Peter Maydell [Tue, 7 Jan 2014 17:19:13 +0000 (17:19 +0000)]
target-arm: Ignore most exceptions from scalbn when doing fixpoint conversion
The VFP fixed point conversion helpers first call float_scalbn and
then convert the result to an integer. This scalbn operation may
set floating point exception flags for:
* overflow & inexact (if it overflows to infinity)
* input denormal squashed to zero
* output denormal squashed to zero
Of these, we only care about the input-denormal flag, since
the output of the whole scale-and-convert operation will be
an integer (so squashed-output-denormal and overflow don't
apply). Suppress the others by saving the pre-scalb exception
flags and only copying across a potential input-denormal flag.
Will Newton [Tue, 7 Jan 2014 17:19:13 +0000 (17:19 +0000)]
target-arm: Rename A32 VFP conversion helpers
The VFP conversion helpers for A32 round to zero as this is the only
rounding mode supported. Rename these helpers to make it clear that
they round to zero and are not suitable for use in the AArch64 code.
Will Newton [Tue, 7 Jan 2014 17:19:13 +0000 (17:19 +0000)]
target-arm: Prepare VFP_CONV_FIX helpers for A64 uses
Make the VFP_CONV_FIX helpers a little more flexible in
preparation for the A64 uses. This requires two changes:
* use the correct softfloat conversion function based on itype
rather than always the int32 one; this is possible now that
softfloat provides int16 versions and necessary for the
future conversion-to-int64 A64 variants. This also allows
us to drop the awkward 'sign' macro argument.
* split the 'fsz' argument which currently controls both
width of the input float type and width of the output
integer type into two; this will allow us to specify the
A64 64-bit-int-to-single conversion function, where the
two widths are different.
We can also drop the (itype##_t) cast now that softfloat
guarantees that all the itype##_to_float* functions take
an integer argument of exactly the correct type.
Peter Maydell [Tue, 7 Jan 2014 17:19:12 +0000 (17:19 +0000)]
softfloat: Add support for ties-away rounding
IEEE754-2008 specifies a new rounding mode:
"roundTiesToAway: the floating-point number nearest to the infinitely
precise result shall be delivered; if the two nearest floating-point
numbers bracketing an unrepresentable infinitely precise result are
equally near, the one with larger magnitude shall be delivered."
Implement this new mode (it is needed for ARM). The general principle
is that the required code is exactly like the ties-to-even code,
except that we do not need to do the "in case of exact tie clear LSB
to round-to-even", because the rounding operation naturally causes
the exact tie to round up in magnitude.
Peter Maydell [Tue, 7 Jan 2014 17:19:12 +0000 (17:19 +0000)]
softfloat: Refactor code handling various rounding modes
Refactor the code in various functions which calculates rounding
increments given the current rounding mode, so that instead of a
set of nested if statements we have a simple switch statement.
This will give us a clean place to add the case for the new
tiesAway rounding mode.
Peter Maydell [Tue, 7 Jan 2014 17:19:11 +0000 (17:19 +0000)]
softfloat: Factor out RoundAndPackFloat16 and NormalizeFloat16Subnormal
In preparation for adding conversions between float16 and float64,
factor out code currently done inline in the float16<=>float32
conversion functions into functions RoundAndPackFloat16 and
NormalizeFloat16Subnormal along the lines of the existing versions
for the other float types.
Note that we change the handling of zExp from the inline code
to match the API of the other RoundAndPackFloat functions; however
we leave the positioning of the binary point between bits 22 and 23
rather than shifting it up to the high end of the word.
This patch re-implements the routine to use the float64_to_uint64_round_to_zero
routine. If saturation occurs we ignore any flags set by the
conversion function and raise only Invalid.
This contribution can be licensed under either the softfloat-2a or -2b
license.
This patch re-implements float64_to_uint32 to re-use the
float64_to_uint64 routine (instead of float64_to_int64). For the
saturation case, we ignore any flags which the conversion routine
has set and raise only the invalid flag.
This contribution can be licensed under either the softfloat-2a or -2b
license.
Peter Maydell [Tue, 7 Jan 2014 17:17:50 +0000 (17:17 +0000)]
softfloat: Fix factor 2 error for scalbn on denormal inputs
If the input to float*_scalbn() is denormal then it represents
a number 0.[mantissabits] * 2^(1-exponentbias) (and the actual
exponent field is all zeroes). This means that when we convert
it to our unpacked encoding the unpacked exponent must be one
greater than for a normal number, which represents
1.[mantissabits] * 2^(e-exponentbias) for an exponent field e.
This meant we were giving answers too small by a factor of 2 for
all denormal inputs.
Note that the float-to-int routines also have this behaviour
of not adjusting the exponent for denormals; however there it is
harmless because denormals will all convert to integer zero anyway.
Peter Maydell [Tue, 7 Jan 2014 17:17:49 +0000 (17:17 +0000)]
softfloat: Only raise Invalid when conversions to int are out of range
We implement a number of float-to-integer conversions using conversion
to an integer type with a wider range and then a check against the
narrower range we are actually converting to. If we find the result to
be out of range we correctly raise the Invalid exception, but we must
also suppress other exceptions which might have been raised by the
conversion function we called.
This won't throw away exceptions we should have preserved, because for
the 'core' exception flags the IEEE spec mandates that the only valid
combinations of exception that can be raised by a single operation are
Inexact + Overflow and Inexact + Underflow. For the non-IEEE softfloat
flag for input denormals, we can guarantee that that flag won't have
been set for out of range float-to-int conversions because a squashed
denormal by definition goes to plus or minus zero, which is always in
range after conversion to integer zero.
This bug has been fixed for some of the float-to-int conversion routines
by previous patches; fix it for the remaining functions as well, so
that they all restore the pre-conversion status flags prior to raising
Invalid.
Peter Maydell [Tue, 7 Jan 2014 17:17:49 +0000 (17:17 +0000)]
softfloat: Make the int-to-float functions take exact-width types
Currently the int-to-float functions take types which are specified
as "at least X bits wide", rather than "exactly X bits wide". This is
confusing and unhelpful since it means that the callers have to include
an explicit cast to [u]intXX_t to ensure the correct behaviour. Fix
them all to take the exactly-X-bits-wide types instead.
Note that this doesn't change behaviour at all since at the moment
we happen to define the 'int32' and 'uint32' types as exactly 32 bits
wide, and the 'int64' and 'uint64' types as exactly 64 bits wide.
Peter Maydell [Tue, 7 Jan 2014 17:17:49 +0000 (17:17 +0000)]
softfloat: Add 16 bit integer to float conversions
Add the float to 16 bit integer conversion routines. These can be
trivially implemented in terms of the int32_to_float* routines, but
providing them makes our API more symmetrical and can simplify callers.
Will Newton [Tue, 7 Jan 2014 17:17:48 +0000 (17:17 +0000)]
softfloat: Add float to 16bit integer conversions.
ARMv8 requires support for converting 32 and 64bit floating point
values to signed and unsigned 16bit integers.
Signed-off-by: Will Newton <[email protected]>
[PMM: updated not to incorrectly set Inexact for Invalid inputs] Signed-off-by: Peter Maydell <[email protected]> Reviewed-by: Richard Henderson <[email protected]>
Peter Maydell [Mon, 6 Jan 2014 11:47:21 +0000 (11:47 +0000)]
softfloat: Fix exception flag handling for float32_to_float16()
Our float32 to float16 conversion routine was generating the correct
numerical answers, but not always setting the right set of exception
flags. Fix this, mostly by rearranging the code to more closely
resemble RoundAndPackFloat*, and in particular:
* non-IEEE halfprec always raises Invalid for input NaNs
* we need to check for the overflow case before underflow
* we weren't getting the tininess-detected-after-rounding
case correct (somewhat academic since only ARM uses halfprec
and it is always tininess-detected-before-rounding)
* non-IEEE halfprec overflow raises only Invalid, not
Invalid + Inexact
* we weren't setting Inexact when we should
Also add some clarifying comments about what the code is doing.
Christoffer Dall [Sat, 21 Dec 2013 06:09:33 +0000 (22:09 -0800)]
hw: arm_gic: Introduce gic_set_priority function
To make the code slightly cleaner to look at and make the save/restore
code easier to understand, introduce this function to set the priority of
interrupts.
