Re: [PATCH] scripts/gdb/linux/symbols.py: address changes to module_sect_attrs

[Antonio Quartulli]

On 21/02/2025 17:52, Jan Kiszka wrote:
> On 21.02.25 14:03, Antonio Quartulli wrote:
> > When loading symbols from kernel modules we used to iterate
> > from 0 to module_sect_attrs::nsections, in order to
> > retrieve their name and address.
> >
> > However module_sect_attrs::nsections has been removed from
> > the struct by a previous commit.
> >
> > Re-arrange the iteration by accessing all items in
> > module_sect_attrs::grp::bin_attrs[] until NULL is found
> > (it's a NULL terminated array).
> >
> > At the same time the symbol address cannot be extracted
> > from module_sect_attrs::attrs[]::address anymore because
> > it has also been deleted. Fetch it from
> > module_sect_attrs::grp::bin_attrs[]::private as described
> > in 4b2c11e4aaf7.
> >
> > Fixes: d8959b947a8d ("module: sysfs: Drop member 'module_sect_attrs::nsections'")
> > Fixes: 4b2c11e4aaf7 ("module: sysfs: Drop member 'module_sect_attr::address'")
> > Cc: Thomas Weißschuh <linux@xxxxxxxxxxxxxx>
> > Signed-off-by: Antonio Quartulli <antonio@xxxxxxxxxxxxx>
> > ---
> >   scripts/gdb/linux/symbols.py | 13 +++++++++----
> >   1 file changed, 9 insertions(+), 4 deletions(-)
> >
> > diff --git a/scripts/gdb/linux/symbols.py b/scripts/gdb/linux/symbols.py
> > index f6c1b063775a..610a3dd3c7b4 100644
> > --- a/scripts/gdb/linux/symbols.py
> > +++ b/scripts/gdb/linux/symbols.py
> > @@ -15,6 +15,7 @@ import gdb
> >   import os
> >   import re
> >   
> > +from itertools import count
> >   from linux import modules, utils, constants
> >   
> >   
> > @@ -95,10 +96,14 @@ lx-symbols command."""
> >           except gdb.error:
> >               return str(module_addr)
> >   
> > -        attrs = sect_attrs['attrs']
> > -        section_name_to_address = {
> > -            attrs[n]['battr']['attr']['name'].string(): attrs[n]['address']
> > -            for n in range(int(sect_attrs['nsections']))}
> > +        section_name_to_address = {}
> > +        for i in count():
> > +            # this is a NULL terminated array
> > +            if sect_attrs['grp']['bin_attrs'][i] == 0x0:
> > +                break
> > +
> > +            attr = sect_attrs['grp']['bin_attrs'][i].dereference()
> > +            section_name_to_address[attr['attr']['name']] = attr['private']
>
> You dropped that .string() from the name - I don't remember the details
> anymore but we have it all around when picking up strings from C
> structures. Was there a particular reason to do that?

Ouch. That was not intentional and my test did not explode, therefore I 
assumed the code was correct.

I presume it may explode if 'name' is changed into something not a 
string. In this case .string() would throw an exception and block the 
execution.

I will send v2 shortly with .string().

> >   
> >           textaddr = section_name_to_address.get(".text", module_addr)
> >           args = []
>
> Thanks for picking up this task so quickly!

Well, I needed gdb to break into ovpn.ko :-)

Cheers,

--
Antonio Quartulli

CEO and Co-Founder
Mandelbit Srl
https://www.mandelbit.com