richard Henderson <rth@twiddle.net>
Message-id: 51AC998B.7070506@huawei.com
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
add preliminary support for TCG target aarch64.
Signed-off-by: Claudio Fontana <claudio.fontana@huawei.com>
Reviewed-by: Richard Henderson <rth@twiddle.net>
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Message-id: 51A5C596.3090108@huawei.com
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
We've got a compile-time check for the condition in exec/cpu-defs.h.
Reviewed-by: Andreas Färber <afaerber@suse.de>
Reviewed-by: liguang <lig.fnst@cn.fujitsu.com>
Signed-off-by: Richard Henderson <rth@twiddle.net>
When setcond2 is rewritten into setcond, the state of the destination
temp should be reset, so that a copy of the previous value is not
used instead of the result.
Reported-by: Michael Tokarev <mjt@tls.msk.ru>
Reviewed-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
Avoid the mini constant pool for armv7, and avoid replicating
the test for pre-v7.
Signed-off-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
Found by inspection, since the effect of the bug was simply to
send all memory ops through the slow path.
Signed-off-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
Move the slow path out of line, as the TODO's mention.
This allows the fast path to be unconditional, which can
speed up the fast path as well, depending on the core.
Signed-off-by: Richard Henderson <rth@twiddle.net>
Work better with branch predition when we have movw+movt,
as the size of the code is the same. Perhaps re-evaluate
when we have a proper constant pool.
Reviewed-by: Aurelien Jarno <aurelien@aurel32.net>
Signed-off-by: Richard Henderson <rth@twiddle.net>
The schedule was fully serial, with no possibility for dual issue.
The old schedule had a minimal issue of 7 cycles; the new schedule
has a minimal issue of 5 cycles.
Signed-off-by: Richard Henderson <rth@twiddle.net>
Share code between qemu_ld and qemu_st to process the tlb.
Reviewed-by: Aurelien Jarno <aurelien@aurel32.net>
Signed-off-by: Richard Henderson <rth@twiddle.net>
Use even more primitive helper functions to avoid lots of duplicated code.
Reviewed-by: Aurelien Jarno <aurelien@aurel32.net>
Signed-off-by: Richard Henderson <rth@twiddle.net>
Make the code more readable by only having one copy of the magic
numbers, swapping registers as needed prior to that. Speed the
compiler by not applying the rd == rn avoidance for v6 or later.
Reviewed-by: Aurelien Jarno <aurelien@aurel32.net>
Signed-off-by: Richard Henderson <rth@twiddle.net>
R12 is call clobbered, while R8 is call saved. This change
gives tcg one more call saved register for real data.
Reviewed-by: Aurelien Jarno <aurelien@aurel32.net>
Signed-off-by: Richard Henderson <rth@twiddle.net>
We get to re-use the _rIN and _rIK subroutines to handle the various
combinations of add vs sub. Fold the << 21 into the opcode enum values
so that we can explicitly add TO_CPSR as desired.
Reviewed-by: Aurelien Jarno <aurelien@aurel32.net>
Signed-off-by: Richard Henderson <rth@twiddle.net>
This greatly improves code generation for addition of small
negative constants.
Reviewed-by: Aurelien Jarno <aurelien@aurel32.net>
Signed-off-by: Richard Henderson <rth@twiddle.net>
This greatly improves the code we can produce for deposit
without armv7 support.
Reviewed-by: Aurelien Jarno <aurelien@aurel32.net>
Signed-off-by: Richard Henderson <rth@twiddle.net>
This makes it easier to verify changes to the code
generating the prologue.
[Aurelien: change the format from %i to %zu]
Reviewed-by: Aurelien Jarno <aurelien@aurel32.net>
Signed-off-by: Richard Henderson <rth@twiddle.net>
We were not allocating TCG_STATIC_CALL_ARGS_SIZE, so this meant that
any helper with more than 4 arguments would clobber the saved regs.
Realizing that we're supposed to have this memory pre-allocated means
we can clean up the tcg_out_arg functions, which were trying to do
more stack allocation.
Allocate stack memory for the TCG temporaries while we're at it.
Signed-off-by: Richard Henderson <rth@twiddle.net>
On 32-bit TCG targets, when emulating deposit_i64 with a mov_i32 +
deposit_i32, care should be taken to not overwrite the low part of
the second argument before the deposit when it is the same the
destination.
This fixes the shld instruction in qemu-system-x86_64, which in turns
fixes booting "system rescue CD version 2.8.0" on this target.
Reported-by: Michael S. Tsirkin <mst@redhat.com>
Reviewed-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
The TCG optimizer does great work when inserting constants, being able
to fold the open-coded deposit expansion to just an AND or an OR. Avoid
a bit the regression caused by having the deposit opcode by expanding
deposit of zero as an AND.
Reviewed-by: Aurelien Jarno <aurelien@aurel32.net>
Signed-off-by: Richard Henderson <rth@twiddle.net>
Glibc 2.16 includes an easy way to get feature bits previously
buried in /proc or the program startup auxiliary vector. Use it.
Reviewed-by: Aurelien Jarno <aurelien@aurel32.net>
Signed-off-by: Richard Henderson <rth@twiddle.net>
There are a few simple special cases that should be handled first.
Break these out to subroutines to avoid code duplication.
Reviewed-by: Aurelien Jarno <aurelien@aurel32.net>
Signed-off-by: Richard Henderson <rth@twiddle.net>
It takes half the cycles to read one CR register instead of all 8.
This is a backward compatible addition to the ISA, so chips prior
to Power 2.00 spec will simply continue to read the entire CR register.
Reviewed-by: Aurelien Jarno <aurelien@aurel32.net>
Signed-off-by: Richard Henderson <rth@twiddle.net>
Nothing else in the call chain ensures that these
constants don't have garbage in the high bits.
Reviewed-by: Aurelien Jarno <aurelien@aurel32.net>
Signed-off-by: Richard Henderson <rth@twiddle.net>
The optimization/bug being fixed is that tcg_out_cmp was not applying the
right type to loading a constant, in the case it can't be implemented
directly. Rather than recomputing the TCGType enum from the arch64 bool,
pass around the original TCGType throughout.
Reviewed-by: Aurelien Jarno <aurelien@aurel32.net>
Signed-off-by: Richard Henderson <rth@twiddle.net>
The mul_i32 pattern was loading non-16-bit constants into a register,
when we can get the middle-end to do that for us. The mul_i64 pattern
was not considering that MULLI takes 64-bit inputs.
Reviewed-by: Aurelien Jarno <aurelien@aurel32.net>
Signed-off-by: Richard Henderson <rth@twiddle.net>
Since we have special code to handle and/or/xor with a constant,
apply the same to andc/orc/eqv with a constant.
Reviewed-by: Aurelien Jarno <aurelien@aurel32.net>
Signed-off-by: Richard Henderson <rth@twiddle.net>
Using a table to look up insns of the right width and sign.
Include support for the Power 2.06 LDBRX and STDBRX insns.
Reviewed-by: Aurelien Jarno <aurelien@aurel32.net>
Signed-off-by: Richard Henderson <rth@twiddle.net>
Handle constants in common code; we'll want to reuse that later.
Reviewed-by: Aurelien Jarno <aurelien@aurel32.net>
Signed-off-by: Richard Henderson <rth@twiddle.net>
Improve constant addition -- previously we'd emit useless addi with 0.
Use new constraints to force the driver to pull full 64-bit constants
into a register.
Reviewed-by: Aurelien Jarno <aurelien@aurel32.net>
Signed-off-by: Richard Henderson <rth@twiddle.net>
We'll need a zero, and Z makes more sense for that. Make sure we
have a full compliment of signed and unsigned 16 and 32-bit tests.
Reviewed-by: Aurelien Jarno <aurelien@aurel32.net>
Signed-off-by: Richard Henderson <rth@twiddle.net>
The test for using movi32 was sub-optimal for TCG_TYPE_I32, comparing
a signed 32-bit quantity against an unsigned 32-bit quantity.
When possible, use addi+oris for 32-bit unsigned constants. Otherwise,
standardize on addi+oris+ori instead of addis+ori+rldicl.
Reviewed-by: Aurelien Jarno <aurelien@aurel32.net>
Signed-off-by: Richard Henderson <rth@twiddle.net>
We weren't ignoring the high 32 bits during a NE comparison.
Reviewed-by: Aurelien Jarno <aurelien@aurel32.net>
Signed-off-by: Richard Henderson <rth@twiddle.net>
* 'tci' of git://qemu.weilnetz.de/qemu:
tci: Make tcg temporaries local to tcg_qemu_tb_exec
tci: Delete unused tb_ret_addr
tci: Avoid code before declarations
tci: Use a local variable for env
tci: Use 32-bit signed offsets to loads/stores
We're moving away from the temporaries stored in env. Make sure we can
differentiate between temp stores and possibly bogus stores for extra
call arguments. Move TCG_AREG0 and TCG_REG_CALL_STACK out of the way
of the parameter passing registers.
Signed-off-by: Richard Henderson <rth@twiddle.net>
Signed-off by: Stefan Weil <sw@weilnetz.de>
Since the change to tcg_exit_req, the first insn of every TB is
a load with a negative offset from env.
Signed-off-by: Richard Henderson <rth@twiddle.net>
Signed-off by: Stefan Weil <sw@weilnetz.de>
When the TCG condition codes were re-organized last year,
we failed to update all of the "old-style" tests for unsigned.
Signed-off-by: Richard Henderson <rth@twiddle.net>
This can save one insn, if the constant has any bits in 32-63 set,
but no bits in 21-31 set. It never results in more insns.
Signed-off-by: Richard Henderson <rth@twiddle.net>
Since we're always in 64-bit mode, load address performs a full
64-bit add. Use that for 3-address addition, as well as for
larger constant addends when we lack extended-immediates facility.
Signed-off-by: Richard Henderson <rth@twiddle.net>
Since we have a free temporary and can always just load the constant, we
ought to do so, rather than spending the same effort constraining the const.
Signed-off-by: Richard Henderson <rth@twiddle.net>
We only support 64-bit code generation for s390x.
Don't clutter the code with ifdefs that suggest otherwise.
Signed-off-by: Richard Henderson <rth@twiddle.net>
Set TCG_TARGET_CALL_STACK_OFFSET properly for the abi. Allocate the
standard TCG_STATIC_CALL_ARGS_SIZE. And while we're at it, allocate
space for CPU_TEMP_BUF_NLONGS.
