From: Vladislav Shpilevoy <v.shpilevoy@tarantool.org>
To: tarantool-patches@freelists.org,
Kirill Shcherbatov <kshcherbatov@tarantool.org>
Subject: [tarantool-patches] Re: [PATCH v1 2/4] sql: use 64b bitmasks instead of 32b where possible
Date: Fri, 15 Feb 2019 20:05:05 +0300 [thread overview]
Message-ID: <cc945579-d4f0-1fc9-6888-9a411098c710@tarantool.org> (raw)
In-Reply-To: <26ee3cd45826f7de3290d14b37524d181ef320af.1549629707.git.kshcherbatov@tarantool.org>
Thanks for the patch! See 1 comment below.
On 08/02/2019 13:52, Kirill Shcherbatov wrote:
> In some cases(like foreign keys) the SQL code is still used
> 32-bit bit mask, while 64-bit bit masks will perform better
> column optimizations. There was refactored code to work with 64b
> bitmasks where required.
> The 32b bitmasks are still used to specify constant OP_Function
> arguments because this change would require changing the P1 type
> of the VDBE p1 argument, which is not desirable. Moreover, the
> 64 function's arguments is an explicit overkill.
>
> Part of #3571
> ---
> src/box/alter.cc | 10 ++-------
> src/box/sql/delete.c | 8 +++----
> src/box/sql/resolve.c | 19 ++++++-----------
> src/box/sql/sqliteInt.h | 46 +++++++++++++++++++++++------------------
> src/box/sql/trigger.c | 14 ++++++-------
> src/box/sql/update.c | 21 ++++++++-----------
> 6 files changed, 53 insertions(+), 65 deletions(-)
>
> diff --git a/src/box/sql/delete.c b/src/box/sql/delete.c
> index f9c42fdec..5e134ab0c 100644
> --- a/src/box/sql/delete.c
> +++ b/src/box/sql/delete.c
> @@ -484,10 +484,8 @@ sql_generate_row_delete(struct Parse *parse, struct Table *table,
> */
> sqlite3VdbeAddOp2(v, OP_Copy, reg_pk, first_old_reg);
> for (int i = 0; i < (int)table->def->field_count; i++) {
> - testcase(mask != 0xffffffff && iCol == 31);
> - testcase(mask != 0xffffffff && iCol == 32);
> - if (mask == 0xffffffff
> - || (i <= 31 && (mask & MASKBIT32(i)) != 0)) {
> + if (i >= COLUMN_MASK_SIZE ||
> + column_mask_fieldno_is_set(mask, i)) {
Now I see, why did you define COLUMN_MASK_SIZE. And still the problem
is in the way how you understand the last bit.
column_mask_fieldno_is_set() should check, if i >= COLUMN_MASK_SIZE, and in such
a case 'and' it with the last bit. In your code the mask is not 'smart' -
you consider fields >= 63 as set even if the last bit of the mask is 0. It is
worthy of note that literally every place of column_mask_fieldno_is_set() usage
also checks for i >= COLUMN_MASK_SIZE.
> sqlite3ExprCodeGetColumnOfTable(v, table->def,
> cursor, i,
> first_old_reg +
next prev parent reply other threads:[~2019-02-15 17:05 UTC|newest]
Thread overview: 17+ messages / expand[flat|nested] mbox.gz Atom feed top
2019-02-08 12:52 [tarantool-patches] [PATCH v1 0/4] sql: replace 32 bit column mask Kirill Shcherbatov
2019-02-08 12:52 ` [tarantool-patches] [PATCH v1 1/4] box: introduce new helpers in column_mask.h Kirill Shcherbatov
2019-02-15 17:05 ` [tarantool-patches] " Vladislav Shpilevoy
2019-02-20 13:42 ` Kirill Shcherbatov
2019-02-22 17:51 ` Vladislav Shpilevoy
2019-02-22 18:01 ` Konstantin Osipov
2019-02-22 18:22 ` Konstantin Osipov
2019-02-08 12:52 ` [tarantool-patches] [PATCH v1 2/4] sql: use 64b bitmasks instead of 32b where possible Kirill Shcherbatov
2019-02-15 17:05 ` Vladislav Shpilevoy [this message]
2019-02-20 13:42 ` [tarantool-patches] " Kirill Shcherbatov
2019-02-08 12:52 ` [tarantool-patches] [PATCH v1 3/4] sql: got rid of redundant MASKBIT32 definition Kirill Shcherbatov
2019-02-15 17:05 ` [tarantool-patches] " Vladislav Shpilevoy
2019-02-20 13:42 ` Kirill Shcherbatov
2019-02-08 12:52 ` [tarantool-patches] [PATCH v1 4/4] sql: got rid of redundant bitmask helpers Kirill Shcherbatov
2019-02-15 17:05 ` [tarantool-patches] " Vladislav Shpilevoy
2019-02-20 13:42 ` Kirill Shcherbatov
2019-02-22 17:52 ` Vladislav Shpilevoy
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=cc945579-d4f0-1fc9-6888-9a411098c710@tarantool.org \
--to=v.shpilevoy@tarantool.org \
--cc=kshcherbatov@tarantool.org \
--cc=tarantool-patches@freelists.org \
--subject='[tarantool-patches] Re: [PATCH v1 2/4] sql: use 64b bitmasks instead of 32b where possible' \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox