Fix Support for ARM CBAR and reset-hivecs
v5 cover Fix Support for ARM CBAR and reset-hivecs
Hi All,
This patch series adds support for two board configurable ARM CPU properties - Configuration Base Address Register and the hivecs-on-reset.
The CBAR is needed to fix Zynq and Highbank which both were broken for linux boot. This series provides the fixes.
I have added these properties as qdev properties rather than object properties to pick up the desired writable-until-realize semantic.
To test the semantics of these new qdev properties I hacked Highbank to illegally set the property after realize:
hw/arm/highbank.c +0 -0
@@ -236,13 +236,13 @@ static void calxeda_init(QEMUMachineInitArgs
*args, enum cxmachines machine)
cpu = ARM_CPU(object_new(object_class_get_name(oc)));
- object_property_set_int(OBJECT(cpu), MPCORE_PERIPHBASE, "reset-cbar",
- &err);
+ object_property_set_bool(OBJECT(cpu), true, "realized", &err);
if (err) {
error_report("%s", error_get_pretty(err));
exit(1);
}
- object_property_set_bool(OBJECT(cpu), true, "realized", &err);
+ object_property_set_int(OBJECT(cpu), MPCORE_PERIPHBASE, "reset-cbar",
+ &err);
if (err) {
error_report("%s", error_get_pretty(err));
exit(1);Booting with this bug I now get:
$ qemu-system-arm -M highbank ... qemu-system-arm: Attempt to set property 'reset-cbar' on anonymous device (type 'cortex-a9-arm-cpu') after it was realized
whereas the previous versions would have silent failed.
Regards,
Peter
changed since v4:
Reimplemented using qdev properties
dropped QOM patches
rebased against mainline
changed since v3:
Rebased against target-arm patch queue (2013/12/11) Added Antonys reset hivecs patches
changed since v2:
Fixed comment in p8 (PMM review)
Enabled CBAR for a15 (PMM review)
Typo sweep
Changed since v1:
Fix QOM to support writeable dynamic properties Use dynamic props instead (PMM/AF discussion) Use error_report (AF reivew)
Use reset- prefix on propname (AF review) Fix machine model namings or the MPCore PERIPHBASE
Antony Pavlov (2):
ARM: cpu: add "reset_hivecs" property
ARM: arm_cpu_reset: make it possible to use high vectors for reset_exc
Peter Crosthwaite (8):
target-arm/helper.c: Allow cp15.c15 dummy override
target-arm: Define and use ARM_FEATURE_CBAR target-arm/cpu: Convert reset CBAR to a property arm/highbank: Use object_new() rather than cpu_arm_init() arm/highbank: Fix CBAR initialisation arm/xilinx_zynq: Use object_new() rather than cpu_arm_init() arm/xilinx_zynq: Implement CBAR initialisation arm/highbank.c: Fix MPCore periphbase name
hw/arm/highbank.c | 33 +++++++++++++++++++++------------ hw/arm/xilinx_zynq.c | 21 +++++++++++++++++---- target-arm/cpu-qom.h | 1 + target-arm/cpu.c | 47 ++++++++++++++++++++++++++++++++++++++--------- target-arm/cpu.h | 1 + target-arm/helper.c | 12 +++++++++++- 6 files changed, 89 insertions(+), 26 deletions(-)
Signature · 1 line
1.8.5.1
PATCH v5 1/10 target-arm/helper.c: Allow cp15.c15 dummy override
The cp15.c15 space is implementation defined. Currently there is a dummy placeholder register RAZing it. Allow overriding of this RAZ so implementations of specific registers can take precedence.
---
target-arm/helper.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-)
target-arm/helper.c +2 -1
@@ -1338,7 +1338,8 @@ static const ARMCPRegInfo dummy_c15_cp_reginfo[] = {*/ { .name = "C15_IMPDEF", .cp = 15, .crn = 15, .crm = CP_ANY, .opc1 = CP_ANY, .opc2 = CP_ANY, - .access = PL1_RW, .type = ARM_CP_CONST | ARM_CP_NO_MIGRATE, + .access = PL1_RW, + .type = ARM_CP_CONST | ARM_CP_NO_MIGRATE | ARM_CP_OVERRIDE, .resetvalue = 0 }, REGINFO_SENTINEL };
Signature · 1 line
1.8.5.1
PATCH v5 2/10 target-arm: Define and use ARM_FEATURE_CBAR
In reply to Peter CrosthwaiteSome processors (notably A9 within Highbank) define and use the CP15 configuration base address (CBAR). This is vendor specific so its best implemented as a CPU property (otherwise we would need vendor specific child classes for every ARM implementation).