Christoffer Dall [Sat, 21 Dec 2013 06:09:32 +0000 (22:09 -0800)]
arm_gic: Rename GIC_X_TRIGGER to GIC_X_EDGE_TRIGGER
TRIGGER can really mean mean anything (e.g. was it triggered, is it
level-triggered, is it edge-triggered, etc.). Rename to EDGE_TRIGGER to
make the code comprehensible without looking up the data structure.
introduces an array aarch64_cpus which is zero
size if this code is built without CONFIG_USER_ONLY.
In particular an attempt to iterate over this array produces a warning
under gcc 4.8.2:
CC aarch64-softmmu/target-arm/cpu64.o
/scm/qemu/target-arm/cpu64.c: In function ‘aarch64_cpu_register_types’:
/scm/qemu/target-arm/cpu64.c:124:5: error: comparison of unsigned
expression < 0 is always false [-Werror=type-limits]
for (i = 0; i < ARRAY_SIZE(aarch64_cpus); i++) {
^
cc1: all warnings being treated as errors
This is the result of ARRAY_SIZE being an unsigned type,
causing "i" to be promoted to unsigned int as well.
As zero size arrays are a gcc extension, it seems
cleanest to add a dummy element with NULL name,
and test for it during registration.
Don't conditionalise GEM instantiation on networking attachments. The
device should always be present even if not attached to a network.
This allows for probing of the device by expectant guests (such as
OS's). This is needed because sysbus (or AXI in Xilinx's real hw case)
is not self identifying so the guest has no dynamic way of detecting
device absence.
Also allows for testing of the GEM in loopback mode with -net none.
If the UART back-end blocks, buffer in the Tx FIFO to try again later.
This stops the IO-thread busy waiting on char back-ends (which causes
all sorts of performance problems).
The can_receive logic was only taking into account the RxFIFO
occupancy. RxFIFO population is only used for the echo and normal modes
however. Improve the logic to correctly return the true number of
receivable characters based on the current mode:
Normal mode: RxFIFO vacancy.
Remote loopback: TxFIFO vacancy.
Echo mode: The min of the TxFIFO and RxFIFO vacancies.
Local Loopback: Return non-zero (to implement droppage)
char/cadence_uart: Remove TX timer & add TX FIFO state
This tx timer implementation is flawed. Despite the controller
attempting to time the guest visable assertion of the TX-empty status
bit (and corresponding interrupt) the controller is still transmitting
characters instantaneously. There is also no sense of multiple character
delay.
The only side effect of this timer is assertion of tx-empty status. So
just remove the timer completely and hold tx-empty as permanently
asserted (its reset status). This matches the actual behaviour of
instantaneous transmission.
While we are VMSD version bumping, add the tx_fifo as device state to
prepare for upcomming TxFIFO flow control. Implement the interrupt
generation logic for the TxFIFO occupancy.
Some (interrupt) status register bits relating to the TxFIFO path were
not defined. Define them. This prepares support for proper Tx data path
flow control.
The status register bits are always pure functions of other device
state. Move the generation of these bits to the update_status()
function to simplify. Makes developing much easier as theres now no need
to recheck status bits on all the changes to rx/tx fifo state.
This should be rechecked on bus write accesses as such accesses may
change the underlying state that generates the interrupt. Particular
relevant for when the guest touches the interrupt status or mask.
Alexander Graf [Sat, 4 Jan 2014 22:15:51 +0000 (22:15 +0000)]
target-arm: Give the FPSCR rounding modes names
When setting rounding modes we currently just hardcode the numeric values
for rounding modes in a big switch statement.
With AArch64 support coming, we will need to refer to these rounding modes
at different places throughout the code though, so let's better give them
names so we don't get confused by accident.
This patch adds emulation for the "Floating-point data-processing (3 source)"
group of instructions.
Signed-off-by: Alexander Graf <[email protected]>
[WN: Commit message tweak, merged single and double precision patches.
Implement using muladd as suggested by Richard Henderson.] Signed-off-by: Will Newton <[email protected]>
[PMM: pull field decode up a level, use register accessors] Signed-off-by: Peter Maydell <[email protected]> Reviewed-by: Richard Henderson <[email protected]>
This patch adds emulation for the "Floating-point data-processing (2 source)"
group of instructions.
