We can call tb_htable_lookup even when the tb_jmp_cache is completely
empty. Therefore, un-nest most of the code dependent on tb != NULL
from the read from the cache.
This improves the hit rate of lookup_tb_ptr; for instance, when booting
and immediately shutting down debian-arm, the hit rate improves from
93.2% to 99.4%.
Backports commit b97a879de980e99452063851597edb98e7e8039c from qemu
The new placement of the TB means that we can use one insn
to load the goto_tb destination directly from the TB.
Backports commit 308714e6bc945389c64faf1b9213e2c0d3f03391 from qemu
Since we're no longer using a direct branch, we have no
limit on the branch distance.
Backports commit acb0b292b6d0f49972dc98f742e79ed53973e438 from qemu
The new placement of the TB means that we can use one insn
to load the return value for exit_tb returning the TB pointer.
Backports commit cc74d332ff9a78684374847375ef63fc4bd10436 from qemu
We are partially initializing tb in tb_alloc. Instead, fully
initialize it in tb_gen_code, which is tb_alloc's only caller.
This saves an unnecessary write to tb->cflags.
Backports commit 2b48e10f888059a98043b4816769fa2a326a1d2c from qemu
Allocating an arbitrarily-sized array of tbs results in either
(a) a lot of memory wasted or (b) unnecessary flushes of the code
cache when we run out of TB structs in the array.
An obvious solution would be to just malloc a TB struct when needed,
and keep the TB array as an array of pointers (recall that tb_find_pc()
needs the TB array to run in O(log n)).
Perhaps a better solution, which is implemented in this patch, is to
allocate TB's right before the translated code they describe. This
results in some memory waste due to padding to have code and TBs in
separate cache lines--for instance, I measured 4.7% of padding in the
used portion of code_gen_buffer when booting aarch64 Linux on a
host with 64-byte cache lines. However, it can allow for optimizations
in some host architectures, since TCG backends could safely assume that
the TB and the corresponding translated code are very close to each
other in memory. See this message by rth for a detailed explanation:
https://lists.gnu.org/archive/html/qemu-devel/2017-03/msg05172.html
Subject: Re: GSoC 2017 Proposal: TCG performance enhancements
Backports commit 6e3b2bfd6af488a896f7936e99ef160f8f37e6f2 from qemu
Add helpers to gather cache info from the host at init-time.
For now, only export the host's I/D cache line sizes, which we
will use to improve cache locality to avoid false sharing.
Backports commit b255b2c8a5484742606e8760870ba3e14d0c9605 from qemu
V flag for subtraction is:
v = (res ^ src1) & (src1 ^ src2)
(see COMPUTE_CCR() in target/m68k/helper.c)
But gen_flush_flags() uses:
v = (res ^ src2) & (src1 ^ src2)
The problem has been found with the following program:
.global _start
_start:
move.l #-2147483648,%d0
subq.l #1,%d0
jvc 1f
move.l #1,%d1
move.l #1,%d0
trap #0
1:
move.l #0,%d1
move.l #1,%d0
trap #0
It works fine (exit(1)) on real hardware, and with "-singlestep".
"-singlestep" uses gen_helper_flush_flags(), whereas
without "-singlestep", V flag is computed directly in
gen_flush_flags().
This patch updates gen_flush_flags() to have the same result
as with gen_helper_flush_flags().
Backports commit 043b936ef6fe53396b3c6b8f5562ea3e238a071d from qemu
Running Windows with icount causes a crash in instruction of write cr.
This patch fixes it.
Reading and writing cr cause an icount read because there are called
cpu_get_apic_tpr and cpu_set_apic_tpr functions. So, there is need
gen_io_start()/gen_io_end() calls.
Backports commit 5b003a40bb1ab14d0398e91f03393d3c6b9577cd from qemu
This speeds up SMM switches. Later on it may remove the need to take
the BQL, and it may also allow to reuse code between TCG and KVM.
Backports commit f8c45c6550b9ff1e1f0b92709ff3213a79870879 from qemu
It really only plays with the dispatchers, so the parameter list does
not need that complexity. This helps for readability at least.
Backports commit 003a0cf2cd1828a1141a874428571267b117f765 from qemu
In theory this would re-enable usage of QEMU on an armv4 host.
Whether this is worthwhile is debatable -- we've been unconditionally
issuing the armv5t BX instruction in the prologue since 2011 without
complaint. Possibly we should simply require an armv6 host.
Backports commit 702a947484eb3e615183dafc93de590ab0679f60 from qemu
Instead of unconditionally exiting to the exec loop, use the
gen_jr helper to jump to the target if it is valid.
Perf impact: see next commit's log.
Backports commit fe62089563ffc6a42f16ff28a6b6be34d2697766 from qemu
Instead of unconditionally exiting to the exec loop, use the
lookup_and_goto_ptr helper to jump to the target if it is valid.
Perf impact: see next commit's log.