This patch prepares support for converting CBAR reset value to a CPU property by moving the CP registration out of the CPU init fn, as registration will need to happen at realize time to pick up any property updates. The easiest way to do this is via definition of a new ARM_FEATURE to flag the existence of the register.
---
changed since v3:
Move new feature to end of list
changed since v2:
msg typo: existence
Enable CBAR for a15 as well
target-arm/cpu.c | 12 +++--------- target-arm/cpu.h | 1 + target-arm/helper.c | 9 +++++++++ 3 files changed, 13 insertions(+), 9 deletions(-)
target-arm/cpu.c +3 -9
@@ -616,6 +616,7 @@ static void cortex_a9_initfn(Object *obj)* and valid configurations; we don't model A9UP). */ set_feature (&cpu->env,ARM_FEATURE_V7MP ); +set_feature (&cpu->env,ARM_FEATURE_CBAR ); cpu->midr =0x410fc090 ; cpu->reset_fpsid =0x41033090 ; cpu->mvfr0 =0x11110222 ;
@@ -638,15 +639,7 @@ static void cortex_a9_initfn(Object *obj) cpu->clidr = (1 <<27 ) | (1 <<24 ) |3 ; cpu->ccsidr[0 ] =0xe00fe015 ;/* 16k L1 dcache. */ cpu->ccsidr[1 ] =0x200fe015 ;/* 16k L1 icache. */ - { -ARMCPRegInfo cbar = { - .name ="CBAR" , .cp =15 , .crn =15 , .crm =0 , .opc1 =4 , - .opc2 =0 , .access =PL1_R |PL3_W , .resetvalue = cpu->reset_cbar, - .fieldoffset = offsetof(CPUARMState, cp15.c15_config_base_address) - }; -define_one_arm_cp_reg (cpu, &cbar); -define_arm_cp_regs (cpu, cortexa9_cp_reginfo); - } +define_arm_cp_regs (cpu, cortexa9_cp_reginfo); }#ifndef CONFIG_USER_ONLY
@@ -685,6 +678,7 @@ static void cortex_a15_initfn(Object *obj)set_feature (&cpu->env,ARM_FEATURE_ARM_DIV );set_feature (&cpu->env,ARM_FEATURE_GENERIC_TIMER );set_feature (&cpu->env,ARM_FEATURE_DUMMY_C15_REGS ); +set_feature (&cpu->env,ARM_FEATURE_CBAR );set_feature (&cpu->env,ARM_FEATURE_LPAE ); cpu->kvm_target =QEMU_KVM_ARM_TARGET_CORTEX_A15 ; cpu->midr =0x412fc0f1 ;
target-arm/cpu.h +1 -0
@@ -467,6 +467,7 @@ enum arm_features {ARM_FEATURE_LPAE ,/* has Large Physical Address Extension */ ARM_FEATURE_V8 ,ARM_FEATURE_AARCH64 ,/* supports 64 bit mode */ + ARM_FEATURE_CBAR, /* has cp15 CBAR */ };static inline int arm_feature (CPUARMState *env,int feature)
target-arm/helper.c +9 -0
@@ -1745,6 +1745,15 @@ void register_cp_regs_for_features(ARMCPU *cpu)define_one_arm_cp_reg (cpu, &auxcr); } +if (arm_feature (env,ARM_FEATURE_CBAR )) { +ARMCPRegInfo cbar = { + .name ="CBAR" , .cp =15 , .crn =15 , .crm =0 , .opc1 =4 , .opc2 =0 , + .access =PL1_R |PL3_W , .resetvalue = cpu->reset_cbar, + .fieldoffset = offsetof(CPUARMState, cp15.c15_config_base_address) + }; +define_one_arm_cp_reg (cpu, &cbar); + } +/* Generic registers whose values depend on the implementation */ {ARMCPRegInfo sctlr = {
Signature · 1 line
1.8.5.1
PATCH v5 3/10 target-arm/cpu: Convert reset CBAR to a property
In reply to Peter CrosthwaiteThe reset value of the CP15 CBAR is a vendor (machine) configurable property. If ARM_FEATURE_CBAR is set, add it as a property at post_init time.