Signed-off-by: Alexander Graf <[email protected]>
[WN: Commit message tweak, merge single and double precision patches. Rebase
and update to new infrastructure. Incorporate FMIN/FMAX support patch by
Michael Matz.] Signed-off-by: Will Newton <[email protected]>
[PMM:
* added convenience accessors for FP s and d regs
* pulled the field decode and opcode validity check up a level] Signed-off-by: Peter Maydell <[email protected]> Reviewed-by: Richard Henderson <[email protected]>
Peter Maydell [Sat, 4 Jan 2014 22:15:49 +0000 (22:15 +0000)]
target-arm: Use VFP_BINOP macro for min, max, minnum, maxnum
Use the VFP_BINOP macro to provide helpers for min, max, minnum
and maxnum, rather than hand-rolling them. (The float64 max
version is not used by A32 but will be needed for A64.)
Peter Maydell [Sat, 4 Jan 2014 22:15:49 +0000 (22:15 +0000)]
target-arm: A64: Fix vector register access on bigendian hosts
The A64 128 bit vector registers are stored as a pair of
uint64_t values in the register array. This means that if
we're directly loading or storing a value of size less than
64 bits we must adjust the offset appropriately to account
for whether the host is bigendian or not. Provide utility
functions to abstract away the offsetof() calculations for
the FP registers.
For do_fp_st() we can sidestep most of the issues for 64 bit
and smaller reg-to-mem transfers by always doing a 64 bit
load from the register and writing just the piece we need
to memory.
The AArch64 linux-user support was written before but merged after
commit 4ce6243dc621 which cleaned up the handling of the clone()
syscall argument order, so we failed to notice that AArch64 also needs
TARGET_CLONE_BACKWARDS to be defined. Add this define so that clone
and fork syscalls work correctly.
Michael Matz [Sat, 4 Jan 2014 22:15:47 +0000 (22:15 +0000)]
target-arm: A64: support for ld/st/cl exclusive
This implement exclusive loads/stores for aarch64 along the lines of
arm32 and ppc implementations. The exclusive load remembers the address
and loaded value. The exclusive store throws an an exception which uses
those values to check for equality in a proper exclusive region.
This is not actually the architecture mandated semantics (for either
AArch32 or AArch64) but it is close enough for typical guest code
sequences to work correctly, and saves us from having to monitor all
guest stores. It's fairly easy to come up with test cases where we
don't behave like hardware - we don't for example model cache line
behaviour. However in the common patterns this works, and the existing
32 bit ARM exclusive access implementation has the same limitations.
AArch64 also implements new acquire/release loads/stores (which may be
either exclusive or non-exclusive). These imposes extra ordering
constraints on memory operations (ie they act as if they have an implicit
barrier built into them). As TCG is single-threaded all our barriers
are no-ops, so these just behave like normal loads and stores.
Peter Maydell [Sat, 4 Jan 2014 22:15:47 +0000 (22:15 +0000)]
target-arm: Widen exclusive-access support struct fields to 64 bits
In preparation for adding support for A64 load/store exclusive instructions,
widen the fields in the CPU state struct that deal with address and data values
for exclusives from 32 to 64 bits. Although in practice AArch64 and AArch32
exclusive accesses will be generally separate there are some odd theoretical
corner cases (eg you should be able to do the exclusive load in AArch32, take
an exception to AArch64 and successfully do the store exclusive there), and it's
also easier to reason about.
The changes in semantics for the variables are:
exclusive_addr -> extended to 64 bits; -1ULL for "monitor lost",
otherwise always < 2^32 for AArch32
exclusive_val -> extended to 64 bits. 64 bit exclusives in AArch32 now
use the high half of exclusive_val instead of a separate exclusive_high
exclusive_high -> is no longer used in AArch32; extended to 64 bits as
it will be needed for AArch64's pair-of-64-bit-values exclusives.
exclusive_test -> extended to 64 bits, as it is an address. Since this is
a linux-user-only field, in arm-linux-user it will always have the top
32 bits zero.
exclusive_info -> stays 32 bits, as it is neither data nor address, but
simply holds register indexes etc. AArch64 will be able to fit all its
information into 32 bits as well.