Signed-off-by: Richard Henderson <rth@twiddle.net>
In TCG, "target" means the host architecture for which TCG generates
the code. Using "guest" rather than "target" to make the document more
consistent.
Signed-off-by: Chen Wei-Ren <chenwj@iis.sinica.edu.tw>
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Fix some of the nasty TCG race conditions and crashes by implementing
cpu_exit() as setting a flag which is checked at the start of each TB.
This avoids crashes if a thread or signal handler calls cpu_exit()
while the execution thread is itself modifying the TB graph (which
may happen in system emulation mode as well as in linux-user mode
with a multithreaded guest binary).
This fixes the crashes seen in LP:668799; however there are another
class of crashes described in LP:1098729 which stem from the fact
that in linux-user with a multithreaded guest all threads will
use and modify the same global TCG date structures (including the
generated code buffer) without any kind of locking. This means that
multithreaded guest binaries are still in the "unsupported"
category.
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Blue Swirl <blauwirbel@gmail.com>
Document tcg_qemu_tb_exec(). In particular, its return value is a
combination of a pointer to the next translation block and some
extra information in the low two bits. Provide some #defines for
the values passed in these bits to improve code clarity.
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Blue Swirl <blauwirbel@gmail.com>
Switch the default for qemu_log logging output from "/tmp/qemu.log"
to stderr. This is an incompatible change in some sense, but logging
is mostly used for debugging purposes so it shouldn't affect production
use. The previous behaviour can be obtained by adding "-D /tmp/qemu.log"
to the command line.
This change requires us to:
* update all the documentation/help text (we take the opportunity
to smooth out minor inconsistencies between the phrasing in
linux-user/bsd-user/system help messages)
* make linux-user and bsd-user defer to qemu-log for the default
logging destination rather than overriding it themselves
* ensure that all logfile closing is done via qemu_log_close()
and that that function doesn't close stderr
as well as the obvious change to the behaviour of do_qemu_set_log()
when no logfile name has been specified.
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Markus Armbruster <armbru@redhat.com>
Message-id: 1361901160-28729-1-git-send-email-peter.maydell@linaro.org
Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
We even had the encoding of smull already handy...
Cc: Andrzej Zaborowski <balrogg@gmail.com>
Signed-off-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Blue Swirl <blauwirbel@gmail.com>
We're going to have use for this shortly in implementing other helpers.
Signed-off-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Blue Swirl <blauwirbel@gmail.com>
Commit 0b0d3320db (TCG: Final globals
clean-up) moved code_gen_prologue but forgot to update ppc code.
This broke the build on 32-bit ppc. ppc64 is unaffected.
Cc: Evgeny Voevodin <evgenyvoevodin@gmail.com>
Cc: Blue Swirl <blauwirbel@gmail.com>
Signed-off-by: Andreas Färber <andreas.faerber@web.de>
Signed-off-by: Blue Swirl <blauwirbel@gmail.com>
Rename the public-facing function cpu_set_log to qemu_set_log. This
requires us to rename the internal-only qemu_set_log() to
do_qemu_set_log().
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Andreas Färber <afaerber@suse.de>
Signed-off-by: Blue Swirl <blauwirbel@gmail.com>
It's worth to clean-up translation blocks variables and move them
into one context as was suggested by Swirl.
Also if we use this context directly inside tcg_ctx, then it
speeds up code generation a bit.
Signed-off-by: Evgeny Voevodin <evgenyvoevodin@gmail.com>
Signed-off-by: Blue Swirl <blauwirbel@gmail.com>
Silence a (legitimate) complaint about missing parentheses:
tcg/arm/tcg-target.c: In function ‘tcg_out_qemu_ld’:
tcg/arm/tcg-target.c:1148:5: error: suggest parentheses around
comparison in operand of ‘&’ [-Werror=parentheses]
tcg/arm/tcg-target.c: In function ‘tcg_out_qemu_st’:
tcg/arm/tcg-target.c:1357:5: error: suggest parentheses around
comparison in operand of ‘&’ [-Werror=parentheses]
which meant that we would mistakenly always assert if running
a QEMU built with debug enabled on ARM.
Signed-off-by: Peter Maydell <peter.maydelL@linaro.org>
Signed-off-by: Blue Swirl <blauwirbel@gmail.com>
This adds two optimizations using the non-zero bit mask. In some cases
involving shifts or ANDs the value can become zero, and can thus be
optimized to a move of zero. Second, useless zero-extension or an
AND with constant can be detected that would only zero bits that are
already zero.
The main advantage of this optimization is that it turns zero-extensions
into moves, thus enabling much better copy propagation (around 1% code
reduction). Here is for example a "test $0xff0000,%ecx + je" before
optimization:
mov_i64 tmp0,rcx
movi_i64 tmp1,$0xff0000
discard cc_src
and_i64 cc_dst,tmp0,tmp1
movi_i32 cc_op,$0x1c
ext32u_i64 tmp0,cc_dst
movi_i64 tmp12,$0x0
brcond_i64 tmp0,tmp12,eq,$0x0
and after (without patch on the left, with on the right):
movi_i64 tmp1,$0xff0000 movi_i64 tmp1,$0xff0000
discard cc_src discard cc_src
and_i64 cc_dst,rcx,tmp1 and_i64 cc_dst,rcx,tmp1
movi_i32 cc_op,$0x1c movi_i32 cc_op,$0x1c
ext32u_i64 tmp0,cc_dst
movi_i64 tmp12,$0x0 movi_i64 tmp12,$0x0
brcond_i64 tmp0,tmp12,eq,$0x0 brcond_i64 cc_dst,tmp12,eq,$0x0
Other similar cases: "test %eax, %eax + jne" where eax is already 32-bit
(after optimization, without patch on the left, with on the right):
discard cc_src discard cc_src
mov_i64 cc_dst,rax mov_i64 cc_dst,rax
movi_i32 cc_op,$0x1c movi_i32 cc_op,$0x1c
ext32u_i64 tmp0,cc_dst
movi_i64 tmp12,$0x0 movi_i64 tmp12,$0x0
brcond_i64 tmp0,tmp12,ne,$0x0 brcond_i64 rax,tmp12,ne,$0x0
"test $0x1, %dl + je":
movi_i64 tmp1,$0x1 movi_i64 tmp1,$0x1
discard cc_src discard cc_src
and_i64 cc_dst,rdx,tmp1 and_i64 cc_dst,rdx,tmp1
movi_i32 cc_op,$0x1a movi_i32 cc_op,$0x1a
ext8u_i64 tmp0,cc_dst
movi_i64 tmp12,$0x0 movi_i64 tmp12,$0x0
brcond_i64 tmp0,tmp12,eq,$0x0 brcond_i64 cc_dst,tmp12,eq,$0x0
In some cases TCG even outsmarts GCC. :) Here the input code has
"and $0x2,%eax + movslq %eax,%rbx + test %rbx, %rbx" and the optimizer,
thanks to copy propagation, does the following:
movi_i64 tmp12,$0x2 movi_i64 tmp12,$0x2
and_i64 rax,rax,tmp12 and_i64 rax,rax,tmp12
mov_i64 cc_dst,rax mov_i64 cc_dst,rax
ext32s_i64 tmp0,rax -> nop
mov_i64 rbx,tmp0 -> mov_i64 rbx,cc_dst
and_i64 cc_dst,rbx,rbx -> nop
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Blue Swirl <blauwirbel@gmail.com>
Add a "mask" field to the tcg_temp_info struct. A bit that is zero
in "mask" will always be zero in the corresponding temporary.
Zero bits in the mask can be produced from moves of immediates,
zero-extensions, ANDs with constants, shifts; they can then be
be propagated by logical operations, shifts, sign-extensions,
negations, deposit operations, and conditional moves. Other
operations will just reset the mask to all-ones, i.e. unknown.
[rth: s/target_ulong/tcg_target_ulong/]
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Blue Swirl <blauwirbel@gmail.com>
The next patch will add to the TCG optimizer a field that should be
non-zero in the default case. Thus, replace the memset of the
temps array with a loop. Only the state field has to be up-to-date,
because others are not used except if the state is TCG_TEMP_COPY
or TCG_TEMP_CONST.
[rth: Extracted the loop to a function.]
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Blue Swirl <blauwirbel@gmail.com>
Commit 7f6f0ae5b9 added two assertions.
One of these assertions is not needed:
The pointer ts is never NULL because it is initialized with the
address of an array element.
Reviewed-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Stefan Weil <sw@weilnetz.de>
Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
Existing compile-time detection is spotty at best. Convert
it all to runtime detection instead.
Signed-off-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Blue Swirl <blauwirbel@gmail.com>
In dead_temp, local temps should always be marked as back to memory,
even if they have not been allocated (i.e. they are discared before
cross a basic block).
It fixes the following assertion in target-xtensa:
qemu-system-xtensa: tcg/tcg.c:1665: temp_save: Assertion `s->temps[temp].val_type == 2 || s->temps[temp].fixed_reg' failed.
Aborted
Reported-by: Max Filippov <jcmvbkbc@gmail.com>
Tested-by: Max Filippov <jcmvbkbc@gmail.com>
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
The bswap16 TCG opcode assumes that the high bytes of the temp equal
to 0 before calling it. The ARM backend implementation takes this
assumption to slightly optimize the generated code.
The same implementation is called for implementing the cross-endian
qemu_st16 opcode, where this assumption is not true anymore. One way to
fix that would be to zero the high bytes before calling it. Given the
store instruction just ignore them, it is possible to provide a slightly
more optimized version. With ARMv6+ the rev16 instruction does the work
correctly. For lower ARM versions the patch provides a version which
behaves correctly with non-zero high bytes, but fill them with junk.
Cc: Andrzej Zaborowski <balrogg@gmail.com>
Cc: Peter Maydell <peter.maydell@linaro.org>
Cc: qemu-stable@nongnu.org
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
The TCG arm backend considers likely that the offset to the TLB
entries does not exceed 12 bits for mem_index = 0. In practice this is
not true for at least the MIPS target.
The current patch fixes that by loading the bits 23-12 with a separate
instruction, and using loads with address writeback, independently of
the value of mem_idx. In total this allow a 24-bit offset, which is a
lot more than needed.
Cc: Andrzej Zaborowski <balrogg@gmail.com>
Cc: Peter Maydell <peter.maydell@linaro.org>
Cc: qemu-stable@nongnu.org
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
The operations for INDEX_op_deposit_i32 and INDEX_op_deposit_i64
are now supported and enabled by default.