---
Changed since v4:
Re-implemented as qdev property.
Removed PMM reviewed-by (major implementation changes) Changed since v3:
typo s/Value/value (PMM review)
Change since v1:
Re-implement as dynamic property
target-arm/cpu.c | 17 +++++++++++++++++ 1 file changed, 17 insertions(+)
target-arm/cpu.c +17 -0
@@ -21,6 +21,7 @@#include "cpu.h" #include "qemu-common.h" #include "hw/qdev-properties.h" +#include "qapi/qmp/qerror.h" #if !defined(CONFIG_USER_ONLY )#include "hw/loader.h" #endif
@@ -231,6 +232,21 @@ static void arm_cpu_initfn(Object *obj) } } +static Property arm_cpu_reset_cbar_property = +DEFINE_PROP_UINT32 ("reset-cbar" ,ARMCPU , reset_cbar,0 ); + +static void arm_cpu_post_init (Object *obj) +{ +ARMCPU *cpu =ARM_CPU (obj); +Error *err =NULL ; + +if (arm_feature (&cpu->env,ARM_FEATURE_CBAR )) { +qdev_property_add_static (DEVICE (obj), &arm_cpu_reset_cbar_property, + &err); +assert_no_error (err); + } +} +static void arm_cpu_finalizefn (Object *obj) {ARMCPU *cpu =ARM_CPU (obj);
@@ -993,6 +1009,7 @@ static const TypeInfo arm_cpu_type_info = { .parent = TYPE_CPU, .instance_size = sizeof(ARMCPU), .instance_init = arm_cpu_initfn, + .instance_post_init = arm_cpu_post_init, .instance_finalize = arm_cpu_finalizefn, .abstract = true, .class_size = sizeof(ARMCPUClass),
Signature · 1 line
1.8.5.1
PATCH v5 4/10 arm/highbank: Use object_new() rather than cpu_arm_init()
In reply to Peter CrosthwaiteTo allow the machine model to set device properties before CPU realization.
---
changed since v1:
use error_report rather than fprintf(stderr
hw/arm/highbank.c | 12 +++++++++--- 1 file changed, 9 insertions(+), 3 deletions(-)
hw/arm/highbank.c +9 -3
@@ -26,6 +26,7 @@#include "hw/boards.h" #include "sysemu/blockdev.h" #include "exec/address-spaces.h" +#include "qemu/error-report.h" #define SMP_BOOT_ADDR 0x100#define SMP_BOOT_REG 0x40
@@ -229,10 +230,15 @@ static void calxeda_init(QEMUMachineInitArgs *args, enum cxmachines machine) }for (n =0 ; n < smp_cpus; n++) { +ObjectClass *oc =cpu_class_by_name (TYPE_ARM_CPU , cpu_model);ARMCPU *cpu; - cpu =cpu_arm_init (cpu_model); -if (cpu ==NULL ) { -fprintf (stderr,"Unable to find CPU definition\n" ); +Error *err =NULL ; + + cpu =ARM_CPU (object_new (object_class_get_name (oc))); + +object_property_set_bool (OBJECT (cpu), true,"realized" , &err); +if (err) { +error_report ("%s" ,error_get_pretty (err));exit (1 ); }
Signature · 1 line
1.8.5.1
PATCH v5 5/10 arm/highbank: Fix CBAR initialisation
In reply to Peter CrosthwaiteFix the CBAR initialisation by using the newly defined static property. CBAR is now set before realization, so the intended value is now actually used.