Note that the refactoring of gen_store_exclusive() coincidentally fixes
a minor bug where ldrexd would incorrectly update the first CPU register
even if the load for the second register faulted.
Peter Maydell [Sat, 4 Jan 2014 22:15:45 +0000 (22:15 +0000)]
target-arm: Widen thread-local register state fields to 64 bits
The common pattern for system registers in a 64-bit capable ARM
CPU is that when in AArch32 the cp15 register is a view of the
bottom 32 bits of the 64-bit AArch64 system register; writes in
AArch32 leave the top half unchanged. The most natural way to
model this is to have the state field in the CPU struct be a
64 bit value, and simply have the AArch32 TCG code operate on
a pointer to its lower half.
For aarch64-linux-user the only registers we need to share like
this are the thread-local-storage ones. Widen their fields to
64 bits and provide the 64 bit reginfo struct to make them
visible in AArch64 state. Note that minor cleanup of the AArch64
system register encoding space means We can share the TPIDR_EL1
reginfo but need split encodings for TPIDR_EL0 and TPIDRRO_EL0.
Since we're touching almost every line in QEMU that uses the
c13_tls* fields in this patch anyway, we take the opportunity
to rename them in line with the standard ARM architectural names
for these registers.
Peter Maydell [Sat, 4 Jan 2014 22:15:45 +0000 (22:15 +0000)]
target-arm: A64: Implement MRS/MSR/SYS/SYSL
The AArch64 equivalent of the traditional AArch32
cp15 coprocessor registers is the set of instructions
MRS/MSR/SYS/SYSL, which cover between them both true
system registers and the "operations with side effects"
such as cache maintenance which in AArch32 are mixed
in with other cp15 registers. Implement these instructions
to look in the cpregs hashtable for the register or
operation.
Since we don't yet populate the cpregs hashtable with
any registers with the "AA64" bit set, everything will
still UNDEF at this point.
MSR/MRS is the first user of is_jmp = DISAS_UPDATE, so
fix an infelicity in its handling where the main loop
was requiring the caller to do the update of PC rather
than just doing it itself.
Peter Maydell [Sat, 4 Jan 2014 22:15:44 +0000 (22:15 +0000)]
target-arm: Remove ARMCPU/CPUARMState from cpregs APIs used by decoder
The cpregs APIs used by the decoder (get_arm_cp_reginfo() and
cp_access_ok()) currently take either a CPUARMState* or an ARMCPU*.
This is problematic for the A64 decoder, which doesn't pass the
environment pointer around everywhere the way the 32 bit decoder
does. Adjust the parameters these functions take so that we can
copy only the relevant info from the CPUARMState into the
DisasContext and then use that.
Peter Maydell [Sat, 4 Jan 2014 22:15:44 +0000 (22:15 +0000)]
target-arm: Update generic cpreg code for AArch64
Update the generic cpreg support code to also handle AArch64:
AArch64-visible registers coexist in the same hash table with
AArch32-visible ones, with a bit in the hash key distinguishing
them.
Peter Maydell [Sun, 22 Dec 2013 22:32:30 +0000 (22:32 +0000)]
target-arm: Pull "add one cpreg to hashtable" into its own function
define_one_arm_cp_reg_with_opaque() has a set of nested loops which
insert a cpreg entry into the hashtable for each of the possible
opc/crn/crm values allowed by wildcard specifications. We're about
to add an extra loop to this nesting, so pull the core of the loop
(which adds a single entry to the hashtable) out into its own
function for clarity.
Peter Maydell [Mon, 23 Dec 2013 23:27:31 +0000 (23:27 +0000)]
target-arm: A64: implement FMOV
Implement FMOV, ie non-converting moves between general purpose
registers and floating point registers. This is a subtype of
the floating point <-> integer instruction class.
Alexander Graf [Mon, 23 Dec 2013 23:27:30 +0000 (23:27 +0000)]
target-arm: A64: add support for 3 src data proc insns
This patch adds emulation for the "Data-processing (3 source)"
family of instructions, namely MADD, MSUB, SMADDL, SMSUBL, SMULH,
UMADDL, UMSUBL, UMULH.
