[v2] Use macro instead of plain text

There is TYPE_ICH9_AHCI definition in ahci.h when QOMify it, seems it
is missed.

Signed-off-by: Cao jin <caoj.fnst@cn.fujitsu.com>
---
 hw/i386/pc_q35.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

Cao jin <caoj.fnst@cn.fujitsu.com> writes:

> There is TYPE_ICH9_AHCI definition in ahci.h when QOMify it, seems it
> is missed.
>
> Signed-off-by: Cao jin <caoj.fnst@cn.fujitsu.com>
> ---
>  hw/i386/pc_q35.c | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/hw/i386/pc_q35.c b/hw/i386/pc_q35.c
> index 9a12068..aa34a07 100644
> --- a/hw/i386/pc_q35.c
> +++ b/hw/i386/pc_q35.c
> @@ -249,7 +249,7 @@ static void pc_q35_init(MachineState *machine)
>      ahci = pci_create_simple_multifunction(host_bus,
>                                             PCI_DEVFN(ICH9_SATA1_DEV,
>                                                       ICH9_SATA1_FUNC),
> -                                           true, "ich9-ahci");
> +                                           true, TYPE_ICH9_AHCI);
>      idebus[0] = qdev_get_child_bus(&ahci->qdev, "ide.0");
>      idebus[1] = qdev_get_child_bus(&ahci->qdev, "ide.1");
>      g_assert(MAX_SATA_PORTS == ICH_AHCI(ahci)->ahci.ports);

I suspect there are more instances of the same problem.

To find all similar macro definitions:

    $ git-grep -Eh '^ *# *define +TYPE_[A-Za-z0-9_]+ +"'

Extract the strings:

    $ git-grep -Eh '^ *# *define +TYPE_[A-Za-z0-9_]+ +"' | sed 's/^[^"]*\("[^"]*"\).*/\1/'

Search the source for them:

    $ for i in `git-grep -Eh '^ *# *define +TYPE_[A-Za-z0-9_]+ +"' | sed 's/^[^"]*\("[^"]*"\).*/\1/'`; do echo "= $i ="; git-grep -F "$i"; done

Results in more than 1000 hits for me.  Names like "device" are of
course very prone to false positives.

Not sure how interested we are in consistent use of these macros...

On 18 December 2015 at 12:52, Markus Armbruster <armbru@redhat.com> wrote:
> Cao jin <caoj.fnst@cn.fujitsu.com> writes:
>
>> There is TYPE_ICH9_AHCI definition in ahci.h when QOMify it, seems it
>> is missed.

> I suspect there are more instances of the same problem.

> Not sure how interested we are in consistent use of these macros...

In a lot of cases we'd need to find somewhere appropriate to put the
TYPE_* macro which might well mean creating a new header file just
for the purpose. That seems like overkill...

thanks
-- PMM

On 12/18/2015 08:52 PM, Markus Armbruster wrote:
> Cao jin <caoj.fnst@cn.fujitsu.com> writes:
>
>> There is TYPE_ICH9_AHCI definition in ahci.h when QOMify it, seems it
>> is missed.
>>
>> Signed-off-by: Cao jin <caoj.fnst@cn.fujitsu.com>
>> ---
>>   hw/i386/pc_q35.c | 2 +-
>>   1 file changed, 1 insertion(+), 1 deletion(-)
>>
>> diff --git a/hw/i386/pc_q35.c b/hw/i386/pc_q35.c
>> index 9a12068..aa34a07 100644
>> --- a/hw/i386/pc_q35.c
>> +++ b/hw/i386/pc_q35.c
>> @@ -249,7 +249,7 @@ static void pc_q35_init(MachineState *machine)
>>       ahci = pci_create_simple_multifunction(host_bus,
>>                                              PCI_DEVFN(ICH9_SATA1_DEV,
>>                                                        ICH9_SATA1_FUNC),
>> -                                           true, "ich9-ahci");
>> +                                           true, TYPE_ICH9_AHCI);
>>       idebus[0] = qdev_get_child_bus(&ahci->qdev, "ide.0");
>>       idebus[1] = qdev_get_child_bus(&ahci->qdev, "ide.1");
>>       g_assert(MAX_SATA_PORTS == ICH_AHCI(ahci)->ahci.ports);
>
> I suspect there are more instances of the same problem.
>
> To find all similar macro definitions:
>
>      $ git-grep -Eh '^ *# *define +TYPE_[A-Za-z0-9_]+ +"'
>
> Extract the strings:
>
>      $ git-grep -Eh '^ *# *define +TYPE_[A-Za-z0-9_]+ +"' | sed 's/^[^"]*\("[^"]*"\).*/\1/'
>
> Search the source for them:
>
>      $ for i in `git-grep -Eh '^ *# *define +TYPE_[A-Za-z0-9_]+ +"' | sed 's/^[^"]*\("[^"]*"\).*/\1/'`; do echo "= $i ="; git-grep -F "$i"; done
>
> Results in more than 1000 hits for me.  Names like "device" are of
> course very prone to false positives.
>

Yes, often seeing "plain text" style when browsing code, and after 
seeing PMM`s explanation, I finally see why there are lots of these...

> Not sure how interested we are in consistent use of these macros...
>

I didn`t pay any attention to this before. It comes to me when I make 
the V2 "msi/msix" patch. as I said in that cover-letter: special case:)

we can forget this patch now.

>
> .
>

Hi Peter,

Thanks for the explanation, it make me aware that why there are lots of 
"plain_text_for_device":)

forget it please.

On 12/18/2015 09:01 PM, Peter Maydell wrote:
> On 18 December 2015 at 12:52, Markus Armbruster <armbru@redhat.com> wrote:
>> Cao jin <caoj.fnst@cn.fujitsu.com> writes:
>>
>>> There is TYPE_ICH9_AHCI definition in ahci.h when QOMify it, seems it
>>> is missed.
>
>> I suspect there are more instances of the same problem.
>
>> Not sure how interested we are in consistent use of these macros...
>
> In a lot of cases we'd need to find somewhere appropriate to put the
> TYPE_* macro which might well mean creating a new header file just
> for the purpose. That seems like overkill...
>
> thanks
> -- PMM
>
>
>