So I have kind of tested this. I booted an ARM kernel on Highbank with the stock Highbank DTB. It doesn't boot (and I will be doing something wrong), but before this patch I got this:
------------[ cut here ]------------ WARNING: CPU: 0 PID: 0 at /workspaces/pcrost/public/linux2.git/arch/arm/mm/ioremap.c:301 __arm_ioremap_pfn_caller+0x180/0x198() CPU: 0 PID: 0 Comm: swapper/0 Tainted: G W 3.13.0-rc1-next-20131126-dirty #2 [<c0015164>] (unwind_backtrace) from [<c00118c0>] (show_stack+0x10/0x14) [<c00118c0>] (show_stack) from [<c02bd5fc>] (dump_stack+0x78/0x90) [<c02bd5fc>] (dump_stack) from [<c001f110>] (warn_slowpath_common+0x68/0x84) [<c001f110>] (warn_slowpath_common) from [<c001f1f4>] (warn_slowpath_null+0x1c/0x24) [<c001f1f4>] (warn_slowpath_null) from [<c0017c6c>] (__arm_ioremap_pfn_caller+0x180/0x198) [<c0017c6c>] (__arm_ioremap_pfn_caller) from [<c0017cd8>] (__arm_ioremap_caller+0x54/0x5c) [<c0017cd8>] (__arm_ioremap_caller) from [<c0017d10>] (__arm_ioremap+0x18/0x1c) [<c0017d10>] (__arm_ioremap) from [<c03913c0>] (highbank_init_irq+0x34/0x8c) [<c03913c0>] (highbank_init_irq) from [<c038c228>] (init_IRQ+0x28/0x2c) [<c038c228>] (init_IRQ) from [<c03899ec>] (start_kernel+0x234/0x398) [<c03899ec>] (start_kernel) from [<00008074>] (0x8074) ---[ end trace 3406ff24bd97382f ]---
Which disappears with this patch.
---
changed since v2:
Fix msg typos
changed since v1:
use error report rather than fprintf(stderr
hw/arm/highbank.c | 8 +++++--- 1 file changed, 5 insertions(+), 3 deletions(-)
hw/arm/highbank.c +5 -3
@@ -236,14 +236,16 @@ static void calxeda_init(QEMUMachineInitArgs *args, enum cxmachines machine) cpu =ARM_CPU (object_new (object_class_get_name (oc))); +object_property_set_int (OBJECT (cpu),GIC_BASE_ADDR ,"reset-cbar" , &err); +if (err) { +error_report ("%s" ,error_get_pretty (err)); +exit (1 ); + }object_property_set_bool (OBJECT (cpu), true,"realized" , &err);if (err) {error_report ("%s" ,error_get_pretty (err));exit (1 ); } - -/* This will become a QOM property eventually */ - cpu->reset_cbar =GIC_BASE_ADDR ; cpu_irq[n] =qdev_get_gpio_in (DEVICE (cpu),ARM_CPU_IRQ ); }
Signature · 1 line
1.8.5.1
PATCH v5 6/10 arm/xilinx_zynq: Use object_new() rather than cpu_arm_init()
In reply to Peter CrosthwaiteTo allow the machine model to set device properties before CPU realization.
---
changed since v1:
use error report rather than fprintf(stderr
hw/arm/xilinx_zynq.c | 12 +++++++++--- 1 file changed, 9 insertions(+), 3 deletions(-)
hw/arm/xilinx_zynq.c +9 -3
@@ -25,6 +25,7 @@#include "sysemu/blockdev.h" #include "hw/loader.h" #include "hw/ssi.h" +#include "qemu/error-report.h" #define NUM_SPI_FLASHES 4#define NUM_QSPI_FLASHES 2
@@ -102,6 +103,7 @@ static void zynq_init(QEMUMachineInitArgs *args)const char *kernel_filename = args->kernel_filename;const char *kernel_cmdline = args->kernel_cmdline;const char *initrd_filename = args->initrd_filename; +ObjectClass *cpu_oc;ARMCPU *cpu;MemoryRegion *address_space_mem =get_system_memory ();MemoryRegion *ext_ram =g_new (MemoryRegion,1 );
@@ -110,15 +112,19 @@ static void zynq_init(QEMUMachineInitArgs *args)SysBusDevice *busdev;qemu_irq pic[64 ];NICInfo *nd; +Error *err =NULL ;int n;if (!cpu_model) { cpu_model ="cortex-a9" ; } + cpu_oc =cpu_class_by_name (TYPE_ARM_CPU , cpu_model); - cpu =cpu_arm_init (cpu_model); -if (!cpu) { -fprintf (stderr,"Unable to find CPU definition\n" ); + cpu =ARM_CPU (object_new (object_class_get_name (cpu_oc))); + +object_property_set_bool (OBJECT (cpu), true,"realized" , &err); +if (err) { +error_report ("%s" ,error_get_pretty (err));exit (1 ); }
Signature · 1 line
1.8.5.1
PATCH v5 7/10 arm/xilinx_zynq: Implement CBAR initialisation
In reply to Peter CrosthwaiteFix the CBAR initialisation by using the newly defined static property. Zynq will now correctly init the CBAR to the SCU base address.
