pg.ddx.io  pgsql-hackers@postgresql.org mailing list archive  
help / color / mirror / Atom feed
From: Anton A. Melnikov <a.melnikov@postgrespro.ru>
To: Tom Lane <tgl@sss.pgh.pa.us>
Cc: John Naylor <johncnaylorls@gmail.com>
Cc: Jeff Davis <pgsql@j-davis.com>
Cc: Ants Aasma <ants.aasma@cybertec.at>
Cc: Heikki Linnakangas <hlinnaka@iki.fi>
Cc: Junwang Zhao <zhjwpku@gmail.com>
Cc: jian he <jian.universality@gmail.com>
Cc: Andres Freund <andres@anarazel.de>
Cc: Gurjeet Singh <gurjeet@singh.im>
Cc: pgsql-hackers@postgresql.org
Subject: Re: Change GUC hashtable to use simplehash?
Date: Thu, 16 Jan 2025 13:54:53 +0300
Message-ID: <16198d33-a4e2-46db-b6c3-494367f5816d@postgrespro.ru> (raw)
In-Reply-To: <1520769.1736991361@sss.pgh.pa.us>
References: <c95e1821b6a69ec0c40eb0252e037b68a62cdae3.camel@j-davis.com>
	<48e8f8bbe0be9c789f98776c7438244ab7a7cc63.camel@j-davis.com>
	<CANwKhkP7pCiW_5fAswLhs71-JKGEz1c1+PC0a_w1fwY4iGMqUA@mail.gmail.com>
	<CANWCAZbXooNxa1oX4+N-QVTknRiiUBs17ai8rcWYihs_g=ShEQ@mail.gmail.com>
	<CANWCAZbg_XeSeY0a_PqWmWqeRATvzTzUNYRLeT+bzs+YQdC92g@mail.gmail.com>
	<CANWCAZbwvp7oUEkbw-xP4L0_S_WNKq-J-ucP4RCNDPJnrakUPw@mail.gmail.com>
	<458350a908d9607f7fc2a15407f4405cc454c970.camel@j-davis.com>
	<CANWCAZYxrJJ4EDw=o5vu_gAs7dS_PK6ry0rqkVu9opGpAJjoQQ@mail.gmail.com>
	<2b0df3af97fac73a40aba16a58837e2cee594b1f.camel@j-davis.com>
	<CANWCAZYLvWOm-1XLT8wGoXGgNBEV_NPMWJY4gM6cXa7LHVDjtw@mail.gmail.com>
	<a3a959f6-14b8-4819-ac04-eaf2aa2e868d@postgrespro.ru>
	<CANWCAZZ0ptKxEyO504OjZOr3Co=WaSEQG5eUU-TFmqD4twhe_w@mail.gmail.com>
	<CANWCAZZjUtfq148s=55L1LKHu+imeKn=iFH5UvgtxRZcCK9fbg@mail.gmail.co! m>
	<0647027b-9c9a-4f16-8f7c-3f9f3eb9451e@postgrespro.ru>
	<1520769.1736991361@sss.pgh.pa.us>

Hi!

On 16.01.2025 04:36, Tom Lane wrote:
> "Anton A. Melnikov" <a.melnikov@postgrespro.ru> writes:
>> Seems it is possible to exclude much less code from checking
>> under valgrind and get the same result by replacing the only
>> function call pg_rightmost_one_pos64() with a valgrind-safe
>> code. See the attached patch, please.
> 
> There is no place anywhere in our code base where we hide unsafe
> code from valgrind rather than fixing said code.  This does not
> seem like a place to start such an ugly practice.  Performance
> does not trump everything else.

Thanks for remark. Agreed.

> I'd be inclined to just remove the pg_rightmost_one_pos64 call
> in favor of the other coding you suggest.

Here is a patch like that.


