2016-03-29 06:42:07

by Daniel Wagner

Subject: [PATCH] scripts/gdb: Fix loading of modules information

From: Daniel Wagner <[email protected]>

The module_core symbols is not available due to

7523e4dc5057 ("module: use a structure to encapsulate layout.")

The base address of the module is now stored inside
struct module_layout.

Signed-off-by: Daniel Wagner <[email protected]>
Cc: Jiri Kosina <[email protected]>
Cc: Rusty Russell <[email protected]>
---
scripts/gdb/linux/modules.py | 3 ++-
scripts/gdb/linux/symbols.py | 3 ++-
2 files changed, 4 insertions(+), 2 deletions(-)

diff --git a/scripts/gdb/linux/modules.py b/scripts/gdb/linux/modules.py
index 25db8cf..cffe078 100644
--- a/scripts/gdb/linux/modules.py
+++ b/scripts/gdb/linux/modules.py
@@ -74,7 +74,8 @@ class LxLsmod(gdb.Command):

for module in module_list():
gdb.write("{address} {name:<19} {size:>8} {ref}".format(
- address=str(module['module_core']).split()[0],
+ module_layout=module['core_layout']
+ addressr=str(module_layout['base'])
name=module['name'].string(),
size=str(module['core_size']),
ref=str(module['refcnt']['counter'])))
diff --git a/scripts/gdb/linux/symbols.py b/scripts/gdb/linux/symbols.py
index 627750c..8650ede 100644
--- a/scripts/gdb/linux/symbols.py
+++ b/scripts/gdb/linux/symbols.py
@@ -108,7 +108,8 @@ lx-symbols command."""

def load_module_symbols(self, module):
module_name = module['name'].string()
- module_addr = str(module['module_core']).split()[0]
+ module_layout = module['core_layout']
+ module_addr = str(module_layout['base'])

module_file = self._get_module_file(module_name)
if not module_file and not self.module_files_updated:
--
2.5.5

2016-03-29 06:58:42

by Daniel Wagner

Subject: [PATCH v2] scripts/gdb: Fix loading of modules information

From: Daniel Wagner <[email protected]>

The module_core symbols is not available due to

7523e4dc5057 ("module: use a structure to encapsulate layout.")

The base address of the module is now stored inside
struct module_layout.

Signed-off-by: Daniel Wagner <[email protected]>
Cc: Jiri Kosina <[email protected]>
Cc: Rusty Russell <[email protected]>
---
Obviously, it's too early to send patches. This times it even works.

scripts/gdb/linux/modules.py | 3 ++-
scripts/gdb/linux/symbols.py | 3 ++-
2 files changed, 4 insertions(+), 2 deletions(-)

diff --git a/scripts/gdb/linux/modules.py b/scripts/gdb/linux/modules.py
index 25db8cf..a1ec0f3 100644
--- a/scripts/gdb/linux/modules.py
+++ b/scripts/gdb/linux/modules.py
@@ -73,8 +73,9 @@ class LxLsmod(gdb.Command):
" " if utils.get_long_type().sizeof == 8 else ""))

for module in module_list():
+ module_layout=module['core_layout']
gdb.write("{address} {name:<19} {size:>8} {ref}".format(
- address=str(module['module_core']).split()[0],
+ address=str(module_layout['base']),
name=module['name'].string(),
size=str(module['core_size']),
ref=str(module['refcnt']['counter'])))
diff --git a/scripts/gdb/linux/symbols.py b/scripts/gdb/linux/symbols.py
index 627750c..8650ede 100644
--- a/scripts/gdb/linux/symbols.py
+++ b/scripts/gdb/linux/symbols.py
@@ -108,7 +108,8 @@ lx-symbols command."""

def load_module_symbols(self, module):
module_name = module['name'].string()
- module_addr = str(module['module_core']).split()[0]
+ module_layout = module['core_layout']
+ module_addr = str(module_layout['base'])

module_file = self._get_module_file(module_name)
if not module_file and not self.module_files_updated:
--
2.5.5

2016-03-29 07:00:10

by Jan Kiszka

Subject: Re: [PATCH] scripts/gdb: Fix loading of modules information

On 2016-03-29 08:41, Daniel Wagner wrote:
> From: Daniel Wagner <[email protected]>
>
> The module_core symbols is not available due to
>
> 7523e4dc5057 ("module: use a structure to encapsulate layout.")
>
> The base address of the module is now stored inside
> struct module_layout.
>
> Signed-off-by: Daniel Wagner <[email protected]>
> Cc: Jiri Kosina <[email protected]>
> Cc: Rusty Russell <[email protected]>
> ---
> scripts/gdb/linux/modules.py | 3 ++-
> scripts/gdb/linux/symbols.py | 3 ++-
> 2 files changed, 4 insertions(+), 2 deletions(-)
>
> diff --git a/scripts/gdb/linux/modules.py b/scripts/gdb/linux/modules.py
> index 25db8cf..cffe078 100644
> --- a/scripts/gdb/linux/modules.py
> +++ b/scripts/gdb/linux/modules.py
> @@ -74,7 +74,8 @@ class LxLsmod(gdb.Command):
>
> for module in module_list():
> gdb.write("{address} {name:<19} {size:>8} {ref}".format(
> - address=str(module['module_core']).split()[0],
> + module_layout=module['core_layout']
> + addressr=str(module_layout['base'])

This looks untested.

