mirrors/qemu
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity
stable-7.2
qemu/target/arm
History
Peter Maydell 318a0b0556 target/arm: Handle denormals correctly for FMOPA (widening) 
The FMOPA (widening) SME instruction takes pairs of half-precision
floating point values, widens them to single-precision, does a
two-way dot product and accumulates the results into a
single-precision destination.  We don't quite correctly handle the
FPCR bits FZ and FZ16 which control flushing of denormal inputs and
outputs.  This is because at the moment we pass a single float_status
value to the helper function, which then uses that configuration for
all the fp operations it does.  However, because the inputs to this
operation are float16 and the outputs are float32 we need to use the
fp_status_f16 for the float16 input widening but the normal fp_status
for everything else.  Otherwise we will apply the flushing control
FPCR.FZ16 to the 32-bit output rather than the FPCR.FZ control, and
incorrectly flush a denormal output to zero when we should not (or
vice-versa).

(In commit 207d30b5fd we tried to fix the FZ handling but
didn't get it right, switching from "use FPCR.FZ for everything" to
"use FPCR.FZ16 for everything".)
(Mjt: it is commit d5373d7bdb in stable-7.2)

Pass the CPU env to the sme_fmopa_h helper instead of an fp_status
pointer, and have the helper pass an extra fp_status into the
f16_dotadd() function so that we can use the right status for the
right parts of this operation.

Cc: qemu-stable@nongnu.org
Fixes: 207d30b5fd ("target/arm: Use FPST_F16 for SME FMOPA (widening)")
Resolves: https://gitlab.com/qemu-project/qemu/-/issues/2373
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
(cherry picked from commit 55f9f4ee01)
Signed-off-by: Michael Tokarev <mjt@tls.msk.ru>
(Mjt: s/tcg_env/cpu_env/ due to missingv
 8.1.0-1189-gad75a51e84af "tcg: Rename cpu_env to tcg_env")
2024-08-02 14:01:04 +03:00
..
hvf hvf: arm: Fix encodings for ID_AA64PFR1_EL1 and debug System registers 2024-05-30 17:11:59 +03:00
a32-uncond.decode arm tcg cpus: Fix Lesser GPL version number 2020-11-15 16:42:14 +01:00
a32.decode target/arm: Implement ESB instruction 2022-05-09 11:47:54 +01:00
arch_dump.c dump: Replace opaque DumpState pointer with a typed one 2022-10-06 19:30:43 +04:00
arm_ldst.h accel/tcg: Add DisasContextBase argument to translator_ld* 2021-09-14 12:00:20 -07:00
arm-powerctl.c arm/arm-powerctl: rebuild hflags after setting CP15 bits in arm_set_cpu_on() 2019-12-20 14:03:00 +00:00
arm-powerctl.h target/arm/arm-powerctl: Add new arm_set_cpu_on_and_reset() 2019-02-28 11:03:04 +00:00
common-semi-target.h semihosting: Split out common-semi-target.h 2022-06-28 04:35:07 +05:30
cpregs.h target/arm: Move define_debug_regs() to debug_helper.c 2022-07-07 11:37:33 +01:00
cpu64.c target/arm: Disable SVE extensions when SVE is disabled 2024-06-01 07:20:31 +03:00
cpu_tcg.c target/arm: Set TCGCPUOps.restore_state_to_opc for v7m 2022-11-29 18:15:26 -05:00
cpu-param.h target/arm: Enable TARGET_TB_PCREL 2022-10-20 11:28:29 +01:00
