page must be big enough to store two records, this is done because compression
later may very well result in two records residing on the same page. This
change handles the case where only one record fits on a page. We don't
split the page in the middle by default if there is only record on the page.
We only split the page if the tuple to be inserted is less than existing
record. That way the existing record is copied over to the right page
during the split and the new tuple is inserted to the left.
symbols. Use it for all definitions of non-static variables and functions.
lexyy.c, make_flex.sh: Declare yylex as UNIV_INTERN, not static. It is
referenced from pars0grm.c.
Actually, according to
nm .libs/ha_innodb.so|grep -w '[ABCE-TVXYZ]'
the following symbols are still global:
* The vtable for class ha_innodb
* pars0grm.c: The function yyparse() and the variables yychar, yylval, yynerrs
The required changes to the Bison-generated file pars0grm.c will be addressed
in a separate commit, which will add a script similar to make_flex.sh.
The class ha_innodb is renamed from class ha_innobase by a #define. Thus,
there will be no clash with the builtin InnoDB. However, there will be some
overhead for invoking virtual methods of class ha_innodb. Ideas for making
the vtable hidden are welcome. -fvisibility=hidden is not available in GCC 3.
btr_page_get_sure_split_rec(): Remove the check if insert_size
exceeds free_space.
btr_page_split_and_insert(): If a compressed page has already been split,
avoid further splits by inserting the record to an empty page. As a
performance optimization, avoid invoking btr_page_insert_fits() on
compressed tables.
page_cur_set_before_first(), page_cur_set_after_last(),
page_cur_position(): Add const qualifiers to buf_block_t and rec.
A better solution would be to define a const_page_cur_t and a
set of accessors, but it would lead to severe code duplication.
page_rec_get_n_recs_before(): Add const qualifiers.
page_dir_get_nth_slot(): Define as a const-preserving macro.
page_dir_slot_get_rec(), page_dir_slot_get_n_owned(),
page_dir_find_owner_slot(), page_check_dir(): Add const qualifiers.
page_rec_get_next_low(): Add const qualifiers.
page_rec_get_next_const(), page_rec_get_prev_const(): New functions,
based on the const-less page_rec_get_next() and page_rec_get_prev().
page_cur_get_page(), page_cur_get_block(), page_cur_get_page_zip(),
page_cur_get_rec(): Define as const-preserving macros.
page_cur_try_search_shortcut(), page_cur_search_with_match():
Add const qualifiers.
buf_page_get_mutex(): Add a const qualifier to buf_page_t*.
rec_get_next_ptr_const(): Const variant of rec_get_next_ptr().
For some reason, GCC 4.2.1 ignores casts (for removing constness)
in calls to inline functions.
page_align(), ut_align_down(): Make the parameter const void*, but still
return a non-const pointer. This is ugly, but these functions cannot be
replaced with a const-preserving macro in a portable way, given that
the pointer argument is not always pointing to bytes.
buf_block_get_page_zip(): Implement as a const-preserving macro.
buf_frame_get_page_zip(), buf_block_align(): Add const qualifiers.
lock_rec_get_prev(): Silence GCC 4.2.1 warnings.
mlog_write_initial_log_record(), mlog_write_initial_log_record_fast(),
mtr_memo_contains(): Add const qualifier to the pointer.
page_header_get_ptr(): Rewrite as page_header_get_offs(), and
implement as a macro that calls this function.
offsets_[] arrays, as suggested by Vasil.
rec_offs_set_n_alloc(): Declare as a public function. Assert that
n_alloc > REC_OFFS_HEADER_SIZE.
rec_offs_get_n_alloc(): Assert that n_alloc > REC_OFFS_HEADER_SIZE.
Some things still fail in innodb-index.test, and there seems to be
a race condition (data dictionary lock wait) when running with --valgrind.
dfield_t: Add an "external storage" flag, dfield->ext.
dfield_is_null(), dfield_is_ext(), dfield_set_ext(), dfield_set_null():
New functions.
dfield_copy(), dfield_copy_data(): Add const qualifiers, fix in/out comments.
data_write_sql_null(): Use memset().
big_rec_field_t: Replace byte* data with const void* data.
ut_ulint_sort(): Remove.
upd_field_t: Remove extern_storage.
upd_node_t: Replace ext_vec, n_ext_vec with n_ext.
row_merge_copy_blobs(): New function.
row_ins_index_entry(): Add the parameter "ibool foreign" for suppressing
foreign key checks during fast index creation or when inserting into
secondary indexes.
btr_page_insert_fits(): Add const qualifiers.
btr_cur_add_ext(), upd_ext_vec_contains(): Remove.
dfield_print_also_hex(), dfield_print(): Replace if...else if with switch.
