[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[Qemu-devel] Re: [PATCH 07/11] eeprom93xx: Use the new hack macro to avo

From: Juan Quintela
Subject: [Qemu-devel] Re: [PATCH 07/11] eeprom93xx: Use the new hack macro to avoid duplicate field names
Date: Wed, 23 Mar 2011 10:58:04 +0100
User-agent: Gnus/5.13 (Gnus v5.13) Emacs/23.2 (gnu/linux)

Anthony Liguori <address@hidden> wrote:
> I don't fully understand this hack business but we need field to be unique 
> so..
> Signed-off-by: Anthony Liguori <address@hidden>
> ---
>  hw/eeprom93xx.c |    2 +-
>  1 files changed, 1 insertions(+), 1 deletions(-)
> diff --git a/hw/eeprom93xx.c b/hw/eeprom93xx.c
> index cfa695d..f1d75ec 100644
> --- a/hw/eeprom93xx.c
> +++ b/hw/eeprom93xx.c
> @@ -114,7 +114,7 @@ static const VMStateInfo vmstate_hack_uint16_from_uint8 = 
> {
>  };
>  #define VMSTATE_UINT16_HACK_TEST(_f, _s, _t)                           \
> -    VMSTATE_SINGLE_TEST(_f, _s, _t, 0, vmstate_hack_uint16_from_uint8, 
> uint16_t)
> +    VMSTATE_SINGLE_TEST_HACK(_f, _s, _t, 0, vmstate_hack_uint16_from_uint8, 
> uint16_t)
>  static bool is_old_eeprom_version(void *opaque, int version_id)
>  {

After the fact, we need to promote it as "full types".

Basically it is needed when we sent a field with a different size that
we use it on the struct.

if we have

struct FOOState {
       int32_t bar;

and it is sent as

VMSTATE_INT8(bar, ....)

In this case, I went through the whole device, checed that int8_t was
enough and did the change.

But if we have:

struct FOOState {
       int8_t bar;

and it is sent as

VMSTATE_INT32(bar, ....)

Then it is not trivial :-(

We change FOOState to int32 or we break migration format.  Here is where
the _HACK suffix appeared.

I thought it was not going to be needed a lot, but there are several
devices that just sent everything over the wire as uint32, independently
of its type.

Later, Juan.

reply via email to

[Prev in Thread] Current Thread [Next in Thread]