cpu-qom.h target: Introduce and use OBJECT_DECLARE_CPU_TYPE() macro 2022-03-06 22:23:09 +01:00
cpu.c target/arm: Disable SME if SVE is disabled 2023-12-20 19:11:10 +03:00
cpu.h target/arm: Handle m-profile in arm_is_secure 2023-04-12 16:57:32 +03:00
crypto_helper.c crypto: move sm4_sbox from target/arm 2022-04-29 10:47:45 +10:00
debug_helper.c target/arm: Store TCR_EL* registers as uint64_t 2022-07-18 13:20:13 +01:00
gdbstub64.c target/arm: Rename sve_zcr_len_for_el to sve_vqm1_for_el 2022-06-08 19:38:57 +01:00
gdbstub.c Fix 'writeable' typos 2022-06-08 19:38:47 +01:00
helper-a64.c target/arm: Change CPUArchState.aarch64 to bool 2022-04-22 14:44:54 +01:00
helper-a64.h target/arm: Merge mte_check1, mte_checkN 2021-04-30 11:16:49 +01:00
helper-mve.h target/arm: Implement MVE VRINT insns 2021-09-01 11:08:17 +01:00
helper-sme.h target/arm: Handle denormals correctly for FMOPA (widening) 2024-08-02 14:01:04 +03:00
helper-sve.h target/arm: Implement REVD 2022-07-11 13:43:51 +01:00
helper.c target/arm: Ignore SMCR_EL2.LEN and SVCR_EL2.LEN if EL2 is not enabled 2024-08-02 10:24:26 +03:00
helper.h target/arm: Implement SCLAMP, UCLAMP 2022-07-11 13:43:51 +01:00
hvf_arm.h target: Use forward declared type instead of structure type 2022-03-06 22:22:40 +01:00
idau.h Use DECLARE_*CHECKER* macros 2020-09-09 09:27:09 -04:00
internals.h target/arm: Correctly propagate stage 1 BTI guarded bit in a two-stage walk 2023-11-03 19:35:34 +03:00
iwmmxt_helper.c arm tcg cpus: Fix Lesser GPL version number 2020-11-15 16:42:14 +01:00
Kconfig meson: Introduce target-specific Kconfig 2021-07-09 18:21:34 +02:00
kvm64.c arm64: Restore trapless ptimer access 2023-09-13 12:21:22 +03:00
kvm_arm.h target/arm: Initialize debug capabilities only once 2023-05-18 21:09:59 +03:00
kvm-consts.h target/arm: Report KVM's actual PSCI version to guest in dtb 2022-03-02 19:27:37 +00:00
kvm-stub.c target/arm: Avoid bare abort() or assert(0) 2022-05-05 09:35:51 +01:00
kvm.c accel/kvm: Specify default IPA size for arm64 2023-09-11 10:53:50 +03:00
m_helper.c target/arm: Use tlb_set_page_full 2022-10-10 14:52:25 +01:00
m-nocp.decode target/arm: Don't NOCP fault for FPCXT_NS accesses 2021-06-21 16:49:37 +01:00
machine.c target/arm: Add the SME ZA storage to CPUARMState 2022-06-27 11:18:17 +01:00
meson.build target/arm: Trap non-streaming usage when Streaming SVE is active 2022-07-11 13:19:35 +01:00
monitor.c target/arm: Add cpu properties to control pauth 2021-01-19 14:38:51 +00:00
mte_helper.c accel/tcg: Simplify page_get/alloc_target_data 2022-10-26 11:11:28 +10:00
mve_helper.c target/arm: Use expand_pred_b in mve_helper.c 2022-06-08 19:38:58 +01:00
mve.decode target/arm: Implement MVE VRINT insns 2021-09-01 11:08:17 +01:00
neon_helper.c Replace config-time define HOST_WORDS_BIGENDIAN 2022-04-06 10:50:37 +02:00
neon-dp.decode target/arm: Implement vector float32 to bfloat16 conversion 2021-06-03 16:43:26 +01:00
neon-ls.decode target/arm: Remove duplicate 'plus1' function from Neon and SVE decode 2021-07-18 10:59:47 +01:00
