[systemd-devel] asset failure that looks like it's coming from systemd/

[Jonathan Kelly]

On 25/12/21 07:43, Jonathan Kelly wrote:
> On 25/12/21 01:23, Andy Pieters wrote:
>>
>>
>> On Fri, 24 Dec 2021 at 10:11, Jonathan Kelly <jonkelly at fastmail.fm> 
>> wrote:
>>
>>     On 24/12/21 19:45, Andy Pieters wrote:
>>>
>>>
>>>     On Fri, 24 Dec 2021 at 01:53, Jonathan Kelly
>>>     <jonkelly at fastmail.fm> wrote:
>>>
>>>         make[2]: Entering directory '/usr/local/src/unicon/uni/lib'
>>>         ../../bin/unicon -s -c gui.icn
>>>         Assertion '(size_t) r < n' failed at
>>>         src/basic/random-util.c:232,
>>>         function genuine_random_bytes(). Aborting.
>>>         make[2]: *** [../../Makedefs.uni:48: gui.u] Aborted (core
>>>         dumped)
>>>
>>>
>>>     What points you to suspect systems?
>>
>>
>>     the line from the error
>>
>>     Assertion '(size_t) r < n' failed at src/basic/random-util.c:232,
>>
>>     that's from systemd.
>>
>>
>>     https://github.com/systemd/systemd/blob/main/src/basic/random-util.c
>>
>>     look at line 232
>>
>>
>>>
>>>     Have you looked at the coredump? (i.e. coredumpctl)
>>
>>     No. I'm in the middle here ... trying to compile unicon and it
>>     borks ... the unicon people don't know anything about
>>     random-util.c ... it not part of their software.
>>
>>     If it comes to it, I can look at more technical sleuthing, but
>>     I'd need some direction.
>>
>>
>>     Jonathan.
>>
>>
>> After googling this for 1 minute I came across this [1]
>>
>> Arch AUR package is out of date, use  the GIT version as documented 
>> in [1]
>>
>> Not the fault of systemd, the fault of using an old package. They 
>> explain that they stopped using svn in 2019, and having just 
>> downloaded the Arch AUR package it shows svn checkout in the prepare 
>> function...
>>
>> [1] 
>> https://sourceforge.net/p/unicon/discussion/starters/thread/4ea4931e8e/?limit=25 
>>
>
>
> I'm not using the AUR, I'm cloning from GIT
>
>
> git clone git://git.code.sf.net/p/unicon/unicon
> ./configure
> make
>
>
> I do know how to use google: that's how I found out 
> src/basic/random-util.c was part of systemd
>
>
> Jonathan.
>
Just wanted to add, I'm not assuming it's JUST a problem with systemd, 
but rather looking for some clues that I can feed back to the unicon people.

Though, it is an assert that is failing. From looking at the code, it 
seems getrandom() is returning more than requested, which is the size of 
a buffer - that doesn't sound good.

In any case, could someone say what would be triggering a call to 
genuine_random_bytes()?


Thanks

Jonathan.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.freedesktop.org/archives/systemd-devel/attachments/20211225/6e55f90e/attachment.htm>