Signed-off-by: Stefan Weil <sw@weilnetz.de>
Signed-off-by: Blue Swirl <blauwirbel@gmail.com>
mmu access looks something like:
<check tlb>
if miss goto slow_path
<fast path>
done:
...
; end of the TB
slow_path:
<pre process>
mr r3, r27 ; move areg0 to r3
; (r3 holds the first argument for all the PPC32 ABIs)
<call mmu_helper>
b $+8
.long done
<post process>
b done
On ppc32 <call mmu_helper> is:
(SysV and Darwin)
mmu_helper is most likely not within direct branching distance from
the call site, necessitating
a. moving 32 bit offset of mmu_helper into a GPR ; 8 bytes
b. moving GPR to CTR/LR ; 4 bytes
c. (finally) branching to CTR/LR ; 4 bytes
r3 setting - 4 bytes
call - 16 bytes
dummy jump over retaddr - 4 bytes
embedded retaddr - 4 bytes
Total overhead - 28 bytes
(PowerOpen (AIX))
a. moving 32 bit offset of mmu_helper's TOC into a GPR1 ; 8 bytes
b. loading 32 bit function pointer into GPR2 ; 4 bytes
c. moving GPR2 to CTR/LR ; 4 bytes
d. loading 32 bit small area pointer into R2 ; 4 bytes
e. (finally) branching to CTR/LR ; 4 bytes
r3 setting - 4 bytes
call - 24 bytes
dummy jump over retaddr - 4 bytes
embedded retaddr - 4 bytes
Total overhead - 36 bytes
Following is done to trim the code size of slow path sections:
In tcg_target_qemu_prologue trampolines are emitted that look like this:
trampoline:
mfspr r3, LR
addi r3, 4
mtspr LR, r3 ; fixup LR to point over embedded retaddr
mr r3, r27
<jump mmu_helper> ; tail call of sorts
And slow path becomes:
slow_path:
<pre process>
<call trampoline>
.long done
<post process>
b done
call - 4 bytes (trampoline is within code gen buffer
and most likely accessible via
direct branch)
embedded retaddr - 4 bytes
Total overhead - 8 bytes
In the end the icache pressure is decreased by 20/28 bytes at the cost
of an extra jump to trampoline and adjusting LR (to skip over embedded
retaddr) once inside.
Signed-off-by: malc <av1474@comtv.ru>
Add optimized TCG qemu_ld/st generation which locates the code of TLB miss
cases at the end of a block after generating the other IRs.
Currently, this optimization supports only i386 and x86_64 hosts.
Signed-off-by: Yeongkyoon Lee <yeongkyoon.lee@samsung.com>
Signed-off-by: Blue Swirl <blauwirbel@gmail.com>
Commit 9c43b68de6 do not correctly check
for dead outputs when they need to be synced to memory in case of
half-dead operations.
Fix that by applying the same pattern than for the default case.
Tested-by: Stefan Weil <sw@weilnetz.de>
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
MIPS32 and later instruction sets have a multiplication instruction
directly operating on GPRs. It only produces a 32-bit result but
it is exactly what is needed by QEMU.
Reviewed-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
When we allocate a reserved_va for the guest, the kernel will likely
choose an address well above 4G. At which point we must use a pair
of movabsq+addq to form the host address. If we have OS support,
set up a segment register to point to guest_base instead.
Signed-off-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
The current helper flags, TCG_CALL_CONST and TCG_CALL_PURE might be
confusing and doesn't provide enough granularity for some helpers (FP
helpers for example).
This patch changes them into the following helpers flags:
- TCG_CALL_NO_READ_GLOBALS means that the helper does not read globals,
either directly or via an exception. They will not be saved to their
canonical location before calling the helper.
- TCG_CALL_NO_WRITE_GLOBALS means that the helper does not modify any
globals. They will only be saved to their canonical locations before
calling helpers, but they won't be reloaded afterwise.
- TCG_CALL_NO_SIDE_EFFECTS means that the call to the function is
removed if the return value is not used.
It provides convenience flags, to avoid helper definitions longer than
80 characters. It also provides compatibility flags, and updates the
documentation.
Reviewed-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
Operations with side effects (in practice qemu_ld/st ops), only need to
synchronize globals to make sure the CPU state is consistent in case of
exception.
Reviewed-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
Mapping a memory address using a global and accessing it through
ld/st operations is currently broken. As it doesn't make any sense
to do that performance wise, let's forbid that.
Update the TCG documentation, and remove partial support for that.
Reviewed-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
Some branch related ops are marked with TCG_OPF_SIDE_EFFECTS, some other
not. In practice they don't need to, as they are all marked with
TCG_OPF_BB_END, which is handled specifically in all the code.
The call op is marked as TCG_OPF_SIDE_EFFECTS, which might be not true
as there is are specific flags (TCG_CALL_CONST and TCG_CALL_PURE) for
specifying that. On the other hand it always clobber arguments, so mark
it as such even if the call op is handled in a different code path.
Reviewed-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
The liveness analysis ensures that globals and temps are at the correct
state at a basic block end or with an op with side effects. Avoid
looping on all temps, this can be time consuming on targets with a lot
of globals. Keep an assert in debug mode.
Reviewed-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
Start with local temps in TEMP_VAL_MEM state, to make possible a later
check that all the temps are correctly saved back to memory.
Reviewed-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
Always mark dead input arguments as dead, even if the op is at the basic
block end. This will allow to check that all temps are correctly saved.
Reviewed-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
Now that the liveness analysis provides more information, rewrite
tcg_reg_alloc_mov(). This changes the behaviour about propagating
constants and memory accesses. We now take the assumption that once
a value is loaded into a register (from memory or from a constant),
it's better to keep it there than to reload it later. This assumption
is now always almost correct given that we are now sure the
corresponding temp is going to be used later (otherwise it would have
been synchronized and marked as dead already). The assumption is wrong
if one of the op after clobbers some registers including the one
of the holding the temp (this can be avoided by allocating clobbered
registers last, which is what most TCG target do), or in case of lack
of available register.
Reviewed-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
Now that the liveness analysis might mark some output temps as dead, call
temp_dead() if needed.
Reviewed-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
Rework the liveness analysis by tracking temps that need to go back to
memory in addition to dead temps tracking. This allows to mark output
arguments as "need sync", and to synchronize them back to memory as soon
as they are not written anymore. This way even arguments mapping to
globals can be marked as "dead", avoiding moves to a new register when
input and outputs are aliased.
In addition it means that registers are freed as soon as temps are not
used anymore, instead of waiting for a basic block end or an op with side
effects. This reduces register spilling especially on CPUs with few
registers, and spread the mov over all the TB, increasing the
performances on in-order CPUs.
Reviewed-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
Synchronize an output argument when requested by the liveness analysis.
This is needed so that the temp can be declared dead later.
For that, add a new op_sync_args table in which each bit tells if the
corresponding output argument needs to be synchronized with the memory.
Pass it to the tcg_reg_alloc_* functions, and honor this bit. We need to
synchronize the argument before marking it as dead, and we have to make
sure all the infos about the temp are correctly filled.
At the same time change some types from unsigned int to uint16_t when
passing op_dead_args.
Reviewed-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
Add a new function temp_sync() to synchronize the canonical location
of a temp with the value in the corresponding register, but without
freeing the associated register. Rewrite temp_save() to call
temp_sync() followed by temp_dead().
Reviewed-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
Add a new function tcg_reg_sync() to synchronize the canonical location
of a temp with the value in the associated register, but without freeing
it. Rewrite tcg_reg_free() to first call tcg_reg_sync() and then to free
the register.
Reviewed-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
A lot of code is duplicated to mark a temporary as dead. Replace it
by temp_dead(), which in addition marks the temp as saved in memory
for globals and local temps, instead of doing this a posteriori in
temp_save().
Reviewed-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
On x86_64, remove the constraint on the third argument register which
is not needed:
- For loads the helper arguments are env, addr, mem_idx. The addr
value should not be in the two first argument registers as they are
used in tcg_out_tlb_load().
- For stores the helper arguments are env, addr, data, mem_idx.
The addr and data values should not be in the two first argument
registers as they are used in tcg_out_tlb_load(). The data value
should also not be in the two first argument registers, but could
be in the third argument register in which case it would be already
loaded at the right location.
Reviewed-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
Now that CONFIG_TCG_PASS_AREG0 has been removed, it's easier to get
an optimal code for the load/store functions.
First swap the two registers used in tcg_out_tlb_load() so that the
address end-up in the second register instead of the first one. Adjust
tcg_out_qemu_ld() and tcg_out_qemu_st() to respectively call
tcg_out_qemu_ld_direct() and tcg_out_qemu_st_direct() with the correct
registers. Then replace the register shifting by direct load of the
arguments.
Reviewed-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
We had a hack for arm and sparc, allocating code_gen_prologue to a
special section. Which, honestly does no good under certain cases.
We've already got limits on code_gen_buffer_size to ensure that all
TBs can use direct branches between themselves; reuse this limit to
ensure the prologue is also reachable.
As a bonus, we get to avoid marking a page of the main executable's
data segment as executable.
Signed-off-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Blue Swirl <blauwirbel@gmail.com>
Like add2, do operand ordering, constant folding, and dead operand
elimination. The latter happens about 15% of all mulu2 during an
x86_64 bios boot.
Signed-off-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
When x86_64 guest is not in 64-bit mode, the high-part of the 64-bit
add is dead. When the host is 32-bit, we can simplify to 32-bit
arithmetic.
Signed-off-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
We can re-use these for implementing double-word folding.
Signed-off-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
This saves a whole lot of repetitive code sequences.
Signed-off-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
Reduces code duplication and prefers
movcond d, c1, c2, const, s
to
movcond d, c1, c2, s, const
It also prefers
add r, r, c
over
add r, c, r
when both inputs are known constants. This doesn't matter for true add, as
we will fully constant fold that. But it matters for a follow-on patch using
this routine for add2 which may not be fully foldable.
Signed-off-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
Note that in the general reg=reg,reg case we're restricted
to 16-bit insertions. This makes it easy to allow "any"
constant as input, as post-truncation it will fit into the
constant load insn for which we have room in the bundle.
Signed-off-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
It is possible to slightly optimize the TLB access code, by replacing
the movi + and instructions by a deposit instruction.
Reviewed-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
Remove suboptimal register shifting in qemu_ld/st ops, introduced at the
CONFIG_TCG_PASS_AREG0 time.
As mem_idx is now loaded in register R58/R59 for the slow path, we have
to make sure to do it last, to not add additional register constraints.