Observe dfield_is_ext().
the insert buffer bitmap.
ibuf_set_free_bits_func(): Never disable redo logging.
ibuf_update_free_bits_zip(): Remove.
btr_page_reorganize_low(), page_zip_reorganize(): Do not update the insert
buffer bitmap. Instead, document that callers will have to take care of it,
and adapt the callers.
btr_compress(): On error, reset the insert buffer free bits.
btr_cur_insert_if_possible(): Do not modify the insert buffer bitmap.
btr_compress(), btr_cur_optimistic_insert(): On compressed pages,
reset the insert buffer bitmap. Document why.
btr_cur_update_alloc_zip(): Document why it is necessary and sufficient
to reset the insert buffer free bits.
btr_cur_update_in_place(), btr_cur_optimistic_update(),
btr_cur_pessimistic_update(): Update the free bits in the same
mini-transaction. Document that the mini-transaction must be
committed before latching any further pages. Verify that this
is the case in all execution paths.
row_ins_sec_index_entry_by_modify(), row_ins_clust_index_entry_by_modify(),
row_undo_mod_clust_low(): Because these functions call
btr_cur_update_in_place(), btr_cur_optimistic_update(), or
btr_cur_pessimistic_update(), document that the mini-transaction must be
committed before latching any further pages. Verify that this is the case
in all execution paths.
crash recovery is in progress. This avoids a hang when
btr_parse_page_reorganize(), called from an I/O handler thread,
attempts to acquire log_sys->mutex while it is being held by
the main thread (the one that runs innobase_init()). This change
was committed accidentally. It may be unsafe to clear
mtr.modifications, because buf_page_release() at mtr_commit() may
forget to put modified pages to the flush list.
Cleanup: Remove the "type" parameter from many ibuf functions.
Let the caller check that !dict_index_is_clust(). This should avoid
function calls and register spilling.
ibuf_set_free_bits_func(), ibuf_set_free_bits(): Remove the parameter "type".
ibuf_reset_free_bits_with_type(): Rename to ibuf_reset_free_bits().
Remove the parameter "type".
ibuf_update_free_bits_if_full(), ibuf_update_free_bits_zip(),
ibuf_update_free_bits_low(), ibuf_update_free_bits_for_two_pages_low():
Remove the parameter "index".
failing insert. Reorganization will have been attempted in
page_cur_tuple_insert() or page_cur_rec_insert().
page_zip_reorganize(): Recompute the insert buffer free bits for
leaf pages of secondary indexes.
ibuf_data_enough_free_for_insert(): Simplify.
compressed tablespaces.
btr_page_reorganize_low(): Recompute the insert buffer free bits
for compressed pages belonging to a non-clustered index.
btr_cur_optimistic_insert(): Determine if the page was reorganized by
page_cur_tuple_insert(). Do not attempt to reorganize compressed pages,
because that will be attempted by page_cur_tuple_insert().
page_zip_max_ins_size(): New function.
btr_cur_optimistic_insert(), btr_cur_optimistic_delete(),
btr_page_split_and_insert(), btr_compress(): Do not update the
ibuf free bits for non-leaf pages or pages belonging to a clustered index.
The insert buffer only covers operations on leaf pages of secondary indexes.
For pages covered by the insert buffer, limit the max_ins_size to
page_zip_max_ins_size().
buf_page_get_gen(): Merge the insert buffer after decompressing the page.
buf_page_io_complete(): Relax the assertion about ibuf_count. For
compressed-only pages, the insert buffer merge takes place
in buf_page_get_gen().
ibuf_index_page_calc_free_bits(), ibuf_index_page_calc_free_from_bits(),
ibuf_index_page_calc_free(), ibuf_update_free_bits_if_full(),
ibuf_update_free_bits_low(), ibuf_update_free_bits_for_two_pages_low(),
ibuf_set_free_bits_low(): Add the parameter zip_size. Limit the maximum
insert size to page_zip_max_ins_size().
buf_page_get_gen(). This saves one mutex operation per block request.
buf_page_get_gen(), various macros and functions: Add parameter zip_size.
btr_node_ptr_get_child(): Add parameter index.
fil_space_get_latch(): Add optional output parameter zip_size.
fil_space_get_zip_size(): Return 0 for space id==0, because the
system tablespace is never compressed.
fsp_header_init(): Remove the parameter zip_size.
ibuf_free_excess_pages(): Remove the parameter zip_size.
trx_rseg_t, trx_undo_t: Add field zip_size.
xdes_lst_get_next(): Remove, unused.
storage size from 16 to 3 bits.
page_zip_get_size(), page_zip_set_size(): New functions.
Replace direct references to page_zip_des_t:size with calls to
buf_block_get_zip_size(), page_zip_get_size(), and page_zip_set_size().
and block->space with buf_block_get_state(block), buf_block_get_page_no(block),
and buf_block_get_space(block).
enum buf_block_state: Replaces the #define'd buf_block_t.state values.
buf_block_get_state(): New function.
buf_block_get_frame(): Add __attribute__((const)).
rec_get_nth_field_offs_old(): Split from rec_get_nth_field_old().
rec_get_nth_field_old(): Reimplement as a constness-preserving macro.
dict_drop_index_tree(), dict_truncate_index_tree(): Note that rec is
an in/out parameter, as the B-tree root page number will be updated.
btr_check_node_ptr(): Replace page_t* parameter with buf_block_t*.
btr_free_externally_stored_field(): Add const qualifier to rec.
