Re: [PATCH v2] [gdb] Detect corrupt link map with zero l_ld and l_addr

[Tom de Vries <tdevries@suse.de>]

On 9/28/26 5:30 PM, Tom de Vries wrote:
>> Returning 0 means we stop reading the shared library list, is that
>> expected?  The known case of this isn't really that the shared library
>> list is corrupted, just that this entry is incomplete.  We could maybe
>> skip over this entry and keep reading the list, to be able to present
>> all the loaded shared libraries to the user?
>>
> 
> Hi Simon,
> 
> thanks for the review.
> 
> I think you're right, using continue instead of return 0 is more 
> appropriate.
> 
>> Is the warning printed every time the library list is read (like at each
>> stop)?  Does that become annoying for the user?
>>
>> In the end maybe we don't care too much about these concerns, because
>> it's for one specific version of one specific distro, and only when
>> debugging programs that use multiple namespaces (which are quite rare I
>> think).
>>
> 
> It could be annoying, agreed.
> 
> I'm treating this though as a corruption that is triggered by one 
> specific version of one specific distro, but might happen otherwise.
> 
> So I think there is information in how many times the warning triggers.
> 
> We could keep track for which shared libs the warning triggers, and emit 
> a summary instead.  But I don't think it's worth the bother at this point.
> 

I've tested the continue approach, and submitted a v3 ( 
https://sourceware.org/pipermail/gdb-patches/2026-September/230704.html ).

Thanks,
- Tom

> Thanks,
> - Tom