Reviewed-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
Implement movcond_i32/64 on ia64 hosts. It is not possible to have
immediate compare arguments without adding a new bundle, but it is
possible to have 22-bit immediate value arguments.
Reviewed-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
Use stack instead of temp_buf array in CPUState for TCG temps.
Signed-off-by: Blue Swirl <blauwirbel@gmail.com>
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
Implement movcond_i32 for ARM, as the sequence
mov dst, v2 (implicitly done by the tcg common code)
cmp c1, c2
movCC dst, v1
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
The code to emit either an immediate cmp or a register cmp insn is
duplicated in several places; factor it out into its own function.
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
Now that we're always sparcv9, we can not bother using Bicc for
32-bit branches and BPcc for 64-bit branches and instead always
use BPcc.
New interfaces allow less direct use of tcg_out32 and raw numbers
inside the qemu_ld/st routines.
Signed-off-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Blue Swirl <blauwirbel@gmail.com>
We must care not to clobber the high parts before we consume them.
Signed-off-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Blue Swirl <blauwirbel@gmail.com>
The set of comparisons that can immediately use the carry are LTU/GEU,
not LTU/LEU. Don't swap operands when we need a temp register; the
register may already be in use from setcond2.
Signed-off-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Blue Swirl <blauwirbel@gmail.com>
The datalo variable is still live in the miss path. Use another
when reconstructing the full data value.
Signed-off-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Blue Swirl <blauwirbel@gmail.com>
Like brcond2, use tcg_high_cond. Use movcc instead of branches.
Signed-off-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Blue Swirl <blauwirbel@gmail.com>
GUEST_BASE support is now supported by all TCG backends, and is
now mandatory. Drop the now-pointless TCG_TARGET_HAS_GUEST_BASE
define (set by every backend) and the error if it is unset.
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Riku Voipio <riku.voipio@linaro.org>
The pointer entry 'temps' always refers to the array entry 'static_temps'.
Removing the pointer and renaming 'static_temps' to 'temps' reduces the
size of TCGContext (4 or 8 byte) and allows better code generation.
Signed-off-by: Stefan Weil <sw@weilnetz.de>
Reviewed-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Blue Swirl <blauwirbel@gmail.com>
* 'trivial-patches' of git://github.com/stefanha/qemu:
versatilepb: Use symbolic indices for ARM PIC
qdev: kill bogus comment
qemu-barrier: Fix compiler version check for future gcc versions
hw: Add missing 'static' attribute for QEMUMachine
cleanup useless return sentence
qemu-sockets: Fix compiler warning (regression for MinGW)
vnc: Fix spelling (hellmen -> hellman) in comment
slirp: Fix spelling in comment (enought -> enough, insure -> ensure)
tcg/arm: Use tcg_out_mov_reg rather than inline equivalent code
cpu: Add missing 'static' attribute to qemu_global_mutex
configure: Support empty target list (--target-list=)
hw: Fix return value check for bdrv_read, bdrv_write
The table that was recently added for hppa is generally usable.
And with the renumbering of the TCG_COND constants it's not too
difficult to compute rather than have a table.
Signed-off-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
There are several cases that can be handled easier inside both
translators and code generators if we have out-of-band values
for conditions. It's easy enough to handle ALWAYS and NEVER in
the natural way inside the tcg middle-end.
Signed-off-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
Before we rearrange the TCG_COND enumeration, add a predicate for
the (single) use of comparisons vs TCGCond.
Signed-off-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
The TCG jmp operation doesn't really make sense in the QEMU context, it
is unused, it is not implemented by some targets, and it is wrongly
implemented by some others.
This patch simply removes it.
Reviewed-by: Richard Henderson <rth@twiddle.net>
Acked-by: Blue Swirl <blauwirbel@gmail.com>
Acked-by: Stefan Weil<sw@weilnetz.de>
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
Use the recently introduced tcg_out_mov_reg() function rather than
the equivalent inline code.
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Aurelien Jarno <aurelien@aurel32.net>
Signed-off-by: Stefan Hajnoczi <stefanha@gmail.com>
Support for helper functions with 5 arguments was missing
in the code generator and in the interpreter.
There is no need to pass the constant TCG_AREG0 from the
code generator to the interpreter. Remove that code for
the INDEX_op_qemu_st* opcodes.
Signed-off-by: Stefan Weil <sw@weilnetz.de>
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
The movcond_i32 op has to be protected with TCG_TARGET_HAS_movcond_i32
to fix the build with -march < i686.
Thanks to Richard Henderson for the hint.
Reported-by: Alex Barcelo <abarcelo@ac.upc.edu>
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
When movcond_i32 is available we can further reduce the generated
op count from 12 to 6, and the generated code size on i686 from
88 to 74 bytes.
Signed-off-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
Avoiding 64-bit arithmetic (outside of the compare) reduces the
generated op count from 15 to 12, and the generated code size on
i686 from 105 to 88 bytes.
Signed-off-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
Checking that we don't try for idx != [01] is trivial. Checking
that we don't issue more than one of any index requires a tad
more data and some ifdefs protecting that new variable.
Signed-off-by: Richard Henderson <rth@twiddle.net>
Cc: Max Filippov <jcmvbkbc@gmail.com>
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
Given these are constants, checking once here means everything
after can assume they're correct.
Signed-off-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
Like the C assert macro, except only enabled for CONFIG_DEBUG_TCG,
and without having to set _NDEBUG and disable all other asserts at
the same time.
The use of __builtin_unreachable (when available) gives the compiler
the same information, which may (or may not) help it optimize better.
Signed-off-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
For tcg_gen_concat_i32_i64 we only use deposit if the host supports it.
For tcg_gen_concat32_i64 even if the host does not, as we get identical
code before and after.
Note that this relies on the ANDI -> EXTU patch for the identity claim.
Signed-off-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
Note that xori_i64 failed to perform even the minimal
optimizations promised by the README.
Signed-off-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
Note that andi_i64 failed to perform even the minimal
optimizations promised by the README.
Signed-off-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
The README file documented the operand ordering of the tcg_gen_*
functions. Since we're documenting opcodes here, use the true
operand ordering.
Signed-off-by: Richard Henderson <rth@twiddle.net>
Cc: malc <av1474@comtv.ru>
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
Fix the MIPS32(R2) cpu detection so that it also works with
-march=octeon. Thanks to Andrew Pinski for the hint.
Cc: Andrew Pinski <apinski@cavium.com>
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
* 'tcg-sparc' of git://repo.or.cz/qemu/rth:
tcg-sparc: Preserve branch destinations during retranslation
tcg-sparc: Fix and enable direct TB chaining.
tcg-sparc: Add %g/%o registers to alloc_order
tcg-sparc: Use defines for temporaries.
tcg-sparc: Mask shift immediates to avoid illegal insns.
tcg-sparc: Clean up cruft stemming from attempts to use global registers.
tcg-sparc: Change AREG0 in generated code to %i0.
tcg-sparc: Support GUEST_BASE.
tcg-sparc: Fix qemu_ld/st to handle 32-bit host.
tcg-sparc: Assume v9 cpu always, i.e. force v8plus in 32-bit mode.
tcg-sparc: Don't MAP_FIXED on top of the program
tcg-sparc: Fix ADDX opcode.
tcg-sparc: Hack in qemu_ld/st64 for 32-bit.
linux-user: Use memcpy in get_user/put_user.
The TCG targets no longer need individual implementations.
Since commit 6a18ae2d29,
'flags' is no longer used in tcg_target_get_call_iarg_regs_count.
The remaining tcg_target_get_call_iarg_regs_count is trivial and only
called once. Therefore the patch eliminates it completely.
Signed-off-by: Stefan Weil <sw@weilnetz.de>
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
32 bit x86 hosts don't need registers for helper function arguments
because they use the default stack based calling convention.
Removing the registers allows simpler code for function
tcg_target_get_call_iarg_regs_count.
Signed-off-by: Stefan Weil <sw@weilnetz.de>
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
While 64 bit hosts use the first three registers which are also used
as function input parameters, 32 bit hosts use TCG_REG_EAX and
TCG_REG_EDX which are not used in parameter passing.
After defining new register macros for the registers used in L
constraint, the patch replaces most occurrences of
tcg_target_call_iarg_regs[0], tcg_target_call_iarg_regs[1] and
tcg_target_call_iarg_regs[2] by those new macros.
tcg_target_call_iarg_regs remains unchanged when it is used for input
arguments (only with 64 bit hosts) before tcg_out_calli.
A comment related to those registers was fixed, too.
Signed-off-by: Stefan Weil <sw@weilnetz.de>
[aurel32: build fix on i386, small optimization for i386 in the prologue]
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
TCG uses 6 registers for function arguments on 64 bit Linux hosts,
but only 4 registers on W64 hosts.
Commit 2999a0b200 increased the number
of arguments for some important helper functions from 4 to 5
which triggered a bug for W64 hosts: QEMU aborts when executing
helper_lcall_real in the guest's BIOS because function
tcg_target_get_call_iarg_regs_count always returned 6.
As W64 has only 4 registers for arguments, the 5th argument must be
passed on the stack using a correct stack offset.
Signed-off-by: Stefan Weil <sw@weilnetz.de>
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
Commit 25c4d9cc changed all TCGOpcode enums to be available, so we don't
need to #ifdef #endif the one that are available only on some targets.
This makes the code easier to read.
Reviewed-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
The "op a, a, b" form is better handled on non-RISC host than the "op
a, b, a" form, so swap the arguments to this form when possible, and
when b is not a constant.
This reduces the number of generated instructions by a tiny bit.
Reviewed-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
When both argument of brcond/movcond/setcond are the same or when one
of the two values is a constant equal to zero, it's possible to do
further optimizations.
Reviewed-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
Now that it's possible to detect copies, we can optimize the case
the "op r, a, a => movi r, 0". This helps in the computation of
overflow flags when one of the two args is 0.
Reviewed-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
Now that we can easily detect all copies, we can optimize the
"op r, a, a => mov r, a" case a bit more.
Reviewed-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
It is possible to due copy propagation for all operations, even the one
that have side effects or clobber arguments (it only concerns input
arguments). That said, the call operation should be handled differently
due to the variable number of arguments.
Reviewed-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
The copy propagation pass tries to keep track what is a copy of what
and what has copy of what, and in addition it keep a circular list of
of all the copies. Unfortunately this doesn't fully work: a mov from
a temp which has a state "COPY" changed it into a state "HAS_COPY".
Later when this temp is used again, it is considered has not having
copy and thus no propagation is done.