Alex Bennée [Mon, 23 Dec 2013 23:27:29 +0000 (23:27 +0000)]
target-arm: A64: add support for add, addi, sub, subi
Implement the non-carry forms of addition and subtraction
(immediate, extended register and shifted register).
This includes the code to calculate NZCV if the instruction
calls for setting the flags.
Alex Bennée [Mon, 23 Dec 2013 23:27:29 +0000 (23:27 +0000)]
target-arm: A64: add support for ld/st with index
This adds support for the pre/post-index ld/st forms with immediate
offsets as well as the un-scaled immediate form (which are all
variations on the same 9-bit immediate instruction form).
Alexander Graf [Fri, 20 Dec 2013 10:01:50 +0000 (11:01 +0100)]
PPC: Fix compilation with TCG debug
The recent VSX patches broken compilation of QEMU when configurated
with --enable-debug, as it was treating "target long" TCG variables
as "i64" which is not true for 32bit targets.
This patch fixes all the places that the compiler has found to use
the correct variable type and if necessary manually cast.
Aurelien Jarno [Sat, 21 Dec 2013 15:46:07 +0000 (16:46 +0100)]
Merge tag 'signed-s390-for-upstream' of git://github.com/agraf/qemu
Patch queue for s390 - 2013-12-18
This covers mostly minor bug fixes and implements the SIGP START
hypercall which allows to start a remote CPU without changing its
state.
Cornelia Huck (1):
s390x/kvm: Fix diagnose handling.
Thomas Huth (7):
s390x/kvm: Removed duplicated SIGP defines
s390x/kvm: Removed s390_store_status stub
s390x/kvm: Fix coding style in handle_sigp()
s390x/kvm: Implemented SIGP START
s390x/kvm: Simplified the calculation of the SIGP order code
s390x/kvm: Fixed condition code for unknown SIGP orders
s390x/ioinst: CHSC has to set a condition code
* tag 'signed-s390-for-upstream' of git://github.com/agraf/qemu:
s390x/ioinst: CHSC has to set a condition code
s390x/kvm: Fixed condition code for unknown SIGP orders
s390x/kvm: Simplified the calculation of the SIGP order code
s390x/kvm: Implemented SIGP START
s390x/kvm: Fix coding style in handle_sigp()
s390x/kvm: Removed s390_store_status stub
s390x/kvm: Removed duplicated SIGP defines
s390x/kvm: Fix diagnose handling.
Anthony Liguori [Fri, 20 Dec 2013 01:03:17 +0000 (17:03 -0800)]
Merge remote-tracking branch 'agraf/tags/signed-ppc-for-upstream' into staging
Patch queue for ppc - 2013-12-20
Alexander Graf (3):
PPC: Use default pci bus name for grackle and heathrow
roms: Flush icache when writing roms to guest memory
PPC: Add VSX to hflags
Alexey Kardashevskiy (5):
powerpc: add PVR mask support
target-ppc: move POWER7+ to a separate family
spapr-rtas: replace return code constants with macros
spapr-rtas: add ibm, (get|set)-system-parameter
spapr: make sure RMA is in first mode of first memory node
Greg Kurz (1):
target-ppc: add stubs for KVM breakpoints
Paolo Bonzini (1):
spapr: tie spapr-nvram to -pflash
Paul Mackerras (1):
spapr: limit numa memory regions by ram size
* agraf/tags/signed-ppc-for-upstream: (32 commits)
spapr: limit numa memory regions by ram size
spapr: make sure RMA is in first mode of first memory node
device_tree: qemu_fdt_setprop: Rename val_array arg
device_tree: s/qemu_devtree/qemu_fdt globally
PPC: Add VSX to hflags
Add xxsldwi
Add xxspltw
Add xxsel
Add xxmrgh/xxmrgl
Add Power7 VSX Logical Instructions
Add VSX Vector Move Instructions
Add VSX Scalar Move Instructions
roms: Flush icache when writing roms to guest memory
spapr: tie spapr-nvram to -pflash
PPC: Use default pci bus name for grackle and heathrow
spapr-rtas: add ibm, (get|set)-system-parameter
spapr-rtas: replace return code constants with macros
target-ppc: move POWER7+ to a separate family
Add stxvw4x
Add stxsdx
...
spapr: make sure RMA is in first mode of first memory node
The SPAPR specification says that the RMA starts at the LPAR's logical
address 0 and is the first logical memory block reported in
the LPAR’s device tree.
