> This can crash at line:414, if either tuple is invalid memory(but I
> think it's not because we have already accessed this memory in above
> if check) or dtup is invalid (this is also not possible because
> brin_new_memtuple has already accessed this).
I was testing with the brin correlation patch [1] applied. I cannot
crash it without the patch either. I am sorry for not testing it
before. The patch make BRIN selectivity estimation function access
more information.
[1] https://www.postgresql.org/message-id/attachment/50164/brin-correlation-v3.patch