From: | Michael Paquier <michael(dot)paquier(at)gmail(dot)com> |
---|---|
To: | PostgreSQL mailing lists <pgsql-hackers(at)postgresql(dot)org> |
Subject: | NULL checks of deferenced pointers in picksplit method of intarray |
Date: | 2015-01-30 12:06:45 |
Message-ID: | CAB7nPqRr+kWGutuFR+OLhZir=81h8MAPoTjw7ijktTNwCi4Q9Q@mail.gmail.com |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-hackers |
Hi all,
Coverity is pointing out that _int_split.c has unnecessary checks for
deferenced pointers in 5 places.
- if (inter_d != (ArrayType *) NULL)
- pfree(inter_d);
+ pfree(inter_d);
In this case inter_d is generated by inner_int_inter, a routine that
always generates an ArrayType with at least new_intArrayType.
In two places there is as well this pattern:
- if (datum_l)
- pfree(datum_l);
- if (union_dr)
- pfree(union_dr);
+ pfree(datum_l);
+ pfree(union_dr);
And that one:
- if (datum_r)
- pfree(datum_r);
- if (union_dl)
- pfree(union_dl);
+ pfree(datum_r);
+ pfree(union_dl);
union_dr and union_dl are generated by inner_int_union which never
returns NULL. Similarly, datum_r and datum_l are created with
copy_intArrayType the first time, which never returns NULL, and their
values are changed at each loop step. Also, as far as I understood
from this code, no elements manipulated are NULL, perhaps this is
worth an assertion?
Attached is a patch to adjust those things.
Regards,
--
Michael
Attachment | Content-Type | Size |
---|---|---|
20150130_intarray_fix_dereferences.patch | text/x-diff | 1.1 KB |
From | Date | Subject | |
---|---|---|---|
Next Message | Sawada Masahiko | 2015-01-30 12:38:10 | Re: Proposal: knowing detail of config files via SQL |
Previous Message | David Rowley | 2015-01-30 11:37:19 | Re: Performance improvement for joins where outer side is unique |