So SLOF only maps the first block and that block needs to span
the full RMA.
This makes sure that the RMA area is where SLOF expects it.
Looking at the implementation, this doesn't really have a lot to do
with arrays. Its just a pointer to a buffer and is passed through
to the wrapped fn (qemu_fdt_setprop) unchanged. So rename to make it
consistent with libfdt, which in the wrapped function just calls it
"val".
Alexander Graf [Wed, 18 Dec 2013 08:21:02 +0000 (09:21 +0100)]
PPC: Add VSX to hflags
We generate different code depending on whether MSR_VSX is set or
clear, so it needs to be part of our hflags too which indicate whether
we're still in the same translation block cache bucket.
Tom Musta [Fri, 1 Nov 2013 13:21:21 +0000 (08:21 -0500)]
Add xxsel
This patch adds the VSX Select (xxsel) instruction.
The xxsel instruction has four VSR operands. Thus the xC
instruction decoder is added.
The xxsel instruction is massively overloaded in the opcode
table since only bits 26 and 27 are opcode bits. This
overloading is done in matrix fashion with two macros
(GEN_XXSEL_ROW and GEN_XX_SEL).
A common generator macro (VSX_SCALAR_MOVE) is added since these
instructions vary only slightly from each other.
Macros to support VSX XX2 and XX3 form opcodes are also added.
These macros handle the overloading of "opcode 2" space (instruction
bits 26:30) caused by AX and BX bits (29 and 30, respectively).
V3: Per feedback from Paolo Bonzini, moved the sign mask into a
temporary and used andc.
Alexander Graf [Wed, 11 Dec 2013 13:17:44 +0000 (14:17 +0100)]
roms: Flush icache when writing roms to guest memory
We use the rom infrastructure to write firmware and/or initial kernel
blobs into guest address space. So we're basically emulating the cache
off phase on very early system bootup.
That phase is usually responsible for clearing the instruction cache for
anything it writes into cachable memory, to ensure that after reboot we
don't happen to execute stale bits from the instruction cache.
So we need to invalidate the icache every time we write a rom into guest
address space. We do not need to do this for every DMA since the guest
expects it has to flush the icache manually in that case.
This fixes random reboot issues on e5500 (booke ppc) for me.
Paolo Bonzini [Fri, 22 Nov 2013 09:27:40 +0000 (10:27 +0100)]
spapr: tie spapr-nvram to -pflash
spapr-nvram's drive property is currently connected to a non-existent
"-machine nvram=<drivename>" option. Instead, tie it to -pflash like
other non-volatile RAM devices. This provides the following possibilities
for adding a backend for the sPAPR non-volatile RAM:
Alexander Graf [Wed, 4 Dec 2013 11:42:32 +0000 (12:42 +0100)]
PPC: Use default pci bus name for grackle and heathrow
There's no good reason to call our bus "pci" rather than let the default
bus name take over ("pci.0").
The big downside to calling it different from anyone else is that tools
that pass -device get confused. They are looking for a bus "pci.0" rather
than "pci".
To make life easier for everyone, let's just drop the name override.
This adds very basic handlers for ibm,get-system-parameter and
ibm,set-system-parameter RTAS calls.
The only parameter handled at the moment is
"platform-processor-diagnostics-run-mode" which is always disabled and
does not support changing. This is expected to make
"ppc64_cpu --run-mode=1" happy.
So far POWER7+ was a part of POWER7 family. However it has a different
PVR base value so in order to support PVR masks, it needs a separate
family class.
This adds a new family class, PVR base and mask values and moves
Power7+ v2.1 CPU to a new family. The class init function is copied
from the POWER7 family.
This defines a firmware name for the new family as "PowerPC,POWER7+"
instead of previously used "PowerPC,POWER7" from the POWER7 family.
The reason for that is that the Sapphire firmware (a h0st firmware)
uses "PowerPC,POWER7+" already and since no specification defines
exactly the CPU nodes naming in the device tree, we better stay
in sync with the host firmware.
projects / qemu.git / log