From: Andres Freund <andres@anarazel.de>
Date: Mon, 24 Oct 2022 16:44:16 -0700
Subject: [PATCH v5 08/15] bufmgr: Support multiple in-progress IOs by using
 resowner

Commit message should describe why we couldn't support multiple
in-progress IOs before, I think (e.g. we couldn't be sure that we
cleared IO_IN_PROGRESS if something happened).

@@ -4709,8 +4704,6 @@ TerminateBufferIO(BufferDesc *buf, bool
clear_dirty, uint32 set_flag_bits)
 {
     uint32        buf_state;

I noticed that the comment above TermianteBufferIO() says
 * TerminateBufferIO: release a buffer we were doing I/O on
 *    (Assumptions)
 *    My process is executing IO for the buffer

Can we still say this is an assumption? What about when it is being
cleaned up after being called from AbortBufferIO()

diff --git a/src/backend/utils/resowner/resowner.c
b/src/backend/utils/resowner/resowner.c
index 19b6241e45..fccc59b39d 100644
--- a/src/backend/utils/resowner/resowner.c
+++ b/src/backend/utils/resowner/resowner.c
@@ -121,6 +121,7 @@ typedef struct ResourceOwnerData

     /* We have built-in support for remembering: */
     ResourceArray bufferarr;    /* owned buffers */
+    ResourceArray bufferioarr;    /* in-progress buffer IO */
     ResourceArray catrefarr;    /* catcache references */
     ResourceArray catlistrefarr;    /* catcache-list pins */
     ResourceArray relrefarr;    /* relcache references */
@@ -441,6 +442,7 @@ ResourceOwnerCreate(ResourceOwner parent, const char *name)

Maybe worth mentioning in-progress buffer IO in resowner README? I know
it doesn't claim to be exhaustive, so, up to you.

Also, I realize that existing code in this file has the extraneous
parantheses, but maybe it isn't worth staying consistent with that?
as in: &(owner->bufferioarr)

+ */
+void
+ResourceOwnerRememberBufferIO(ResourceOwner owner, Buffer buffer)
+{
+    ResourceArrayAdd(&(owner->bufferioarr), BufferGetDatum(buffer));
+}
+