Currently the migration code incorrectly treats a subsection with
no .needed function pointer as if it was the subsection list
terminator -- it is ignored and so is everything after it.
Work around this by giving various M profile vmstate structs
a 'needed' function that always returns true.
We reuse m_needed() for this, since it's always true here.

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
---
Not strictly a regression as it only affects M profile CPUs
with the security extensions, and migration of those was
broken anyway in 2.12 due to a different bug.
---
 target/arm/machine.c | 3 +++
 1 file changed, 3 insertions(+)

diff --git a/target/arm/machine.c b/target/arm/machine.c
index 2e28d086bdf..ff4ec22bf75 100644
--- a/target/arm/machine.c
+++ b/target/arm/machine.c
@@ -184,6 +184,7 @@ static const VMStateDescription vmstate_m_faultmask_primask = {
     .name = "cpu/m/faultmask-primask",
     .version_id = 1,
     .minimum_version_id = 1,
+    .needed = m_needed,
     .fields = (VMStateField[]) {
         VMSTATE_UINT32(env.v7m.faultmask[M_REG_NS], ARMCPU),
         VMSTATE_UINT32(env.v7m.primask[M_REG_NS], ARMCPU),
@@ -230,6 +231,7 @@ static const VMStateDescription vmstate_m_scr = {
     .name = "cpu/m/scr",
     .version_id = 1,
     .minimum_version_id = 1,
+    .needed = m_needed,
     .fields = (VMStateField[]) {
         VMSTATE_UINT32(env.v7m.scr[M_REG_NS], ARMCPU),
         VMSTATE_END_OF_LIST()
@@ -240,6 +242,7 @@ static const VMStateDescription vmstate_m_other_sp = {
     .name = "cpu/m/other-sp",
     .version_id = 1,
     .minimum_version_id = 1,
+    .needed = m_needed,
     .fields = (VMStateField[]) {
         VMSTATE_UINT32(env.v7m.other_sp, ARMCPU),
         VMSTATE_END_OF_LIST()
-- 
2.17.1

Peter Maydell <peter.maydell@linaro.org> wrote:
> Currently the migration code incorrectly treats a subsection with
> no .needed function pointer as if it was the subsection list
> terminator -- it is ignored and so is everything after it.
> Work around this by giving various M profile vmstate structs
> a 'needed' function that always returns true.
> We reuse m_needed() for this, since it's always true here.
>
> Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
> Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>

Reviewed-by: Juan Quintela <quintela@redhat.com>


> ---
> Not strictly a regression as it only affects M profile CPUs
> with the security extensions, and migration of those was
> broken anyway in 2.12 due to a different bug.
> ---
>  target/arm/machine.c | 3 +++
>  1 file changed, 3 insertions(+)
>
> diff --git a/target/arm/machine.c b/target/arm/machine.c
> index 2e28d086bdf..ff4ec22bf75 100644
> --- a/target/arm/machine.c
> +++ b/target/arm/machine.c
> @@ -184,6 +184,7 @@ static const VMStateDescription vmstate_m_faultmask_primask = {
>      .name = "cpu/m/faultmask-primask",
>      .version_id = 1,
>      .minimum_version_id = 1,
> +    .needed = m_needed,
>      .fields = (VMStateField[]) {
>          VMSTATE_UINT32(env.v7m.faultmask[M_REG_NS], ARMCPU),
>          VMSTATE_UINT32(env.v7m.primask[M_REG_NS], ARMCPU),
> @@ -230,6 +231,7 @@ static const VMStateDescription vmstate_m_scr = {
>      .name = "cpu/m/scr",
>      .version_id = 1,
>      .minimum_version_id = 1,
> +    .needed = m_needed,
>      .fields = (VMStateField[]) {
>          VMSTATE_UINT32(env.v7m.scr[M_REG_NS], ARMCPU),
>          VMSTATE_END_OF_LIST()
> @@ -240,6 +242,7 @@ static const VMStateDescription vmstate_m_other_sp = {
>      .name = "cpu/m/other-sp",
>      .version_id = 1,
>      .minimum_version_id = 1,
> +    .needed = m_needed,
>      .fields = (VMStateField[]) {
>          VMSTATE_UINT32(env.v7m.other_sp, ARMCPU),
>          VMSTATE_END_OF_LIST()

But having 3 subsections with the same needed function ... we could have
a single subsection.

Later, Juan.

On 7 August 2018 at 15:40, Juan Quintela <quintela@redhat.com> wrote:
> But having 3 subsections with the same needed function ... we could have
> a single subsection.

I wanted to keep the change small here since we're very
late in the release cycle, so I kept it to a change
that restores the intent of the code, ie "always transfer
the subsection".

thanks
-- PMM

Peter Maydell <peter.maydell@linaro.org> wrote:
> On 7 August 2018 at 15:40, Juan Quintela <quintela@redhat.com> wrote:
>> But having 3 subsections with the same needed function ... we could have
>> a single subsection.
>
> I wanted to keep the change small here since we're very
> late in the release cycle, so I kept it to a change
> that restores the intent of the code, ie "always transfer
> the subsection".

ok with minimizing changes.  But I can't understand why that is going
that way.  Seeing this, I am making a patch that at registration time
check that we are not doing something stupid, like not having needed
function, bad names, whatever.

Thanks.