RFR (S): 8237777 "Dumping core ..." is shown despite claiming that "# No core dump will be written."

[gerard ziemski]

hi Yasumasa,


On 4/23/20 6:56 PM, Yasumasa Suenaga wrote:
> Hi Gerard,
>
> On 2020/04/24 0:31, gerard ziemski wrote:
>>

[trimmed]

>> The main point here is that rlimit==0 check must be handled early 
>> enough, so that we don't end up saying we will dump the core, when we 
>> might be unable to do so, regardless of what other extra mechanism 
>> say about the pattern of the core file name.
>>
>> Before I go off and code this, however, I'd like a consensus.
>
> I think /proc/sys/kernel/core_pattern should be checked (starts with 
> '|' or not) at first on Linux because some tools (e.g. ABRT) would 
> dump core image regardless of rlimit.

Hmm, that certainly complicates things a lot.

Thank you for the info.

I think I'll punt on much of the changes and take David's suggestion, 
but perhaps tiny bit step further - please see my reply to David email 
for more...


cheers