2019-02-15 16:12:04

by Felipe Balbi

[permalink] [raw]
Subject: [PATCH] scripts: gdb: redefine MS_RDONLY

Since commit e262e32d6bde0f77fb0c95d977482fc872c51996 "vfs: Suppress
MS_* flag defs within the kernel unless explicitly enabled", the
generated gdb scripts can't be used anymore. That commit moved the
definition of MS_* flags but forgot to update gdb scripts to use the
new location. This patch includes <uapi/linux/mount.h> to
constants.py.in so gdb scripts are functional again.

Fixes: e262e32d6bde "vfs: Suppress MS_* flag defs within the kernel unless explicitly enabled"
Signed-off-by: Felipe Balbi <[email protected]>
---
scripts/gdb/linux/constants.py.in | 1 +
1 file changed, 1 insertion(+)

diff --git a/scripts/gdb/linux/constants.py.in b/scripts/gdb/linux/constants.py.in
index 7aad82406422..c2a93805d911 100644
--- a/scripts/gdb/linux/constants.py.in
+++ b/scripts/gdb/linux/constants.py.in
@@ -13,6 +13,7 @@
*/

#include <linux/fs.h>
+#include <uapi/linux/mount.h>
#include <linux/mount.h>
#include <linux/of_fdt.h>

--
2.20.1



2019-02-16 07:21:39

by Kieran Bingham

[permalink] [raw]
Subject: Re: [PATCH] scripts: gdb: redefine MS_RDONLY

Hi Felipe,

Thank you for the patch,

On 15/02/2019 11:29, Felipe Balbi wrote:
> Since commit e262e32d6bde0f77fb0c95d977482fc872c51996 "vfs: Suppress
> MS_* flag defs within the kernel unless explicitly enabled", the
> generated gdb scripts can't be used anymore. That commit moved the
> definition of MS_* flags but forgot to update gdb scripts to use the
> new location. This patch includes <uapi/linux/mount.h> to
> constants.py.in so gdb scripts are functional again.
>
> Fixes: e262e32d6bde "vfs: Suppress MS_* flag defs within the kernel unless explicitly enabled"
> Signed-off-by: Felipe Balbi <[email protected]>
> ---
> scripts/gdb/linux/constants.py.in | 1 +
> 1 file changed, 1 insertion(+)
>
> diff --git a/scripts/gdb/linux/constants.py.in b/scripts/gdb/linux/constants.py.in
> index 7aad82406422..c2a93805d911 100644
> --- a/scripts/gdb/linux/constants.py.in
> +++ b/scripts/gdb/linux/constants.py.in
> @@ -13,6 +13,7 @@
> */
>
> #include <linux/fs.h>

Do we still need linux/fs.h?

> +#include <uapi/linux/mount.h>

Perhaps this should go at the end to maintain alphabetical ordering?

> #include <linux/mount.h>
> #include <linux/of_fdt.h>
>

We can see from Dan's patch - that the comment above the value
definitions also needs to be updated...

Felipe, you won the patch race - how would you like to handle this? Will
you resubmit with changes? or should we just add the missing Fixes: tag
to Dan's patch?


--
Regards
--
Kieran

2019-02-18 07:18:54

by Felipe Balbi

[permalink] [raw]
Subject: Re: [PATCH] scripts: gdb: redefine MS_RDONLY


Hi,

Kieran Bingham <[email protected]> writes:
> Hi Felipe,
>
> Thank you for the patch,
>
> On 15/02/2019 11:29, Felipe Balbi wrote:
>> Since commit e262e32d6bde0f77fb0c95d977482fc872c51996 "vfs: Suppress
>> MS_* flag defs within the kernel unless explicitly enabled", the
>> generated gdb scripts can't be used anymore. That commit moved the
>> definition of MS_* flags but forgot to update gdb scripts to use the
>> new location. This patch includes <uapi/linux/mount.h> to
>> constants.py.in so gdb scripts are functional again.
>>
>> Fixes: e262e32d6bde "vfs: Suppress MS_* flag defs within the kernel unless explicitly enabled"
>> Signed-off-by: Felipe Balbi <[email protected]>
>> ---
>> scripts/gdb/linux/constants.py.in | 1 +
>> 1 file changed, 1 insertion(+)
>>
>> diff --git a/scripts/gdb/linux/constants.py.in b/scripts/gdb/linux/constants.py.in
>> index 7aad82406422..c2a93805d911 100644
>> --- a/scripts/gdb/linux/constants.py.in
>> +++ b/scripts/gdb/linux/constants.py.in
>> @@ -13,6 +13,7 @@
>> */
>>
>> #include <linux/fs.h>
>
> Do we still need linux/fs.h?

don't know enough about the tool to judge that. Just added the minimal
change to get it working again.

>> +#include <uapi/linux/mount.h>
>
> Perhaps this should go at the end to maintain alphabetical ordering?
>
>> #include <linux/mount.h>
>> #include <linux/of_fdt.h>
>>
>
> We can see from Dan's patch - that the comment above the value
> definitions also needs to be updated...
>
> Felipe, you won the patch race - how would you like to handle this? Will
> you resubmit with changes? or should we just add the missing Fixes: tag
> to Dan's patch?

I don't mind either way, as long as the GDB scripts are working upstream :-)

--
balbi


Attachments:
signature.asc (847.00 B)