This patch fixes that by removing the hiearchy between copies, and thus
only keeping a "COPY" state both meaning "is a copy" and "has a copy".
The decision of which copy to use is deferred to the actual temp
replacement. At this stage there is not one best choice to do, but only
better choices than others. For doing the best choice the operation
would have to be parsed in reversed to know if a temp is going to be
used later or not. That what is done by the liveness analysis. At this
stage it is known that globals will be always live, that local temps
will be dead at the end of the translation block, and that the temps
will be dead at the end of the basic block. This means that this stage
should try to replace temps by local temps or globals and local temps
by globals.
Reviewed-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
The copy propagation doesn't check the types of the temps during copy
propagation. However TCG is using the mov_i32 for the i64 to i32
conversion and thus the two are not equivalent.
With this patch tcg_opt_gen_mov() doesn't consider two temps of
different type as copies anymore.
So far it seems the optimization was not aggressive enough to trigger
this bug, but it will be triggered later in this series once the copy
propagation is improved.
Reviewed-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
TCG_TEMP_ANY has no different meaning than TCG_TEMP_UNDEF, so use
the later instead.
Reviewed-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
movcond operation can be implemented on MIPS32 Release 2 using the MOVN,
MOVZ, SLT and SLTU instructions.
Reviewed-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
deposit operations can be optimized on MIPS32 Release 2 using the INS
instruction.
Reviewed-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
rotr operations can be optimized on MIPS32 Release 2 using the ROTR and
ROTRV instructions. Also implemented rotl operations by subtracting the
shift from 32.
Reviewed-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
bswap operations can be optimized on MIPS32 Release 2 using the ROTR,
WSBH and SEH instructions. We can't use the non-R2 code to implement the
ops due to registers constraints, so don't define the corresponding
TCG_TARGET_HAS_bswap* values.
Also bswap16* operations are supposed to be called with the 16 high bits
zeroed. This is the case everywhere (including for TCG by definition)
except when called from the store helper. Remove the AND instructions from
bswap16* and move it there.
Reviewed-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
MIPS has some conditional branch instructions when comparing with zero.
Use them.
Reviewed-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
Use stack instead of temp_buf array in CPUState for TCG
temps.
Reviewed-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
Instead of int, use the correct TCGArg and TCGReg type: TCGReg when
representing a TCG target register, TCGArg when representing the latter
or a constant.
Reviewed-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
Recent versions of GCC emit warnings when compiling user mode targets.
Kill them by reordering a bit the #ifdef.
Reviewed-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
The 'Z' constraint has been introduced to map the zero register. However
when the op also accept a constant, there is no point to accept the zero
register in addition.
Reviewed-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
The xtensa-test image generates a sra_i32 with count 0x40.
Whether this is accident of tcg constant propagation or
originating directly from the instruction stream is immaterial.
Signed-off-by: Richard Henderson <rth@twiddle.net>
Don't use -ffixed-gN. Don't link statically. Don't save/restore
AREG0 around calls. Don't allocate space on the stack for AREG0 save.
Signed-off-by: Richard Henderson <rth@twiddle.net>
At the same time, split out the tlb load logic to a new function.
Fixes the cases of two data registers and two address registers.
Fixes the signature of, and adds missing, qemu_ld/st opcodes.
Signed-off-by: Richard Henderson <rth@twiddle.net>
Current code doesn't actually work in 32-bit mode at all. Since
no one really noticed, drop the complication of v7 and v8 cpus.
Eliminate the --sparc_cpu configure option and standardize macro
testing on TCG_TARGET_REG_BITS / HOST_LONG_BITS
Signed-off-by: Richard Henderson <rth@twiddle.net>
The CONFIG_TCG_PASS_AREG0 code for calling ld/st helpers
was not respecting the ABI requirement for 64-bit values
being aligned in registers.
Mirror the ARM port in use of helper functions to marshal
arguments into the correct registers.
Signed-off-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
Neither of these functions were performing double-word
compares properly.
Signed-off-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
Commit 6375e09e changed the type of TranslationBlock.tb_next,
but failed to change the type of TCGContext.tb_next.
Signed-off-by: Richard Henderson <rth@twiddle.net>
Reviewed-by: Andreas Färber <afaerber@suse.de>
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
While swapping constants to the second operand, swap
sources matching destinations to the first operand.
Signed-off-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
Implemented with setcond if the target does not provide
the optional opcode.
Signed-off-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
Commit e31b0a7c05 fixed copy propagation on
32-bit host by restricting the copy between different types. This was the
wrong fix.
The real problem is that the all temps states should be reset at the end
of a basic block. This was done by adding such operations in the switch,
but brcond2 was forgotten (that's why the crash was only observed on 32-bit
hosts).
Fix that by looking at the TCG_OPF_BB_END instead. We need to keep the case
for op_set_label as temps might be modified through another path.
Cc: Blue Swirl <blauwirbel@gmail.com>
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
Given the copy propagation breakage on 32-bit hosts has been fixed
commit e31b0a7c05 can be reverted.
Cc: Blue Swirl <blauwirbel@gmail.com>
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
set_label is effectively the end of a basic block, as no optimization
can be made accross it. It was treated as such in the liveness analysis
code, but as a special case.
Mark it with TCG_OPF_BB_END flag so that this information can be used
by other parts of the TCG code, and remove the special case in the liveness
analysis code.
Cc: Blue Swirl <blauwirbel@gmail.com>
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
On x86, it is possible to move a constant value to memory. Add code to
handle a constant argument to load/store ops.
Reviewed-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
Now that CONFIG_TCG_PASS_AREG0 is enabled for all targets,
remove dead code and support for !CONFIG_TCG_PASS_AREG0 case.
Remove dyngen-exec.h and all references to it. Although included by
hw/spapr_hcall.c, it does not seem to use it.
Remove unused HELPER_CFLAGS.
Signed-off-by: Blue Swirl <blauwirbel@gmail.com>
Reviewed-by: Richard Henderson <rth@twiddle.net>
optimizer.c contains some cases were the break is appearing in both the
if and the else parts. Fix that by moving it to the outer part. Also
move some common code there.
Reviewed-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
brcond and setcond ops are not commutative, but it's easy to compute the
new condition after swapping the arguments. Try to always put the constant
argument in second position like for commutative ops, to help backends to
generate better code.
Reviewed-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
Split expression simplification in multiple parts so that a given op
can appear multiple times. This patch should not change anything.
Reviewed-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
Now that there are two passes of optimization (optimize.c, liveness)
there is no point of outputing the statistics of the liveness part
only. Update the code to take into account both optimizations.
Reviewed-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
The load/store slow path has been broken in e141ab52d:
- We need to move 4 registers for store functions and 3 registers for
load functions and not the reverse.
- According to the s390x calling convention the arguments of a function
should be zero extended. This means that the register shift should be
done with TCG_TYPE_I64 to ensure the higher word is correctly zero
extended when needed.
I am aware that CONFIG_TCG_PASS_AREG0 is being removed and thus that
this patch can be improved, but doing so means it can also be applied to
the 1.1 and 1.2 stable branches.
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
Signed-off-by: Alexander Graf <agraf@suse.de>
The CONFIG_TCG_PASS_AREG0 code for calling ld/st helpers was
broken in that it did not respect the ABI requirement that 64
bit values were passed in even-odd register pairs. The simplest
way to fix this is to implement some new utility functions
for marshalling function arguments into the correct registers
and stack, so that the code which sets up the address and
data arguments does not need to care whether there has been
a preceding env argument.
Based on commit 9716ef3b for ARM by Peter Maydell.
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Store slow path has been broken in e141ab52d:
- the arguments are shifted before the last one (mem_index) is written.
- the shift is done for both slow and fast paths.
Fix that. Also optimize a bit by bundling the move together. This still
can be optimized, but it's better to wait for a decision to be taken on
the arguments order.
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
The CONFIG_TCG_PASS_AREG0 code for calling ld/st helpers was
broken in that it did not respect the ABI requirement that 64
bit values were passed in even-odd register pairs. The simplest
way to fix this is to implement some new utility functions
for marshalling function arguments into the correct registers
and stack, so that the code which sets up the address and
data arguments does not need to care whether there has been
a preceding env argument.
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Blue Swirl <blauwirbel@gmail.com>
If tci_out_label is called in the context of tcg_gen_code_search_pc, we
could be overwriting an already patched relocation with zero -- and not
repatch it because the set_label is past search_pc, causing a QEMU crash
when it tries to branch to a zero label.
Not writing anything to the relocation area seems to be in line with what
other backends do from the couple I looked at (x86, ppc).
Signed-off-by: Scott Wood <scottwood@freescale.com>
Signed-off-by: Blue Swirl <blauwirbel@gmail.com>
Commit eeacee4d86 changed the syntax of tcg_dump_ops, but didn't convert
all users (notably missing the ppc ones) to it. Fix them to the new syntax.
Signed-off-by: Alexander Graf <agraf@suse.de>
Signed-off-by: malc <av1474@comtv.ru>
Don't use global variables directly but via accessor functions. Rename globals.
Convert macros to functions, add GCC format attributes.
Signed-off-by: Blue Swirl <blauwirbel@gmail.com>
powerpc-apple-darwin9-gcc-4.2.1 (GCC) 4.2.1 (Apple Inc. build 5577)
does not define _CALL_DARWIN, leading to unexpected behavior w.r.t.
register clobbering and stack frame layout.
Since _CALL_DARWIN is a reserved identifier, define a custom
TCG_TARGET_CALL_DARWIN based on either _CALL_DARWIN or __APPLE__.
Signed-off-by: Andreas F?rber <andreas.faerber@web.de>
Signed-off-by: malc <av1474@comtv.ru>
In qemu_ld/st load the registers for the helper calls directly rather
than rotating them around afterwards for AREG0.
Also clobber the additional register.
Signed-off-by: Andreas F?rber <afaerber@suse.de>
Signed-off-by: malc <av1474@comtv.ru>
Adjust the tcg_out_qemu_{ld,st}() slow paths to pass AREG0 in r3,
based on patches by malc.
Also adjust the registers clobbered, based on patch by Alex.
Signed-off-by: Andreas Färber <afaerber@suse.de>
Acked-by: Alexander Graf <agraf@suse.de>
[AF: Do not hardcode r3 for AREG0, requested by Alex]
Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
This accounts for the additional addr_reg2 register.
Signed-off-by: Andreas Färber <afaerber@suse.de>
Acked-by: Alexander Graf <agraf@suse.de>
Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
Also assure i64 alignment where necessary.
Alignment code optimization suggested by malc.