> name=module['name'].string(),
> size=str(module['core_size']),
> ref=str(module['refcnt']['counter'])))
> diff --git a/scripts/gdb/linux/symbols.py b/scripts/gdb/linux/symbols.py
> index 627750c..8650ede 100644
> --- a/scripts/gdb/linux/symbols.py
> +++ b/scripts/gdb/linux/symbols.py
> @@ -108,7 +108,8 @@ lx-symbols command."""
>
> def load_module_symbols(self, module):
> module_name = module['name'].string()
> - module_addr = str(module['module_core']).split()[0]
> + module_layout = module['core_layout']
> + module_addr = str(module_layout['base'])
>
> module_file = self._get_module_file(module_name)
> if not module_file and not self.module_files_updated:
>

Thanks, but ad4db3b24a should already be on its way into stable.

Jan

--
Siemens AG, Corporate Technology, CT RDA ITP SES-DE
Corporate Competence Center Embedded Linux

2016-03-29 07:13:58

by Daniel Wagner

Subject: Re: [PATCH] scripts/gdb: Fix loading of modules information

On 03/29/2016 08:59 AM, Jan Kiszka wrote:
> On 2016-03-29 08:41, Daniel Wagner wrote:
>> for module in module_list():
>> gdb.write("{address} {name:<19} {size:>8} {ref}".format(
>> - address=str(module['module_core']).split()[0],
>> + module_layout=module['core_layout']
>> + addressr=str(module_layout['base'])
>
> This looks untested.

Yep, I send out the untested version. Need more coffee...

>> name=module['name'].string(),
>> size=str(module['core_size']),
>> ref=str(module['refcnt']['counter'])))
>> diff --git a/scripts/gdb/linux/symbols.py b/scripts/gdb/linux/symbols.py
>> index 627750c..8650ede 100644
>> --- a/scripts/gdb/linux/symbols.py
>> +++ b/scripts/gdb/linux/symbols.py
>> @@ -108,7 +108,8 @@ lx-symbols command."""
>>
>> def load_module_symbols(self, module):
>> module_name = module['name'].string()
>> - module_addr = str(module['module_core']).split()[0]
>> + module_layout = module['core_layout']
>> + module_addr = str(module_layout['base'])
>>
>> module_file = self._get_module_file(module_name)
>> if not module_file and not self.module_files_updated:
>>
>
> Thanks, but ad4db3b24a should already be on its way into stable.

Ah, I haven't got that fix yet because I am still on a older branch.
Also my google-foo didn't help.

Sorry about that noise.

cheers,
daniel

2016-03-29 07:23:11

by Jan Kiszka

Subject: Re: [PATCH] scripts/gdb: Fix loading of modules information

On 2016-03-29 09:13, Daniel Wagner wrote:
> On 03/29/2016 08:59 AM, Jan Kiszka wrote:
>> On 2016-03-29 08:41, Daniel Wagner wrote:
>>> for module in module_list():
>>> gdb.write("{address} {name:<19} {size:>8} {ref}".format(
>>> - address=str(module['module_core']).split()[0],
>>> + module_layout=module['core_layout']
>>> + addressr=str(module_layout['base'])
>>
>> This looks untested.
>
> Yep, I send out the untested version. Need more coffee...
>
>>> name=module['name'].string(),
>>> size=str(module['core_size']),
>>> ref=str(module['refcnt']['counter'])))
>>> diff --git a/scripts/gdb/linux/symbols.py b/scripts/gdb/linux/symbols.py
>>> index 627750c..8650ede 100644
>>> --- a/scripts/gdb/linux/symbols.py
>>> +++ b/scripts/gdb/linux/symbols.py
>>> @@ -108,7 +108,8 @@ lx-symbols command."""
>>>
>>> def load_module_symbols(self, module):
>>> module_name = module['name'].string()
>>> - module_addr = str(module['module_core']).split()[0]
>>> + module_layout = module['core_layout']
>>> + module_addr = str(module_layout['base'])
>>>
>>> module_file = self._get_module_file(module_name)
>>> if not module_file and not self.module_files_updated:
>>>
>>
>> Thanks, but ad4db3b24a should already be on its way into stable.
>
> Ah, I haven't got that fix yet because I am still on a older branch.
> Also my google-foo didn't help.
>
> Sorry about that noise.

No problem. At least we now know another frequent user.

BTW, Kieran and some other people at Linaro started to think about how
to auto-test for such regressions - probably the only reliable way to
catch them.

Jan

--
Siemens AG, Corporate Technology, CT RDA ITP SES-DE
Corporate Competence Center Embedded Linux

2016-03-31 06:17:28

by Daniel Wagner

Subject: Re: [PATCH] scripts/gdb: Fix loading of modules information

On 03/29/2016 09:22 AM, Jan Kiszka wrote:
> No problem. At least we now know another frequent user.

Let's say sometimes :)

> BTW, Kieran and some other people at Linaro started to think about how
> to auto-test for such regressions - probably the only reliable way to
> catch them.

Good idea. I am happy to test if you got something.

cheers,
daniel