Backports commit 7ad55b4ffd982c80f26f7f3658138d94cdc678e8 from qemu
Instead of exporting goto_ptr directly to TCG frontends, export
tcg_gen_lookup_and_goto_ptr(), which calls goto_ptr with the pointer
returned by the lookup_tb_ptr() helper. This is the only use case
we have for goto_ptr and lookup_tb_ptr, so having this function is
very convenient. Furthermore, it trivially allows us to avoid calling
the lookup helper if goto_ptr is not implemented by the backend.
Backports commit cedbcb01529cb6cf9a2289cdbebbc63f6149fc18 from qemu
We need to coordinate with the TCG_OVERSIZED_GUEST test in cputlb.c,
and allow 64-bit atomics even though sizeof(void *) == 4.
Backports commit 374aae653499f4d405caf32b7fff0c8639113fe4 from qemu
The cp15, CRn=15, opc1=0, CRm=5, opc2=0 instruction invalidates all the
data cache on the cortex-r5. Implementing it as a NOP.
Backports commit 95e9a242e2a393c7d4e5cc04340e39c3a9420f03 from qemu
M profile doesn't implement ARM, and the architecturally required
behaviour for attempts to execute with the Thumb bit clear is to
generate a UsageFault with the CFSR INVSTATE bit set. We were
incorrectly implementing this as generating an UNDEFINSTR UsageFault;
fix this.
Backports commit e13886e3a790b52f0b2e93cb5e84fdc2ada5471a from qemu
Implement the exception return consistency checks
described in the v7M pseudocode ExceptionReturn().
Inspired by a patch from Michael Davidsaver's series, but
this is a reimplementation from scratch based on the
ARM ARM pseudocode.
Backports commit aa488fe3bb5460c6675800ccd80f6dccbbd70159 from qemu
Extract the code from the tail end of arm_v7m_do_interrupt() which
enters the exception handler into a pair of utility functions
v7m_exception_taken() and v7m_push_stack(), which correspond roughly
to the pseudocode PushStack() and ExceptionTaken().
This also requires us to move the arm_v7m_load_vector() utility
routine up so we can call it.
Handling illegal exception returns has some cases where we want to
take a UsageFault either on an existing stack frame or with a new
stack frame but with a specific LR value, so we want to be able to
call these without having to go via arm_v7m_cpu_do_interrupt().
Backports commit 39ae2474e337247e5930e8be783b689adc9f6215 from qemu
All the places in armv7m_cpu_do_interrupt() which pend an
exception in the NVIC are doing so for synchronous
exceptions. We know that we will always take some
exception in this case, so we can just acknowledge it
immediately, rather than returning and then immediately
being called again because the NVIC has raised its outbound
IRQ line.
Backports commit a25dc805e2e63a55029e787a52335e12dabf07dc from qemu
The M profile condition for when we can take a pending exception or
interrupt is not the same as that for A/R profile. The code
originally copied from the A/R profile version of the
cpu_exec_interrupt function only worked by chance for the
very simple case of exceptions being masked by PRIMASK.
Replace it with a call to a function in the NVIC code that
correctly compares the priority of the pending exception
against the current execution priority of the CPU.
Backports commit 7ecdaa4a9635f1ded0dfa9218c25273b6d4dcd44 from qemu
Having armv7m_nvic_acknowledge_irq() return the new value of
env->v7m.exception and its one caller assign the return value
back to env->v7m.exception is pointless. Just make the return
type void instead.
Backports commit a5d8235545e98c1ce02560d5f4f57552d937efe9 from qemu
Implement HFNMIENA support for the M profile MPU. This bit controls
whether the MPU is treated as enabled when executing at execution
priorities of less than zero (in NMI, HardFault or with the FAULTMASK
bit set).
Doing this requires us to use a different MMU index for "running
at execution priority < 0", because we will have different
access permissions for that case versus the normal case.
Backports commit 3bef7012560a7f0ea27b265105de5090ba117514 from qemu
The M series MPU is almost the same as the already implemented R
profile MPU (v7 PMSA). So all we need to implement here is the MPU
register interface in the system register space.
This implementation has the same restriction as the R profile MPU
that it doesn't permit regions to be sized down smaller than 1K.
We also do not yet implement support for MPU_CTRL.HFNMIENA; this
bit should if zero disable use of the MPU when running HardFault,
NMI or with FAULTMASK set to 1 (ie at an execution priority of
less than zero) -- if the MPU is enabled we don't treat these
cases any differently.
Backports commit 29c483a506070e8f554c77d22686f405e30b9114 from qemu
General logic is that operations stopped by the MPU are MemManage,
and those which go through the MPU and are caught by the unassigned
handle are BusFault. Distinguish these by looking at the
exception.fsr values, and set the CFSR bits and (if appropriate)
fill in the BFAR or MMFAR with the exception address.
Backports commit 5dd0641d234e355597be62e5279d8a519c831625 from qemu