Signed-off-by: Andreas Färber <afaerber@suse.de>
Acked-by: Alexander Graf <agraf@suse.de>
Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
For targets where TARGET_LONG_BITS != 32, i.e. 64-bit guests,
addr_reg is moved to r4. For hosts without TCG_TARGET_CALL_ALIGN_ARGS
either data_reg2 or data_reg or a masked version thereof would overwrite
r4. Place it in r5 instead, matching TCG_TARGET_CALL_ALIGN_ARGS hosts.
This fixes immediate crashes of 64-bit guests observed on Darwin/ppc but
not on Darwin/ppc64.
Signed-off-by: Andreas Färber <andreas.faerber@web.de>
Acked-by: malc <av1474@comtv.ru>
Acked-by: Alexander Graf <agraf@suse.de>
Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
w64 uses the registers rcx, rdx, r8 and r9 for function arguments,
so it needs a different declaration of tcg_target_call_iarg_regs.
rax, rcx, rdx, r8, r9, r10 and r11 may be changed by function calls.
rbx, rbp, rdi, rsi, r12, r13, r14 and r15 remain unchanged by function calls.
Reviewed-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Stefan Weil <sw@weilnetz.de>
Not all i386 / x86_64 hosts use ELF.
Ask the compiler whether ELF is used.
On w64, gdb crashes when ELF_HOST_MACHINE is defined.
Cc: Blue Swirl <blauwirbel@gmail.com>
Acked-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Stefan Weil <sw@weilnetz.de>
There two entries of INDEX_op_ld_i64 in the ppc_op_defs. That causes an
assertion failure in tcg_add_target_add_op_defs() when --enable-debug is
used on a ppc64 backend (that's ppc64 host, not target).
Signed-off-by: Li Zhang <zhlcindy@linux.vnet.ibm.com>
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
Signed-off-by: malc <av1474@comtv.ru>
This allows us to actually supply a function name in softmmu builds;
gdb doesn't pick up the minimal symbol table otherwise. Also add a
bit of documentation and statically generate more of the ELF image.
Signed-off-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Blue Swirl <blauwirbel@gmail.com>
This allows us to generate unwind info for the dynamicly generated
code in the code_gen_buffer. Only i386 is converted at this point.
Signed-off-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Blue Swirl <blauwirbel@gmail.com>
Optionally, make memory access helpers take a parameter for CPUState
instead of relying on global env.
On most targets, perform simple moves to reorder registers. On i386,
switch from regparm(3) calling convention to standard stack-based
version.
Signed-off-by: Blue Swirl <blauwirbel@gmail.com>
Use stack based calling convention (GCC default) for interfacing with
generated code instead of register based convention (regparm(3)).
Signed-off-by: Blue Swirl <blauwirbel@gmail.com>
next_tb is the numeric value of a tcg target (= QEMU host) address.
Using tcg_target_ulong instead of unsigned long shows this and makes
the code portable for hosts with an unusual size of long (w64).
The type cast '(long)(next_tb & ~3)' was not needed (casting
unsigned long to long does not change the bits, and nor does
casting long to pointer for most (= all non w64) hosts.
It is removed here.
Macro or function tcg_qemu_tb_exec is used to set next_tb.
The function also returns next_tb. Therefore tcg_qemu_tb_exec
must return a tcg_target_ulong.
Signed-off-by: Stefan Weil <sw@weilnetz.de>
Signed-off-by: Blue Swirl <blauwirbel@gmail.com>
An attempt to allocate a large memory chunk after a small one resulted in
circular links in list of pools. It caused the same memory being
allocated twice for different arrays.
Now pools for large memory chunks are kept in separate list and are
freed during pool reset because current allocator can not reuse them.
Signed-off-by: Kirill Batuzov <batuzovk@ispras.ru>
Signed-off-by: Blue Swirl <blauwirbel@gmail.com>
Scripted conversion:
for file in *.[hc] hw/*.[hc] hw/kvm/*.[hc] linux-user/*.[hc] linux-user/m68k/*.[hc] bsd-user/*.[hc] darwin-user/*.[hc] tcg/*/*.[hc] target-*/cpu.h; do
sed -i "s/CPUState/CPUArchState/g" $file
done
All occurrences of CPUArchState are expected to be replaced by QOM CPUState,
once all targets are QOM'ified and common fields have been extracted.
Signed-off-by: Andreas Färber <afaerber@suse.de>
Reviewed-by: Anthony Liguori <aliguori@us.ibm.com>
tcg_out_label is always called with a third argument of pointer type
which was casted to tcg_target_long.
These casts can be avoided by changing the prototype of tcg_out_label.
There was also a cast to long. For most hosts with
sizeof(long) == sizeof(tcg_target_long) == sizeof(void *) this did not
matter, but for w64 it was wrong. This is fixed now.
Cc: Blue Swirl <blauwirbel@gmail.com>
Cc: Richard Henderson <rth@twiddle.net>
Signed-off-by: Stefan Weil <sw@weilnetz.de>
Signed-off-by: Blue Swirl <blauwirbel@gmail.com>
The TCG targets i386 and tci needed a change of the function
prototype for w64.
This change is currently not needed for the other TCG targets,
but it can be applied to avoid code differences.
Cc: Blue Swirl <blauwirbel@gmail.com>
Cc: Andrzej Zaborowski <balrogg@gmail.com>
Cc: Richard Henderson <rth@twiddle.net>
Cc: Aurelien Jarno <aurelien@aurel32.net>
Cc: Alexander Graf <agraf@suse.de>
Signed-off-by: Stefan Weil <sw@weilnetz.de>
Signed-off-by: Blue Swirl <blauwirbel@gmail.com>
flush_icache_range takes two address parameters which must be large
enough to address any address of the host.
For hosts with sizeof(unsigned long) == sizeof(void *), this patch
changes nothing. All currently supported hosts fall into this category.
For w64 hosts, sizeof(unsigned long) is 4 while sizeof(void *) is 8,
so the use of tcg_target_ulong is needed for i386 and tci (the tcg
targets which work with w64).
Signed-off-by: Stefan Weil <sw@weilnetz.de>
Signed-off-by: Blue Swirl <blauwirbel@gmail.com>
This change makes tcg_target_ulong available in tcg-target.h.
Signed-off-by: Stefan Weil <sw@weilnetz.de>
Signed-off-by: Blue Swirl <blauwirbel@gmail.com>
The standard include files are already included in qemu-common.h.
malloc.h and alloca.h were needed for alloca() which was removed
from TCG code some years ago when switching from dyngen to TCG
(see commit 49516bc0d6).
Signed-off-by: Stefan Weil <sw@weilnetz.de>
Signed-off-by: Stefan Hajnoczi <stefanha@linux.vnet.ibm.com>
ARM still doesn't support 16GB buffers in 32-bit modes, replace the
16GB by 16MB in the comment.
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
Signed-off-by: Stefan Hajnoczi <stefanha@linux.vnet.ibm.com>
On ARM, in Thumb mode r7 is used for the framepointer; this meant
that we would fail to compile in debug mode because we were using r7
for TCG_AREG0. Shift to r6 instead to avoid this clash.
(Bug reported as LP:870990.)
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Andrzej Zaborowski <andrew.zaborowski@intel.com>
On ARM, don't map the code buffer at a fixed location, and fix up the
call/goto tcg routines to let it do long jumps.
Mapping the code buffer at a fixed address could sometimes result in it being
mapped over the top of the heap with pretty random results.
Signed-off-by: Dr. David Alan Gilbert <david.gilbert@linaro.org>
Signed-off-by: Andrzej Zaborowski <andrew.zaborowski@intel.com>
Make tcg_const_ptr() include a cast so that you can pass it a
pointer. This allows us to drop the casts we had in all the places
that use this macro.
Acked-by: Andreas Färber <andreas.faerber@web.de>
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Stefan Hajnoczi <stefanha@linux.vnet.ibm.com>
TCG_TARGET_REG_BITS is declared in tcg.h for all TCG targets.
Signed-off-by: Stefan Weil <sw@weilnetz.de>
Signed-off-by: Stefan Hajnoczi <stefanha@linux.vnet.ibm.com>
This is standard for other tcg targets and improves tci, too.
Signed-off-by: Stefan Weil <sw@weilnetz.de>
Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
In both cases, val is computed, but then not used in the
subsequent line, which then re-computes the quantity in
a different type (int32_t vs unsigned long).
Keep the computation type that's been working so far.
Signed-off-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Blue Swirl <blauwirbel@gmail.com>
* 's390-1.0' of git://repo.or.cz/qemu/agraf:
s390x: initialize virtio dev region
tcg: Use TCGReg for standard tcg-target entry points.
tcg: Standardize on TCGReg as the enum for hard registers
s390x: Add shutdown for TCG s390-virtio machine
s390: Fix cpu shutdown for KVM
s390: fix short kernel command lines
s390: fix reset hypercall to reset the status
s390x: implement SIGP restart and shutdown
s390x: implement rrbe instruction properly
s390x: update R and C bits in storage key
s390x: make ipte 31-bit aware
s390x: add ldeb instruction
Including tcg_out_ld, tcg_out_st, tcg_out_mov, tcg_out_movi.
Signed-off-by: Richard Henderson <rth@twiddle.net>
Reviewed-by: Andreas Färber <afaerber@suse.de>
Reviewed-by: Stefan Weil <sw@weilnetz.de>
Signed-off-by: Alexander Graf <agraf@suse.de>
Most targets did not name the enum; tci used TCGRegister.
Signed-off-by: Richard Henderson <rth@twiddle.net>
Reviewed-by: Andreas Färber <afaerber@suse.de>
Reviewed-by: Stefan Weil <sw@weilnetz.de>
Signed-off-by: Alexander Graf <agraf@suse.de>
tcg/ppc64/tcg-target.c has a couple of places where variables are set
unconditionally, but otherwise used only for softmmu builds, not
userspace only builds. This causes compiler warnings (which are fatal
by default) when compiling for a ppc64 host with gcc 4.6. This patch
fixes the problem by moving the code which defines and sets the
variables into the CONFIG_SOFTMMU guarded regions.
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
Signed-off-by: Alexander Graf <agraf@suse.de>
* 'tci' of git://qemu.weilnetz.de/qemu:
tcg: Add tcg interpreter to configure / make
tcg: Add tci disassembler
tcg: Add interpreter for bytecode
tcg: Add bytecode generator for tcg interpreter
tcg: Make ARRAY_SIZE(tcg_op_defs) globally available
tcg: TCG targets may define tcg_qemu_tb_exec
The error being caused by the failure to copy the other half of
the input to the output after having narrowed the deposit operation.
