Re: [PATCH 2/4] target/arm: Move cortex-m related functions to new file v7m.c

[Auger Eric]

Hi Thomas,

On 9/24/19 1:06 PM, Thomas Huth wrote:
> On 24/09/2019 13.02, Auger Eric wrote:
>> Hi Thomas,
>>
>> On 9/23/19 8:09 PM, Thomas Huth wrote:
>>> On 23/09/2019 16.31, Auger Eric wrote:
>>>> Hi Thomas,
>>>>
>>>> On 9/21/19 5:04 PM, Thomas Huth wrote:
>>>>> We are going to make CONFIG_ARM_V7M optional, so the related cortex-m
>>>>> CPUs should only be created if the switch is enabled. This can best
>>>>> be done if the code resides in a separate file, thus move the related
>>>>> functions to a new file v7m.c which only gets compiled if CONFIG_ARM_V7M
>>>>> is enabled.
>>>>>
>>>>> Signed-off-by: Thomas Huth <address@hidden>
>>>>> ---
>>>>>  target/arm/Makefile.objs |   1 +
>>>>>  target/arm/cpu.c         | 146 -----------------------------
>>>>>  target/arm/v7m.c         | 193 +++++++++++++++++++++++++++++++++++++++
>>>>>  3 files changed, 194 insertions(+), 146 deletions(-)
>>>>>  create mode 100644 target/arm/v7m.c
>>> [...]
>>>>> diff --git a/target/arm/v7m.c b/target/arm/v7m.c
>>>>> new file mode 100644
>>>>> index 0000000000..505043febe
>>>>> --- /dev/null
>>>>> +++ b/target/arm/v7m.c
>>>>> @@ -0,0 +1,193 @@
>>>>> +/*
>>>>> + * ARM v7m helpers.
>>>>> + *
>>>>> + * This code is licensed under the GNU GPL v2 or later.
>>>>> + *
>>>>> + * SPDX-License-Identifier: GPL-2.0-or-later
>>>>> + */
>>>>> +
>>>>> +#include "qemu/osdep.h"
>>>>> +#include "qemu/qemu-print.h"
>>>>> +#include "qemu-common.h"
>>>>> +#include "target/arm/idau.h"
>>>>> +#include "qemu/module.h"
>>>>> +#include "qapi/error.h"
>>>>> +#include "qapi/visitor.h"
>>>>> +#include "cpu.h"
>>>>> +#include "internals.h"
>>>>> +#include "exec/exec-all.h"
>>>>> +#include "hw/qdev-properties.h"
>>>>> +#if !defined(CONFIG_USER_ONLY)
>>>>> +#include "hw/loader.h"
>>>>> +#include "hw/boards.h"
>>>>> +#endif
>>>>> +#include "sysemu/sysemu.h"
>>>>> +#include "sysemu/tcg.h"
>>>>> +#include "sysemu/hw_accel.h"
>>>>> +#include "disas/capstone.h"
>>>>> +#include "fpu/softfloat.h"
>>>>
>>>> I guess some of those headers are not needed.
>>>
>>> Yeah, I just copy-n-pasted from the source file ... I'll check what can
>>> be omitted (if this patch series has a chance at all...)
>>>
>>>>> +
>>>>> +#if !defined(CONFIG_USER_ONLY) || !defined(TARGET_AARCH64)
>>>>> +
>>>>> +static bool arm_v7m_cpu_exec_interrupt(CPUState *cs, int 
>>>>> interrupt_request)
>>>>> +{
>>>>> +    CPUClass *cc = CPU_GET_CLASS(cs);
>>>>> +    ARMCPU *cpu = ARM_CPU(cs);
>>>>> +    CPUARMState *env = &cpu->env;
>>>>> +    bool ret = false;
>>>>> +
>>>>> +    /*
>>>>> +     * ARMv7-M interrupt masking works differently than -A or -R.
>>>>> +     * There is no FIQ/IRQ distinction. Instead of I and F bits
>>>>> +     * masking FIQ and IRQ interrupts, an exception is taken only
>>>>> +     * if it is higher priority than the current execution priority
>>>>> +     * (which depends on state like BASEPRI, FAULTMASK and the
>>>>> +     * currently active exception).
>>>>> +     */
>>>>> +    if (interrupt_request & CPU_INTERRUPT_HARD
>>>>> +        && (armv7m_nvic_can_take_pending_exception(env->nvic))) {
>>>>
>>>> so what is the status wrt m_helper.c which stays unconditionally
>>>> compiled. m_helper functions seem to called from target/arm/translate.c
>>>> mostly. Have you abandoned the stub idea. It may be confusing to have 2
>>>> different helper files. At least a comment explaining where a new helper
>>>> shall go may be useful.
>>>
>>> All the HELPER() functions should definitely stay in m_helper.c. They
>>> are required for linking. Or do you prefer a stub file instead? Then we
>>> could maybe make the whole m_helper.c conditional in the Makefile.objs
>>> instead.
>>
>> I was simply referring to your previous approach:
>>
>> Applying [Qemu-devel] [RFC PATCH 3/3] target/arm: Make m_helper.c
>> optional via CONFIG_ARM_V7M seems to fix the issue
>> https://lists.gnu.org/archive/html/qemu-devel/2019-09/msg00333.html
>>
>> It seems to work fine as it removes the call to armv7m_nvic.c functions
>> so no need to move the interrupt controller code?
> 
> Yes, but then we either need stubs in a separate file, or have to put
> lots of #ifdefs into translate.c ... none of those solutions seem to be
> really perfect :-(
Yep, the separate stub file looked cleaner to me.

> Anyway, Philippe is currently respinning his series (I think), so I'll
> postpone my work now to avoid to interfere with him.
OK

Thanks

Eric
> 
>  Thomas
>