Needed to boot Linux on the xilinx_zynq machine model.
---
changed since v1:
use error report rather than fprintf(stderr
rename SCU_BASE_ADDR to MPCORE_PERIPHBASE
hw/arm/xilinx_zynq.c | 9 ++++++++- 1 file changed, 8 insertions(+), 1 deletion(-)
hw/arm/xilinx_zynq.c +8 -1
@@ -36,6 +36,8 @@#define IRQ_OFFSET 32/* pic interrupts start from index 32 */ +#define MPCORE_PERIPHBASE 0xF8F00000 +static const int dma_irqs[8 ] = {46 ,47 ,48 ,49 ,72 ,73 ,74 ,75 };
@@ -122,6 +124,11 @@ static void zynq_init(QEMUMachineInitArgs *args) cpu =ARM_CPU (object_new (object_class_get_name (cpu_oc))); +object_property_set_int (OBJECT (cpu),MPCORE_PERIPHBASE ,"reset-cbar" , &err); +if (err) { +error_report ("%s" ,error_get_pretty (err)); +exit (1 ); + }object_property_set_bool (OBJECT (cpu), true,"realized" , &err);if (err) {error_report ("%s" ,error_get_pretty (err));
@@ -160,7 +167,7 @@ static void zynq_init(QEMUMachineInitArgs *args)qdev_prop_set_uint32 (dev,"num-cpu" ,1 );qdev_init_nofail (dev); busdev =SYS_BUS_DEVICE (dev); -sysbus_mmio_map (busdev,0 ,0xF8F00000 ); +sysbus_mmio_map (busdev,0 ,MPCORE_PERIPHBASE );sysbus_connect_irq (busdev,0 ,qdev_get_gpio_in (DEVICE (cpu),ARM_CPU_IRQ ));
Signature · 1 line
1.8.5.1
PATCH v5 8/10 arm/highbank.c: Fix MPCore periphbase name
In reply to Peter CrosthwaiteGIC_BASE_ADDR is not the base address of the GIC. Its clear from the code that this is the base address of the MPCore. Rename to MPCORE_PERIPHBASE accordingly.
---
changed since v2: Fixed broken comment (PMM review)
hw/arm/highbank.c | 15 ++++++++------- 1 file changed, 8 insertions(+), 7 deletions(-)
hw/arm/highbank.c +8 -7
@@ -28,11 +28,11 @@#include "exec/address-spaces.h" #include "qemu/error-report.h" -#define SMP_BOOT_ADDR 0x100 -#define SMP_BOOT_REG 0x40 -#define GIC_BASE_ADDR 0xfff10000 +#define SMP_BOOT_ADDR 0x100 +#define SMP_BOOT_REG 0x40 +#define MPCORE_PERIPHBASE 0xfff10000 -#define NIRQ_GIC 160 +#define NIRQ_GIC 160/* Board init. */
@@ -55,7 +55,7 @@ static void hb_write_secondary(ARMCPU *cpu, const struct arm_boot_info *info)0xe1110001 ,/* tst r1, r1 */ 0x0afffffb ,/* beq <wfi> */ 0xe12fff11 ,/* bx r1 */ - GIC_BASE_ADDR /* privbase: gic address. */ + MPCORE_PERIPHBASE /* privbase: MPCore peripheral base address. */ };for (n =0 ; n <ARRAY_SIZE (smpboot); n++) { smpboot[n] =tswap32 (smpboot[n]);
@@ -236,7 +236,8 @@ static void calxeda_init(QEMUMachineInitArgs *args, enum cxmachines machine) cpu =ARM_CPU (object_new (object_class_get_name (oc))); -object_property_set_int (OBJECT (cpu),GIC_BASE_ADDR ,"reset-cbar" , &err); +object_property_set_int (OBJECT (cpu),MPCORE_PERIPHBASE ,"reset-cbar" , + &err);if (err) {error_report ("%s" ,error_get_pretty (err));exit (1 );
@@ -287,7 +288,7 @@ static void calxeda_init(QEMUMachineInitArgs *args, enum cxmachines machine)qdev_prop_set_uint32 (dev,"num-irq" ,NIRQ_GIC );qdev_init_nofail (dev); busdev =SYS_BUS_DEVICE (dev); -sysbus_mmio_map (busdev,0 ,GIC_BASE_ADDR ); +sysbus_mmio_map (busdev,0 ,MPCORE_PERIPHBASE );for (n =0 ; n < smp_cpus; n++) {sysbus_connect_irq (busdev, n, cpu_irq[n]); }
Signature · 1 line
1.8.5.1
PATCH v5 9/10 ARM: cpu: add "reset_hivecs" property
In reply to Peter CrosthwaiteFrom: Antony Pavlov <[email protected]>
Add an ARM CPU property for the reset value of hivecs as it is a board/SoC configurable setting.