Remove an explicit buf_block_align() call, but replace an
mtr_memo_contains() with mtr_memo_contains_page().
row_upd_rec_sys_fields(): Reorder an assertion containing buf_block_align()
so that the costly call can be avoided in some cases.
Replace all page_t* parameters with buf_block_t*, and replace many
rec_t* parameters with heap_no. This eliminates also many
rec_get_heap_no() calls, which became more expensive with the
introduction of ROW_FORMAT=COMPACT in MySQL/InnoDB 5.0.3.
page_rec_get_heap_no(), lock_get_min_heap_no(): New functions.
page_set_max_trx_id(): Replace page_t* parameter with buf_block_t*,
to avoid a buf_block_align() call.
Replace some occurrences of page_get_page_no() with buf_block_get_page_no().
page_cur_delete_rec(): Replace buf_block_align() with page_cur_get_block().
btr_cur_t: Move page_block to page_cur_t::block.
page_cur_get_block(), page_cur_get_page_zip(): New functions.
page_cur_position(): Add parameter block.
Remove many page_zip parameters, now that there is page_cur_get_page_zip().
Replace some page, page_zip parameters with block.
Add some const qualifiers to function parameters and remove casts.
PAGE_HEAP_NO_INFIMUM, PAGE_HEAP_NO_SUPREMUM, PAGE_HEAP_NO_USER_LOW:
New constants.
Replace some cursor code in low-level diagnostic functions with
direct management of rec, because buf_block_t::buf_fix_count may be 0
when the functions are called, and debug assertions would fail.
accessors returning pointers with macros that preserve const qualifiers.
In UNIV_DEBUG builds, retain the accessors and cast away constness there.
dfield_get_type(), dfield_get_data(), dtuple_get_nth_field(),
dict_table_get_nth_col(), dict_table_get_sys_col(): Implement as macro
unless #ifdef UNIV_DEBUG.
rec_get_nth_field(): Replace with rec_get_nth_field_offs() that does not
do pointer arithmetics. Implement rec_get_nth_field() as a macro.
ibuf_reset_free_bits(): Remove, as there already is a similar function
ibuf_reset_free_bits_with_type().
ibuf_reset_free_bits_with_type(), ibuf_set_free_bits(),
ibuf_update_free_bits_if_full(), btr_leaf_page_release(),
buf_page_make_young(): Replace page_t with buf_block_t.
btr_compress(): Replace btr_page_get() with btr_block_get().
with page_get_page_no() and page_get_space_id(). We want to avoid
buf_block_align() calls, and the page_no and space_id are now stamped
on the pages early on.
Replace ut_ad(mtr_memo_contains(mtr, buf_block_align(ptr), ...))
with ut_ad(mtr_memo_contains_page(mtr, ptr, ...)) in order to reduce the
number of buf_block_align() calls.
of externally stored columns, and fix bugs introduced in r873. (Bug #22496)
btr_page_get_sure_split_rec(), btr_page_insert_fits(),
rec_get_converted_size(), rec_convert_dtuple_to_rec(),
rec_convert_dtuple_to_rec_old(), rec_convert_dtuple_to_rec_new():
Add parameters ext and n_ext. Flag external fields during the
conversion.
rec_set_field_extern_bits(), rec_set_field_extern_bits_new(),
rec_offs_set_nth_extern(), rec_set_nth_field_extern_bit_old():
Remove. The bits are set by rec_convert_dtuple_to_rec().
page_cur_insert_rec_low(): Remove the parameters ext and n_ext.
btr_cur_add_ext(): New utility function for updating and sorting ext[].
Low-level functions now expect the array to be in ascending order
for performance reasons. Used in btr_cur_optimistic_insert(),
btr_cur_pessimistic_insert(), and btr_cur_pessimistic_update().
btr_cur_optimistic_insert(): Remove some defensive code, because we cannot
compute the added parameters of rec_get_converted_size().
btr_push_update_extern_fields(): Sort the array. Require the array to
be twice the maximum usage, so that ut_ulint_sort() can be used.
dtuple_convert_big_rec(): Allocate new space for the BLOB pointer,
to avoid overwriting prefix indexes to the same column. Adapt
dtuple_convert_back_big_rec().
row_build_index_entry(): Fetch the columns also for prefix indexes of
the clustered index.
page_zip_apply_log(), page_zip_decompress_clust(): Allow externally
stored fields to lack a locally stored part.
"On compressed pages, return NULL if incl_data > free_space."
btr_page_split_and_insert(): Remove the limit check of n_iterations
on compressed pages. After enough many invocations of
btr_page_get_sure_split_rec(), the record will finally fit.
for (;;) { ... if (condition) { ... return }} loop as
do { ... } while (!condition); ... return.
On compressed pages, return NULL if incl_data > free_space.
guaranteed free space available for inserting one record.
btr_page_get_sure_split_rec(), btr_cur_pessimistic_insert():
Use page_zip_empty_size().
btr_page_split_and_insert(): Relax a debug assertion that there should
be at least two user records on the page. On compressed pages, we may
be able to write only one record.