With the best wishes,

-- 
Anton A. Melnikov
Postgres Professional: http://www.postgrespro.com
The Russian Postgres Company

Attachments:

v2-0001-Add-valgrind-safe-code-to-find-rightmost-bytes.patch (text/x-patch, 1.4K) download From bac26e0c7ee900246d454dd98de45b3ecefc15ec Mon Sep 17 00:00:00 2001 From: "Anton A. Melnikov" <a.melnikov@postgrespro.ru> Date: Thu, 16 Jan 2025 02:56:52 +0300 Subject: [PATCH] Use valgrind-safe code to find a number of rightmost zero bytes. --- src/include/common/hashfn_unstable.h | 16 +++++++++++----- 1 file changed, 11 insertions(+), 5 deletions(-) diff --git a/src/include/common/hashfn_unstable.h b/src/include/common/hashfn_unstable.h index a09c0b7d5ce..fa9b954c6b9 100644 --- a/src/include/common/hashfn_unstable.h +++ b/src/include/common/hashfn_unstable.h @@ -299,15 +299,21 @@ fasthash_accum_cstring_aligned(fasthash_state *hs, const char *str) if (firstbyte64(chunk) != 0) { - size_t remainder; + size_t remainder = 0; uint64 mask; /* - * The byte corresponding to the NUL will be 0x80, so the rightmost - * bit position will be in the range 15, 23, ..., 63. Turn this into - * byte position by dividing by 8. + * The byte corresponding to the NUL terminator will be the rightmost 0x80. + * All zero bytes to the right of it correspond to the tail of the string. + * It remains to count them. */ - remainder = pg_rightmost_one_pos64(zero_byte_low) / BITS_PER_BYTE; + while ((zero_byte_low & 0xFF) == 0) + { + zero_byte_low >>= 8; + ++remainder; + } + + Assert(remainder != 0); /* * Create a mask for the remaining bytes so we can combine them into -- 2.48.1

view thread (103+ messages)  latest in thread

Message-ID: <16198d33-a4e2-46db-b6c3-494367f5816d@postgrespro.ru>
Permalink:  ../16198d33-a4e2-46db-b6c3-494367f5816d@postgrespro.ru/
Also on:    postgresql.org/message-id/16198d33-a4e2-46db-b6c3-494367f5816d@postgrespro.ru

 ·  · 

reply

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Reply to all the recipients from your mail client; the link fills in
  the headers below: reply via email
  Some clients, webmail especially, drop In-Reply-To and References from
  a link, and the reply then starts a new thread: use the mbox below.

  To: pgsql-hackers@postgresql.org
  Cc: a.melnikov@postgrespro.ru, tgl@sss.pgh.pa.us, johncnaylorls@gmail.com, pgsql@j-davis.com, ants.aasma@cybertec.at, hlinnaka@iki.fi, zhjwpku@gmail.com, jian.universality@gmail.com, andres@anarazel.de, gurjeet@singh.im
  Subject: Re: Change GUC hashtable to use simplehash?
  In-Reply-To: <16198d33-a4e2-46db-b6c3-494367f5816d@postgrespro.ru>
  References: <c95e1821b6a69ec0c40eb0252e037b68a62cdae3.camel@j-davis.com> <48e8f8bbe0be9c789f98776c7438244ab7a7cc63.camel@j-davis.com> <CANwKhkP7pCiW_5fAswLhs71-JKGEz1c1+PC0a_w1fwY4iGMqUA@mail.gmail.com> <CANWCAZbXooNxa1oX4+N-QVTknRiiUBs17ai8rcWYihs_g=ShEQ@mail.gmail.com> <CANWCAZbg_XeSeY0a_PqWmWqeRATvzTzUNYRLeT+bzs+YQdC92g@mail.gmail.com> <CANWCAZbwvp7oUEkbw-xP4L0_S_WNKq-J-ucP4RCNDPJnrakUPw@mail.gmail.com> <458350a908d9607f7fc2a15407f4405cc454c970.camel@j-davis.com> <CANWCAZYxrJJ4EDw=o5vu_gAs7dS_PK6ry0rqkVu9opGpAJjoQQ@mail.gmail.com> <2b0df3af97fac73a40aba16a58837e2cee594b1f.camel@j-davis.com> <CANWCAZYLvWOm-1XLT8wGoXGgNBEV_NPMWJY4gM6cXa7LHVDjtw@mail.gmail.com> <a3a959f6-14b8-4819-ac04-eaf2aa2e868d@postgrespro.ru> <CANWCAZZ0ptKxEyO504OjZOr3Co=WaSEQG5eUU-TFmqD4twhe_w@mail.gmail.com> <0647027b-9c9a-4f16-8f7c-3f9f3eb9451e@postgrespro.ru> <1520769.1736991361@sss.pgh.pa.us> <16198d33-a4e2-46db-b6c3-494367f5816d@postgrespro.ru>

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

This inbox is served by DDX for PostgreSQL; see mirroring instructions
for how to clone and mirror all data and code used for this inbox