Re: Consistently use palloc_object() and palloc_array()

Michael Paquier <michael@paquier.xyz>

From: Michael Paquier <michael@paquier.xyz>

To: David Geier <geidav.pg@gmail.com>

Cc: Tom Lane <tgl@sss.pgh.pa.us>, Thomas Munro <thomas.munro@gmail.com>, Chao Li <li.evan.chao@gmail.com>, pgsql-hackers <pgsql-hackers@postgresql.org>

Date: 2025-12-09T22:37:34Z

Lists: pgsql-hackers

Commits

Same data as JSON: GET /api/v1/messages/:b64id/commits the thread's linked commits as JSON, with link sources. API reference →

btree_gist: Fix memory allocation formula
- 5cf03552fbb4 19 (unreleased) landed
Use palloc_object() and palloc_array(), the last change
- 4f7dacc5b82a 19 (unreleased) landed
pg_buffercache: Fix memory allocation formula
- 580b5c2f397f 18.2 landed
- 3f83de20ba2e 19 (unreleased) landed
Fix allocation formula in llvmjit_expr.c
- 0c67dbcc4e39 14.21 landed
- 07ddf6197b78 15.16 landed
- 5a4dc4aabd03 16.12 landed
- 0bab0c3b74af 17.8 landed
- 5b7bbf16db34 18.2 landed
- 06761b6096b6 19 (unreleased) landed
Use palloc_object() and palloc_array() in backend code
- 1b105f9472bd 19 (unreleased) landed
Use palloc_object() and palloc_array() in more areas of the tree
- 0c3c5c3b06a3 19 (unreleased) landed
Use more palloc_object() and palloc_array() in contrib/
- 31d3847a37be 19 (unreleased) landed

Attachments

backend_files.txt (text/plain)

On Fri, Dec 05, 2025 at 04:41:41PM +0900, Michael Paquier wrote:
> Thanks.  That's a lot to digest.

Digesting a bit more now..

-   char       *ret = palloc(sizeof(buf));
+   char       *ret = palloc_array(char, sizeof(buf))

This one in dumputils.c is right, but I am not sure that it is an
improvement compared to the statu-quo.

Here is a list of the files where I have noticed tha addition of
casts, which are not required anymore:
brin_tuple.c.
gistbuildbuffers.c.
genam.c.
nbtdedup.c.
nbtree.c.
nbtsort.c.
index.c
execGrouping.c
execMain.c
relnode.c
partbounds.c
mcv.c
spell.c
arrayfuncs.c
partcache.c

Perhaps you have used a semi-automatic process and missed these?
Okay, some of these relied on a "Data" structure for the size vs
pointer for the allocation, but the results are the same.

-       IndexOrderByDistance *distances =
-           palloc(sizeof(distances[0]) * so->numberOfOrderBys);

Okay, this one in spgscan.c is not the usual project style.  Correct,
still funky.

            b_checkargnulls =
-               palloc(sizeof(LLVMBasicBlockRef *) * op->d.func.nargs);
+               palloc_array(LLVMBasicBlockRef *, op->d.func.nargs);

This one in llvmjit_expr.c was causing a compilation failure.  I am
not exactly sure why, but discarded for now.  I got a reproduction
locally as well as in the CI.

-   node->tsm_state = palloc0(sizeof(BernoulliSamplerData));
+   node->tsm_state = palloc0_object(BernoulliSamplerData); 

One can argue that this one in bernouilli.c is not really necessary,
tsm_state is actually a void *.

Among the 300-ish files changed in the backend, 48 of them had their
builds slightly change.  The list of them is attached.

So that's everything for the trivial changes, with 4601e7f1c708 for
src/backend/ and 0c3c5c3b06a3 for the rest.
--
Michael