MIME-Version: 1.0
References: <6BC5DBAB-6084-4BB8-8450-52E9648AB021@gmail.com>
 <CAAKRu_ZCjHoRPfQ8AbMrFY8TOMCPAvZ0_m9SX7yg0edfTk45-g@mail.gmail.com>
 <7F5BCD7A-764D-4D8D-8E27-6F2CAAEA1CEE@gmail.com>
 <CAAKRu_a04jbDACwzRYwzDND31aPyf7Yvz9TAZrTr=+F5bK1aVA@mail.gmail.com>
 <CALdSSPjcv25jmXm29X-MRWZBae6+HwcWfVH1PE8NfD=EMTnkAg@mail.gmail.com>
 <CAAKRu_bwtBEzDwemyim1r6yYonw7FTyFr1HXG8vywCe-MdbPBQ@mail.gmail.com>
 <4379FDA3-9446-4E2C-9C15-32EFE8D4F31B@yandex-team.ru>
 <CAAKRu_YQd=2KvomM+RHcpeDKj0bq+peJ=3W-fip+pkvzA-Jq9w@mail.gmail.com>
 <7ib3sa55sapwjlaz4sijbiq7iezna27kjvvvar4dpgkmadml6t@gfpkkwmdnepx>
 <CAAKRu_bs+gZ83QDacmBxunPvCGnXJ05hxP2BDPJ3BGwdbGRXzg@mail.gmail.com>
 <bqc4kh5midfn44gnjiqez3bjqv4zogydguvdn446riw45jcf3y@4ez66il7ebvk>
 <CAAKRu_Y1MuANdm1p47Ev13Y9EQz8z+pw-vHOh=3DVdahUTjgXg@mail.gmail.com>
 <BD8B69E7-26D8-4706-9164-597C6AE57812@gmail.com>
In-Reply-To: <BD8B69E7-26D8-4706-9164-597C6AE57812@gmail.com>
From: Melanie Plageman <melanieplageman@gmail.com>
Date: Tue, 3 Mar 2026 10:52:34 -0500
Message-ID: 
 <CAAKRu_a7ByG2LipFADR7UYnLP4BWQKOV1sajqJC4=R37iO05+A@mail.gmail.com>
Subject: Re: eliminate xl_heap_visible to reduce WAL (and eventually set VM
 on-access)
To: Chao Li <li.evan.chao@gmail.com>
Cc: Andres Freund <andres@anarazel.de>, Andrey Borodin <x4mmm@yandex-team.ru>,
	Kirill Reshke <reshkekirill@gmail.com>, Xuneng Zhou <xunengzhou@gmail.com>,
	Robert Haas <robertmhaas@gmail.com>,
	PostgreSQL Hackers <pgsql-hackers@lists.postgresql.org>,
 Heikki Linnakangas <hlinnaka@iki.fi>
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable
Archived-At: 
 <https://www.postgresql.org/message-id/CAAKRu_a7ByG2LipFADR7UYnLP4BWQKOV1sajqJC4%3DR37iO05%2BA%40mail.gmail.com>
Precedence: bulk

On Tue, Mar 3, 2026 at 2:33=E2=80=AFAM Chao Li <li.evan.chao@gmail.com> wro=
te:
>
> 2 - 0003 - Does it make sense to also do the same renaming in PruneFreeze=
Result?

I could do that. Later commits remove them, so I thought it didn't
make sense. If only this commit goes in though, it would make sense.

> -                * Calculate what the snapshot conflict horizon should be=
 for a record
> -                * freezing tuples. We can use the visibility_cutoff_xid =
as our cutoff
> -                * for conflicts when the whole page is eligible to becom=
e all-frozen
> -                * in the VM once we're done with it. Otherwise, we gener=
ate a
> -                * conservative cutoff by stepping back from OldestXmin.
> -                */
> -               if (prstate->set_all_frozen)
> -                       prstate->frz_conflict_horizon =3D prstate->visibi=
lity_cutoff_xid;
> -               else
> -               {
> -                       /* Avoids false conflicts when hot_standby_feedba=
ck in use */
> -                       prstate->frz_conflict_horizon =3D prstate->cutoff=
s->OldestXmin;
> -                       TransactionIdRetreat(prstate->frz_conflict_horizo=
n);
> -               }
> +               Assert(TransactionIdPrecedesOrEquals(prstate->pagefrz.Fre=
ezePageConflictXid,
> +                                                                        =
                prstate->cutoffs->OldestXmin));
> ```
>
> At this point of Assert, can prstate->pagefrz.FreezePageConflictXid be In=
validTransactionId? My understanding is no, in that case, would it make sen=
se to also Assert(prstate->pagefrz.FreezePageConflictXid !=3D InvalidTransa=
ctionId)?

I think it is possible if we are doing some kind of freezing to a
multixact that we reach here and FreezePageConflictXid is
InvalidTransactionId.

> Otherwise, if prstate->pagefrz.FreezePageConflictXid is still possibly be=
 InvalidTransactionId, then the Assert should be changed to something like:
>
> Assert(prstate->pagefrz.FreezePageConflictXid =3D=3D InvalidTransactionId=
 ||
>   TransactionIdPrecedesOrEquals(prstate->pagefrz.FreezePageConflictXid, p=
rstate->cutoffs->OldestXmin)

This is covered by TransactionIdPrecedesOrEquals because
InvalidTransactionId is 0. We assume that in many places throughout
the code.

> I will continue with 0005 tomorrow.

Thanks for the review!

I noticed a serious bug in v35-0017: I pass hscan->modifies_base_rel
to heap_page_prune_opt() as rel_read_only, which is the opposite of
what I want to do -- it should be !hscan->modifies_base_rel. I'm going
to wait to fix it though and post a new v36 once I've batched up more
fixups.

- Melanie