neon-shared.decode target/arm: Remove duplicate 'plus1' function from Neon and SVE decode 2021-07-18 10:59:47 +01:00
op_addsub.h
op_helper.c accel/tcg: Remove will_exit argument from cpu_restore_state 2022-11-01 08:31:41 +11:00
pauth_helper.c compiler.h: replace QEMU_NORETURN with G_NORETURN 2022-04-21 17:03:51 +04:00
psci.c target/arm: Support PSCI 1.1 and SMCCC 1.0 2022-03-02 19:27:36 +00:00
ptw.c target/arm: Correctly propagate stage 1 BTI guarded bit in a two-stage walk 2023-11-03 19:35:34 +03:00
sme_helper.c target/arm: Handle denormals correctly for FMOPA (widening) 2024-08-02 14:01:04 +03:00
sme-fa64.decode target/arm: Mark LD1RO as non-streaming 2022-07-11 13:19:35 +01:00
sme.decode target/arm: Implement SME integer outer product 2022-07-11 13:43:51 +01:00
sve_helper.c target/arm: Fix SVE/SME gross MTE suppression checks 2024-02-16 14:19:18 +03:00
sve_ldst_internal.h target/arm: Use probe_access_full for MTE 2022-10-20 11:27:49 +01:00
sve.decode target/arm: Implement SCLAMP, UCLAMP 2022-07-11 13:43:51 +01:00
syndrome.h target/arm: fix exception syndrome for AArch32 bkpt insn 2024-02-03 16:42:56 +03:00
t16.decode arm tcg cpus: Fix Lesser GPL version number 2020-11-15 16:42:14 +01:00
t32.decode target/arm: Implement ESB instruction 2022-05-09 11:47:54 +01:00
tlb_helper.c target/arm: Explicitly select short-format FSR for M-profile 2023-05-31 09:43:56 +03:00
trace-events docs: fix references to docs/devel/tracing.rst 2021-06-02 06:51:09 +02:00
trace.h trace: switch position of headers to what Meson requires 2020-08-21 06:18:24 -04:00
translate-a32.h target/arm: Define and use new load_cpu_field_low32() 2023-05-18 21:09:59 +03:00
translate-a64.c target/arm: Return correct result for LDG when ATA=0 2023-06-22 10:35:22 +03:00
translate-a64.h target/arm: Export unpredicated ld/st from translate-sve.c 2022-07-11 13:19:35 +01:00
translate-m-nocp.c target/arm: Enable TARGET_TB_PCREL 2022-10-20 11:28:29 +01:00
translate-mve.c target/arm: Change gen_exception_insn* to work on displacements 2022-10-20 11:27:52 +01:00
translate-neon.c target/arm: Fix alignment for VLD4.32 2022-09-22 16:38:27 +01:00
translate-sme.c target/arm: Handle denormals correctly for FMOPA (widening) 2024-08-02 14:01:04 +03:00
translate-sve.c target/arm: Avoid shifts by -1 in tszimm_shr() and tszimm_shl() 2024-08-02 10:24:18 +03:00
translate-vfp.c target/arm: Change gen_exception_insn* to work on displacements 2022-10-20 11:27:52 +01:00
translate.c target/arm: Fix 64-bit SSRA 2023-09-11 10:53:50 +03:00
translate.h target/arm: Enable TARGET_TB_PCREL 2022-10-20 11:28:29 +01:00
vec_helper.c target/arm: Fix VCMLA Dd, Dn, Dm[idx] 2024-07-02 09:03:21 +03:00
vec_internal.h target/arm: Export bfdotadd from vec_helper.c 2022-06-08 19:38:58 +01:00
vfp_helper.c target/arm: Check NaN mode before silencing NaN 2021-07-02 11:48:36 +01:00
vfp-uncond.decode arm tcg cpus: Fix Lesser GPL version number 2020-11-15 16:42:14 +01:00
vfp.decode target/arm: Don't NOCP fault for FPCXT_NS accesses 2021-06-21 16:49:37 +01:00