The existence of the property is conditional on the ARM CPU not being M class.
[ PC Changes:
* Elaborated commit message
* refactored to use qdev_property_add_static
]
---
changed since v4:
Re-implemented using qdev_property_add_static
target-arm/cpu-qom.h | 1 + target-arm/cpu.c | 13 +++++++++++++ 2 files changed, 14 insertions(+)
target-arm/cpu-qom.h +1 -0
@@ -139,6 +139,7 @@ typedef struct ARMCPU {uint32_t ccsidr[16 ];uint32_t reset_cbar;uint32_t reset_auxcr; +bool reset_hivecs; }ARMCPU ;#define TYPE_AARCH64_CPU "aarch64-cpu"
target-arm/cpu.c +13 -0
@@ -235,6 +235,9 @@ static void arm_cpu_initfn(Object *obj)static Property arm_cpu_reset_cbar_property =DEFINE_PROP_UINT32 ("reset-cbar" ,ARMCPU , reset_cbar,0 ); +static Property arm_cpu_reset_hivecs_property = +DEFINE_PROP_BOOL ("reset-hivecs" ,ARMCPU , reset_hivecs, false); +static void arm_cpu_post_init (Object *obj) {ARMCPU *cpu =ARM_CPU (obj);
@@ -245,6 +248,12 @@ static void arm_cpu_post_init(Object *obj) &err);assert_no_error (err); } + +if (!arm_feature (&cpu->env,ARM_FEATURE_M )) { +qdev_property_add_static (DEVICE (obj), &arm_cpu_reset_hivecs_property, + &err); +assert_no_error (err); + } }static void arm_cpu_finalizefn (Object *obj)
@@ -306,6 +315,10 @@ static void arm_cpu_realizefn(DeviceState *dev, Error **errp)set_feature (env,ARM_FEATURE_PXN ); } +if (cpu->reset_hivecs) { + cpu->reset_sctlr |= (1 <<13 ); + } +register_cp_regs_for_features (cpu);arm_cpu_register_gdb_regs_for_features (cpu);
Signature · 1 line
1.8.5.1
PATCH v5 10/10 ARM: arm_cpu_reset: make it possible to use high vectors for reset_exc
In reply to Peter CrosthwaiteFrom: Antony Pavlov <[email protected]>
If hivecs are being used on reset, the CPU should come out of reset at the hivecs reset vector (0xFFFF0000)
[ PC Changes:
* Fixed Grammar error in commit message
* Elaborated commit message.
]
---
target-arm/cpu.c | 5 +++++ 1 file changed, 5 insertions(+)
target-arm/cpu.c +5 -0
@@ -121,6 +121,11 @@ static void arm_cpu_reset(CPUState *s) env->regs[15] = pc & ~1; } } + +if (env->cp15.c1_sys & (1 <<13 )) { + env->regs[15 ] =0xFFFF0000 ; + } + env->vfp.xregs[ARM_VFP_FPEXC ] =0 ;#endif set_flush_to_zero (1 , &env->vfp.standard_fp_status);
Signature · 1 line
1.8.5.1
Quoted 11 lines · Peter Crosthwaite
Hi All,
This patch series adds support for two board configurable ARM CPU properties - Configuration Base Address Register and the hivecs-on-reset.
The CBAR is needed to fix Zynq and Highbank which both were broken for linux boot. This series provides the fixes.
I have added these properties as qdev properties rather than object properties to pick up the desired writable-until-realize semantic.
Looks good to me, thanks. Applied all to target-arm.next.
thanks
-- PMM