SH4 is always using softfloat, so it's possible to have helpers directly
taking float32 or float64 value. This allow to get rid of conversions
through CPU_{Float,Double}U.
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
tcg_gen_exit_tb takes a parameter of type tcg_target_long,
so the type casts of pointer to long should be replaced by
type casts of pointer to tcg_target_long (suggested by Blue Swirl).
These changes are needed for build environments where
sizeof(long) != sizeof(void *), especially for w64.
Signed-off-by: Stefan Weil <weil@mail.berlios.de>
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
All implementations are now the same, and there is only one caller,
so inline the function there.
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Blue Swirl <blauwirbel@gmail.com>
All targets except SH4 have the same cpu_halted() routine, and it has
only one caller. It is therefore a good candidate for inlining.
The difference is the handling of the intr_at_halt, which is necessary
to ignore SR.BL when sleeping. Move intr_at_halt handling out of it, by
setting this variable while executing the sleep instruction, and
clearing it when the CPU has been woken-up by an interrupt, whatever the
state of SR.BL. Also rename this variable in_sleep.
Cc: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
Update the PTEH register to contain the VPN at which an MMU
exception occured as specified by the SH4 reference.
Signed-off-by: Alexandre Courbot <gnurou@gmail.com>
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
Exception index of address read error should be 0x0e0.
Signed-off-by: Alexandre Courbot <gnurou@gmail.com>
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
In cpu_sh4_invalidate_tlb, the UTLB was invalidated twice and the
ITLB left unchaged, probably because of some unfortunate copy/paste.
Signed-off-by: Alexandre Courbot <gnurou@gmail.com>
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
Fix wrong usage of ! and & in MMU related functions. Thanks to Blue
Swirl for reporting the issue.
Reported-by: Blue Swirl <blauwirbel@gmail.com>
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
When a TCG variable is anded with a value and the compared with the same
value, we can simply invert the comparison and compare it with 0. The
generated code is smaller.
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
As exception is not the normal path, don't bother saving PC, before
raising one, instead rely on code retranslation to get the CPU state.
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
Add the ftrv XMTRX,FVn instruction, which computes the 4-row x 4-column
matrix XMTRX by the 4-dimensional vector FVn.
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
Add the fipr FVm,FVn instruction, which computes the inner products of
a 4-dimensional single precision floating-point vector.
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
FPU exception support where not implemented on SH4. Implement them by
clearing the softfloat exceptions flags before an FP instruction (the
SH4 FPU also clear them before an instruction), and calling a function
to update the FPSCR register after an FP instruction. This function
update the corresponding FPSCR bits (both flags and cumulative flags)
and trigger exception if enabled.
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
When the FPSCR.DN bit is set, the SH4 FPU treat denormalized numbers as
zero. Enable the corresponding softfloat option when this bit is set.
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
Illegal instructions in a slot delay should generate a slot illegal
instruction exception instead of an illegal instruction exception.
The current PC should be saved before generating such an exception,
but should not be corrected if in a delay slot, given it's already
done in the exception handler do_interrupt().
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
SH4 is using 16-bit instructions which means most of the constants are
loaded through a constant pool at the end of the subroutine. The same
memory page is therefore accessed in exec and read mode.
With the current implementation, a QEMU TLB entry is set to read or
read/write mode after an UTLB search and to exec mode after an ITLB
search, which causes a lot of TLB exceptions to switch from read or
read/write to exec and vice versa.
This patch optimizes that by already setting the QEMU TLB entry in read
or read/write mode when an UTLB entry is copied into ITLB (during an
ITLB miss). This improve the emulation speed by about 14%.
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
Some Linux kernels seems to implement ITLB/UTLB flushing through by
writing all TLB entries through the memory mapped interface instead
of writing one to MMUCR.TI.
Implement memory mapped ITLB write interface so that such kernels can
boot. This fixes https://bugs.launchpad.net/bugs/700774 .
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
fprintf_function uses format checking with GCC_FMT_ATTR.
Format errors were fixed in
* target-i386/helper.c
* target-mips/translate.c
* target-ppc/translate.c
Cc: Blue Swirl <blauwirbel@gmail.com>
Signed-off-by: Stefan Weil <weil@mail.berlios.de>
Signed-off-by: Blue Swirl <blauwirbel@gmail.com>
The LDST macro is used to generate ldc and stc instructions that work with a
specific register. However, the SGR register only supports stc up to SH4A,
which supports both stc and ldc. This patch creates two sub-macros named LD
and ST that handle generating ldc and stc instructions separately, and
redeclares LDST to use these sub-macro.
Signed-off-by: Alexandre Courbot <gnurou@gmail.com>
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
When building with -DNDEBUG, assert(0) will not stop execution
so it must not be used for abnormal termination.
Use cpu_abort() when in CPU context, abort() otherwise.
Signed-off-by: Blue Swirl <blauwirbel@gmail.com>
QEMU uses a fixed page size for the CPU TLB. If the guest uses large
pages then we effectively split these into multiple smaller pages, and
populate the corresponding TLB entries on demand.
When the guest invalidates the TLB by virtual address we must invalidate
all entries covered by the large page. However the address used to
invalidate the entry may not be present in the QEMU TLB, so we do not
know which regions to clear.
Implementing a full vaiable size TLB is hard and slow, so just keep a
simple address/mask pair to record which addresses may have been mapped by
large pages. If the guest invalidates this region then flush the
whole TLB.
Signed-off-by: Paul Brook <paul@codesourcery.com>
Removes a set of ifdefs from exec.c.
Introduce TARGET_VIRT_ADDR_SPACE_BITS for all targets other
than Alpha. This will be used for page_find_alloc, which is
supposed to be using virtual addresses in the first place.
Signed-off-by: Richard Henderson <rth@twiddle.net>
env->exception_index should be cleared with -1, not 0.
See also 821b19fe92.
Spotted by Igor Kovalenko.
Signed-off-by: Blue Swirl <blauwirbel@gmail.com>
On SH4, the ITLB and UTLB configurations are memory mapped, so loading
ITLB entries from UTLB has to be simulated correctly. For that the QEMU
TLB has to be handle the execute (ITLB) and read/write permissions
(UTLB) seperately.
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
With the current code, the QEMU TLB is setup to match the read/write
mode of the MMU fault. This means when read access is done, the page
is setup in read-only mode. When the page is later accessed in write
mode, an MMU fault happened, and the page is switch in write-only
mode. This flip-flop causes a lot of calls to the MMU code and slow
down the emulation.
This patch changes the MMU emulation, so that the QEMU TLB is setup
to match the UTLB protection key. This impressively increase the
speed of the emulation.
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
There is an ITLB access violation if SR_MD=0 (user mode) while
the high bit of the protection key is 0 (priviledge mode).
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
The mem_idx is wrongly computed. As written in target-sh4/cpu.h, mode 0
corresponds to kernel mode (SR_MD = 1), while mode 1 corresponds to user
mode (SR_MD = 0).
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