Signed-off-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: malc <av1474@comtv.ru>
Unlike other tcg target code generators, this one does not generate
machine code for some cpu. It generates machine independent bytecode
which is interpreted later.
This allows running QEMU on any host.
Interpreted bytecode is slower than direct execution of generated
machine code.
Signed-off-by: Stefan Weil <sw@weilnetz.de>
tcg_op_defs was already a global array.
The tci disassembler also needs ARRAY_SIZE(tcg_op_defs),
so add a new global constant with this value.
Signed-off-by: Stefan Weil <sw@weilnetz.de>
Targets may use a non standard definition of tcg_tb_exec
by defining this macro in their tcg_target.h.
This is used here by ppc. It will be used by the TCG interpreter, too.
Cc: malc <av1474@comtv.ru>
Signed-off-by: Stefan Weil <sw@weilnetz.de>
If the deposit replaces the entire word, optimize to a move.
If we're inserting to the top of the word, avoid the mask of arg2
as we'll be shifting out all of the garbage and shifting in zeros.
If the host is 32-bit, reduce a 64-bit deposit to a 32-bit deposit
when possible.
Signed-off-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Blue Swirl <blauwirbel@gmail.com>
Remove the unused function tcg_out_addi() from the s390 TCG backend;
this brings it into line with other backends.
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Acked-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Blue Swirl <blauwirbel@gmail.com>
Remove the unused function tcg_out_addi() from the ia64 TCG backend;
this brings it into line with other backends.
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Blue Swirl <blauwirbel@gmail.com>
x86 cannot provide an optimized generic deposit implementation. But at
least for a few special cases, namely for writing bits 0..7, 8..15, and
0..15, versions using only a single instruction are feasible.
Introducing such limited support improves emulating 16-bit x86 code on
x86, but also rarer cases where 32-bit or 64-bit code accesses bytes or
words.
Signed-off-by: Jan Kiszka <jan.kiszka@siemens.com>
Signed-off-by: Blue Swirl <blauwirbel@gmail.com>
Remove the unused function tcg_out_addi() from the ARM TCG backend;
this fixes a compilation failure on ARM hosts with newer gcc.
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Blue Swirl <blauwirbel@gmail.com>
These functions are defined in the tcg target specific file
tcg-target.c.
The forward declarations assert that every tcg target uses
the same function prototype.
Signed-off-by: Stefan Weil <weil@mail.berlios.de>
Signed-off-by: Blue Swirl <blauwirbel@gmail.com>
It is now declared for all tcg targets in tcg.h,
so the tcg target specific declarations are redundant.
Signed-off-by: Stefan Weil <weil@mail.berlios.de>
Signed-off-by: Blue Swirl <blauwirbel@gmail.com>
TCG_TARGET_REG_BITS can be determined by the compiler,
so there is no need to declare it for each individual tcg target.
This is especially important for new tcg targets
which will be supported by the tcg interpreter.
Signed-off-by: Stefan Weil <weil@mail.berlios.de>
Signed-off-by: Blue Swirl <blauwirbel@gmail.com>
The second register is only needed for 32 bit hosts.
Cc: Vassili Karpov <av1474@comtv.ru>
Fine-with-me'd-by: Vassili Karpov <av1474@comtv.ru>
Signed-off-by: Stefan Weil <weil@mail.berlios.de>
Signed-off-by: Blue Swirl <blauwirbel@gmail.com>
The second register is only needed for 32 bit hosts.
Cc: Blue Swirl <blauwirbel@gmail.com>
Signed-off-by: Stefan Weil <weil@mail.berlios.de>
Signed-off-by: Blue Swirl <blauwirbel@gmail.com>
The second register is only needed for 32 bit hosts.
Cc: Alexander Graf <agraf@suse.de>
Acked-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Stefan Weil <weil@mail.berlios.de>
Signed-off-by: Blue Swirl <blauwirbel@gmail.com>
The second register is never used for ia64 hosts.
Cc: Aurelien Jarno <aurelien@aurel32.net>
Signed-off-by: Stefan Weil <weil@mail.berlios.de>
Signed-off-by: Blue Swirl <blauwirbel@gmail.com>
The second register is only needed for 32 bit hosts.
Signed-off-by: Stefan Weil <weil@mail.berlios.de>
Signed-off-by: Blue Swirl <blauwirbel@gmail.com>
The ppc64 code generation backend uses an rldicr (Rotate Left Double
Immediate and Clear Right) instruction to implement zero extension of
a 32 bit quantity to a 64 bit quantity (INDEX_op_ext32u_i64). However
this is wrong - this instruction clears specified low bits of the
value, instead of high bits as we require for a zero extension. It
should instead use an rldicl (Rotate Left Double Immediate and Clear
Left) instruction.
Presumably amongst other things, this causes the SLOF firmware image
used with -M pseries to not boot on a ppc64 host.
It appears this bug was exposed by commit
0bf1dbdcc9 (tcg/ppc64: fix 16/32 mixup)
which enabled the use of the op_ext32u_i64 operation on the ppc64
backend.
Signed-off-by: Thomas Huth <thuth@de.ibm.com>
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
Signed-off-by: malc <av1474@comtv.ru>
Move the declaration and initialisation of some variables in
tcg_out_qemu_ld and tcg_out_qemu_st inside CONFIG_SOFTMMU, to
avoid the "variable set but not used" warning of gcc 4.6.
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: malc <av1474@comtv.ru>
Use enum TCGOpcode instead of plain old int so that the name of
current op can be seen in GDB. Add a default case to switch
so that GCC does not complain about unhandled enum cases.
Signed-off-by: Blue Swirl <blauwirbel@gmail.com>
Signed-off-by: Richard Henderson <rth@twiddle.net>
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Edgar E. Iglesias <edgar.iglesias@gmail.com>
By always defining these symbols, we can eliminate a lot of ifdefs.
To allow this to be checked reliably, the semantics of the
TCG_TARGET_HAS_* macros must be changed from def/undef to true/false.
This allows even more ifdefs to be removed, converting them into
C if statements.
Signed-off-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Blue Swirl <blauwirbel@gmail.com>
This allows the simplification of the op_bits function from
tcg/optimize.c.
Signed-off-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Blue Swirl <blauwirbel@gmail.com>
Copy propagation introduced in 22613af4a6
considered only global registers. However, register temps and stack
allocated locals must be handled differently because register temps
don't survive across brcond.
Fix by propagating only within same class of temps.
Tested-by: Stefan Weil <weil@mail.berlios.de>
Signed-off-by: Blue Swirl <blauwirbel@gmail.com>
Fix breakage by a640f03178
and 55c0975c5b.
Some TCG targets don't implement all TCG ops, so make
optimizing those conditional.
Signed-off-by: Blue Swirl <blauwirbel@gmail.com>
Perform constant folding for NOT and EXT{8,16,32}{S,U} operations.
Signed-off-by: Kirill Batuzov <batuzovk@ispras.ru>
Signed-off-by: Blue Swirl <blauwirbel@gmail.com>
Perform actual constant folding for ADD, SUB and MUL operations.
Signed-off-by: Kirill Batuzov <batuzovk@ispras.ru>
Signed-off-by: Blue Swirl <blauwirbel@gmail.com>
Make tcg_constant_folding do copy and constant propagation. It is a
preparational work before actual constant folding.
Signed-off-by: Kirill Batuzov <batuzovk@ispras.ru>
Signed-off-by: Blue Swirl <blauwirbel@gmail.com>
Added file tcg/optimize.c to hold TCG optimizations. Function tcg_optimize
is called from tcg_gen_code_common. It calls other functions performing
specific optimizations. Stub for constant folding was added.
Signed-off-by: Kirill Batuzov <batuzovk@ispras.ru>
Signed-off-by: Blue Swirl <blauwirbel@gmail.com>
cppcheck reports an error:
qemu/tcg/mips/tcg-target.c:1487: error: Invalid number of character (()
The unpatched code won't compile on mips hosts starting with commit
cea5f9a28f.
Cc: Blue Swirl <blauwirbel@gmail.com>
Signed-off-by: Stefan Weil <weil@mail.berlios.de>
Signed-off-by: Blue Swirl <blauwirbel@gmail.com>
Expand the note on the number of TCG ops generated per target insn,
to be clearer about the range of applicability of the 20 op rule
of thumb. Also add a note about the hard MAX_OP_PER_INSTR limit.
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Blue Swirl <blauwirbel@gmail.com>
Use stack instead of temp_buf array in CPUState for TCG temps.
Signed-off-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Blue Swirl <blauwirbel@gmail.com>
Use TCG_REG_CALL_STACK instead of TCG_REG_SP for consistency.
Acked-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Blue Swirl <blauwirbel@gmail.com>
Use stack instead of temp_buf array in CPUState for TCG temps.
On Sparc64, stack pointer is not aligned but there is a fixed bias of 2047,
so don't try to enforce alignment.
Signed-off-by: Blue Swirl <blauwirbel@gmail.com>
Except for specific cases where the use of %esp changes the encoding of
the instruction, it's cleaner to use TCG_REG_CALL_STACK instead of
TCG_REG_ESP.
Signed-off-by: Blue Swirl <blauwirbel@gmail.com>
The code for stack allocation for call arguments is way too simplistic
to actually work on targets with non-trivial stack allocation policies,
e.g. ppc64. We've also already allocated TCG_STATIC_CALL_ARGS_SIZE worth
of stack for calls which should be well more than any helper needs.
Remove broken dynamic stack allocation code and replace it with an assert.
Should dynamic stack allocation ever be needed again, target specific
functions should be added.
Thanks to Richard Henderson for the analysis.
Signed-off-by: Blue Swirl <blauwirbel@gmail.com>
Make functions take a parameter for CPUState instead of relying
on global env. Pass CPUState pointer to TCG prologue, which moves
it to AREG0.
Thanks to Peter Maydell and Laurent Desnogues for the ARM prologue
change.
Revert the hacks to avoid AREG0 use on Sparc hosts.
Move cpu_has_work() and cpu_pc_from_tb() from exec.h to cpu.h.
Compile the file without HELPER_CFLAGS.
Signed-off-by: Blue Swirl <blauwirbel@gmail.com>
Based on a patch from Hans de Goede <hdegoede@redhat.com>
This warning is new in gcc 4.6.
Acked-by: Amit Shah <amit.shah@redhat.com>
Signed-off-by: Christophe Fergeau <cfergeau@redhat.com>
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
When compiling with DEBUG_TCGV enabled, make the TCGv_ptr type distinct
from TCGv_i32/TCGv_i64. This means that using an i32 or i64 TCG op to
manipulate a TCGv_ptr will always be detected at compile time, rather
than only if compiling on a host system with the other word size.
NB: the tcg_add_ptr and tcg_sub_ptr macros have been removed as they
were not used anywhere.
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
The prototypes for the ld/st functions on a 64 bit host declared
the address parameter as a TCGv_i64 rather than a TCGv_ptr. This
worked OK (since the two are aliases), but needs to be fixed to
allow extension of TCG type debugging to i64/i32/ptr mismatches.
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
Use the correct header in the TCG MIPS code to find cacheflush() on OpenBSD
to fix compilation of the MIPS host support for OpenBSD/mips64 based architecures.
Signed-off-by: Brad Smith <brad@comstyle.com>
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
If an op with dead outputs is not removed, because it has side effects
or has multiple output and only one dead, mark the registers as dead
instead of saving them. This avoid a few register spills on TCG targets
with low register count, especially with div2 and mul2 ops, or when a
qemu_ld* result is not used (prefetch emulation for example).
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
If an op is not removed and has dead output arguments, mark it
in op_dead_args similarly to what is done for input arguments.
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
Allow all args to be dead by replacing the input specific op_dead_iargs
variable by op_dead_args. Note this is a purely mechanical change.
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
Although the TCG generated code is always in ARM mode, it is possible
that the host code was compiled by gcc in Thumb mode (this is often the
default for Linux distributions targeting ARM v7 only). Handle this
by using BLX imm when doing a call from ARM into Thumb mode.
Since BLX imm is not a conditionalisable instruction, we make
tcg_out_call() no longer take a condition code; we were only ever
using it with COND_AL anyway.
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Andrzej Zaborowski <andrew.zaborowski@intel.com>
Add support (if CONFIG_DEBUG_TCG is defined) for debugging leakage
of temporary variables. Generally any temporaries created by
a target while it is translating an instruction should be freed
by the end of that instruction; otherwise carefully crafted
guest code could cause TCG to run out of temporaries and assert.
By calling tcg_check_temp_count() after each instruction we can
check that we are not leaking temporaries in this way.
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
Add a comment about cache coherency and retranslation, so that people
developping new targets based on existing ones are warned of the issue.
Acked-by: Edgar E. Iglesias <edgar.iglesias@gmail.com>
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
Improve constant loading in two ways:
- On all ARM versions, it's possible to load 0xffffff00 = -0x100 using
the mvn rd, #0. Fix the conditions.
- On <= ARMv6 versions, where movw and movt are not available, load the
constants using mov and orr with rotations depending on the constant
to load. This is very useful for example to load constants where the
low byte is 0. This reduce the generated code size by about 7%.
Also fix the coding style at the same time.
Cc: Andrzej Zaborowski <balrog@zabor.org>
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
TCG on MIPS was trying to avoid changing the branch offset, but didn't
due to a stupid typo. Fix it.
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
Due to a typo, qemu_st64 doesn't properly byteswap the 32-bit low word of
a 64 bit word before saving it. This patch fixes that.
Acked-by: Andrzej Zaborowski <balrogg@gmail.com>
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
QEMU uses code retranslation to restore the CPU state when an exception
happens. For it to work the retranslation must not modify the generated
code. This is what is currently implemented in ARM TCG.
However on CPU that don't have icache/dcache/memory synchronised like
ARM, this requirement is stronger and code retranslation must not modify
the generated code "atomically", as the cache line might be flushed
at any moment (interrupt, exception, task switching), even if not
triggered by QEMU. The probability for this to happen is very low, and
depends on cache size and associativiy, machine load, interrupts, so the
symptoms are might happen randomly.
This requirement is currently not followed in tcg/arm, for the
load/store code, which basically has the following structure:
1) tlb access code is written
2) conditional fast path code is written
3) branch is written with a temporary target
4) slow path code is written
5) branch target is updated
The cache lines corresponding to the retranslated code is not flushed
after code retranslation as the generated code is supposed to be the
same. However if the cache line corresponding to the branch instruction
is flushed between step 3 and 5, and is not flushed again before the
code is executed again, the branch target is wrong. In the guest, the
symptoms are MMU page fault at a random addresses, which leads to
kernel page fault or segmentation faults.
The patch fixes this issue by avoiding writing the branch target until
it is known, that is by writing only the branch instruction first, and
later only the offset.
This fixes booting linux guests on ARM hosts (tested: arm, i386, mips,
mipsel, sh4, sparc).
Acked-by: Edgar E. Iglesias <edgar.iglesias@gmail.com>
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
The usermode version of qemu_ld doesn't used mem_index,
leading to set-but-not-used warnings.
Signed-off-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Edgar E. Iglesias <edgar@axis.com>
A typo in the usermode address calculation path; R3 used where R2 needed.
Signed-off-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Edgar E. Iglesias <edgar@axis.com>
Use ld4 not ld8 for reading the tlb of 32-bit targets.
Signed-off-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Edgar E. Iglesias <edgar@axis.com>
The port was not properly merged following
86feb1c860
Signed-off-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Edgar E. Iglesias <edgar@axis.com>
Fix compilation error when GUEST_BASE is not defined.
Signed-off-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Edgar E. Iglesias <edgar@axis.com>
The arguments to tcg_gen_helper32 for these functions were not
updated correctly in rev 2bece2c883.
Signed-off-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Edgar E. Iglesias <edgar@axis.com>
fprintf_function uses format checking with GCC_FMT_ATTR.
Cc: Blue Swirl <blauwirbel@gmail.com>
Signed-off-by: Stefan Weil <weil@mail.berlios.de>
Signed-off-by: Blue Swirl <blauwirbel@gmail.com>
When qemu is configured with --enable-debug-tcg,
gcc throws this warning (or error with -Werror):
tcg/tcg.c:1030: error: comparison of unsigned expression >= 0 is always true
Fix it by removing the >= 0 part.
The type cast to 'unsigned' catches negative values of op
(which should never happen).
This is a modification of Hollis Blanchard's patch.
Cc: Hollis Blanchard <hollis@penguinppc.org>
Cc: Blue Swirl <blauwirbel@gmail.com>
Signed-off-by: Stefan Weil <weil@mail.berlios.de>
Signed-off-by: Blue Swirl <blauwirbel@gmail.com>
3b6dac3416 apparently broke the ppc64 TCG target
compilation in the code path without guest base.
Reverting this line fixes the build.
Signed-off-by: Andreas F?rber <andreas.faerber@web.de>
Cc: malc <av1474@comtv.ru>
Cc: Aurelien Jarno <aurelien@aurel32.net>
Cc: Richard Henderson <rth@twiddle.net>
Signed-off-by: malc <av1474@comtv.ru>
5da79c86a3 broke compilation on Mac OS X v10.5 ppc.
Apple's GCC 4.0.1 does not define _CALL_DARWIN. Recognize __APPLE__ again as well.
Signed-off-by: Andreas F?rber <andreas.faerber@web.de>
Cc: malc <av1474@comtv.ru>
Cc: Paolo Bonzini <pbonzini@redhat.com>
Cc: J?rgen Lock <nox@jelal.kn-bremen.de>
Cc: Stefan Weil <weil@mail.berlios.de>
Signed-off-by: malc <av1474@comtv.ru>
Original patch from Ulrich Hecht, further work from Alexander Graf
and Richard Henderson.
Cc: Ulrich Hecht <uli@suse.de>
Cc: Alexander Graf <agraf@suse.de>
Signed-off-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
We need not reserve the register unless we're going to use it.
Signed-off-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: malc <av1474@comtv.ru>
Some hosts (amd64, ia64) have an ABI that ignores the high bits
of the 64-bit register when passing 32-bit arguments. Others
require the value to be properly sign-extended for the type.
I.e. "int32_t" must be sign-extended and "uint32_t" must be
zero-extended to 64-bits.
To effect this, extend the "sizemask" parameter to tcg_gen_callN
to include the signedness of the type of each parameter. If the
tcg target requires it, extend each 32-bit argument into a 64-bit
temp and pass that to the function call.
This ABI feature is required by sparc64, ppc64 and s390x.
Signed-off-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
Before gcc 4.2, __builtin___clear_cache doesn't exist, and
afterward the gcc s390 backend implements it as nothing.
Signed-off-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
Both tcg_target_init and tcg_target_qemu_prologue
are unused outside of tcg.c.
Signed-off-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
Mirror tcg_out_movi in having a TYPE parameter. This allows x86_64
to perform the move at the proper width, which may elide a REX prefix.
Introduce a TCG_TYPE_REG enumerator to represent the "native width"
of the host register, and to distinguish the usage from "pointer data"
as represented by the existing TCG_TYPE_PTR.
Update all targets to match.
Signed-off-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
Make fallthru be TLB hit and branch be TLB miss. Doing this
both improves branch prediction and will allow further cleanup.
Signed-off-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
Splitting out these functions will allow further cleanups.
Signed-off-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
Include it in the opcode as an extension, as with P_EXT
or the REX bits in the x86-64 port.
Signed-off-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
(1) The output registers were not marked call-clobbered, even though
they can be modified by called functions.
(2) The thread pointer was not marked reserved.
(3) R4-R6 are call-saved, but not saved by the prologue. Rather than
save them, mark them reserved so that we don't use them.
Signed-off-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
Now that the prologue is generated after GUEST_BASE is fixed,
we can load it as an immediate, and also avoid reserving the
register if it isn't necessary.
Signed-off-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
This will allow backends to make intelligent choices about how
to implement GUEST_BASE.
Signed-off-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
The result is shorter than the mov+add that TCG would
otherwise generate for us.
Signed-off-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
Implement full modrm+sib addressing mode processing.
Use that in qemu_ld/st to output the LEA.
Signed-off-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
Define OPC_GRP3 and EXT3_FOO to match. Use them instead of
bare constants.
Define OPC_GRP5 and rename the existing EXT_BAR to EXT5_BAR to
make it clear which extension should be used with which opcode.
Signed-off-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
Define OPC_CALL_Jz, generated by tcg_out_calli; use the later
throughout. Unify the calls within qemu_st; adjust the stack
with a single pop if applicable.
Define and use EXT_CALLN_Ev for indirect calls.
Signed-off-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
Move tcg_out_push/pop up in the file so that they can be used
by qemu_ld/st. Define a tcg_out_pushi to be used as well.
Signed-off-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
