| Age | Commit message (Collapse) | Author |
|---|
|
This allows bif1/2/3 to share the main part of the code.
The price is that we always need to copy all three temporary registers
when error handling in bodies, but that should be infrequent.
Additionally it makes it a bit harder to read the disasembly since now
the arguments to BIFs are in the reverse order.
|
|
Simplify implementation of GC BIFs
|
|
Summary: This commit simplifies the implementation of the "GC BIFs" so
that they no longer need to do a garbage collection, removing duplicate
code for all GC BIFs in the runtime system, as well as potentially
reducing the size of the loaded BEAM code by using shorter
instructions calling those BIFs.
A GC BIF is a guard BIF that will do a garbage
collection if it needs to build anything on the heap.
For example, `abs/1` is a GC BIF because it might need to
allocate space on the heap (if the result is a floating point
number or the resulting integer is a bignum).
Before R12, a guard BIF (such as `abs/1`) that need to allocate
heap space would allocate outside of process's main heap, in
a heap fragment.
GC BIFs were introduced in R12B to support literals. During garbage
collection it become necessary to quickly test whether a term was
a literal. To make the check simple, guards BIFs were no longer
allowed to create heap fragments. Instead GC BIFs were introduced.
In OTP 19, the implementation of literals was changed to support
storing messages in heap fragments outside of the main heap for a
process. That change again made it allowed for guard BIFs to create
heap fragments when they need to build terms on the heap.
It would even be possible for the guard BIFs to build directly
on the main heap if there is room there, because the compiler
assumes that a new `test_heap/2` instruction must be emitted
when building anything after calling a GC BIF. (We don't do that
in this commit; see below.)
This commit simplifies the implementation of the GC BIFs in
the runtime system.
Each GC BIF had a dual implementation: one that was used when the GC
BIF was called directly and one used when it was called via
`apply/3`. For example, `abs/1` was implemented in `abs_1()` and
`erts_gc_abs_1()`. This commit removes the GC version of each BIF. The
other version that allocates heap space using `HAlloc()` is updated to
use the new `HeapFragOnlyAlloc()` macro that will allocate heap
space in a heap fragment outside of the main heap.
Because the BIFs will allocate outside of the main heap, the same
`bif` instructions used by nonbuilding BIFs can be used for the
(former) GC BIFs. Those instructions don't use the macros that save
and restore the heap and stack pointers (SWAPOUT/SWAPIN). If the
former GC BIFs would build on the main heap, either new instructions
would be needed, or SWAPOUT/SWAPIN instructions would need to be added
to the `bif` instructions.
Instructions that call the former GC BIFs don't need the operand
that specifies the number of live X registers. Therefore, the
instructions that call the BIFs are usually one word shorter.
|
|
* bjorn/erts/remove-faulty-assertion:
erl_bif_list: Remove faulty assertion
|
|
* maint:
Fix reading beyond end of bignum in integer squaring
|
|
Fix reading beyond end of bignum in integer squaring
|
|
The multiplication of two bignums is specially optimized when the two
operands have the same address, because squaring can be done more
efficiently than multiplication of two arbitrary integers. That is,
expressions such as `I * I` will be calculated by squaring the value
of `I`.
The optimized function for squaring would read one word beyond the
end of the bignum in the last iteration of a loop. The garbage
value would never be used. In almost all circumstances that would
be harmless. Only if the read word happened to fall on the start
of an unmapped page would the runtime crash. That is unlikely
to happen because most bignums are stored on a process heap, and
since the stack is located at the other end of the block that the
heap is located in, the word beyond the end of bignum is guaranteed
to be readable.
|
|
* sverker/erts/ets-debug-test-cuddle:
erts: Fix volatile ets test case failures on debug VM
|
|
* maint:
Updated OTP version
Prepare release
|
|
The variable can be an empty list.
|
|
|
|
|
|
Fix --without-APP config option
|
|
* maint:
Updated OTP version
Prepare release
|
|
* maint-20:
Updated OTP version
Prepare release
|
|
* maint:
Don't try to wake up wx thread when not necessary
|
|
* dgud/wx/perf-imp:
Don't try to wake up wx thread when not necessary
|
|
|
|
* henrik/ERIERL-278/OTP-15470:
Handle erroneous length during decode (BER only) without crashing
|
|
jhogberg/john/erts/mark-pooled-free-blocks-unused/OTP-15075
Mark free blocks in pooled carriers as unused
|
|
This lets the OS reclaim the physical memory associated with these
blocks which reduces the impact of long-lived awkward allocations.
A small allocated block will still keep a huge carrier alive, but
the unused part of the carrier will now be available to the OS.
Co-authored-by: Dmytro Lytovchenko <[email protected]>
|
|
That kills performance on at least windows when a the load goes up and
a lot of commands is sent to the driver.
|
|
|
|
|
|
* henrik/ERIERL-278/OTP-15470:
Handle erroneous length during decode (BER only) without crashing
|
|
|
|
Conflicts:
erts/emulator/beam/erl_process.c
|
|
* lukas/OTP-21.1.1/scheduler_pollset/OTP-15475:
erts: Move fds with active true behaviour to own pollset
erts: Fix lists_member_2 reduction count
erts: Allow code_model_small to be set in xcomp setting
erts: Implement delay_send using timer instead of poll
erts: Optimize driver_set_timer(0) to fire at once
erts: Optimize the inet driver multi timers for one timer
erts: Move all inet tcp CONNECTED timers to multi timer
erts: Add erts_io_notify_port_task_executed to check_io msacc state
erts: Add pre-alloc to ALLOC msacc state
erts: Make thr prgr wakeup current or sched 1
erts: Pass thread progress data where possible
|
|
At start of the VM a poll-set that the schedulers
will check is created where fds that have triggered
many (at the moment, many means 10) times without
being deselected inbetween. In this scheduler specific
poll-set fds do not use ONESHOT, which means that the
number of syscalls goes down dramatically for such fds.
This pollset is introduced in order to handle fds that
are used by the erlang distribution and that never
change their state from {active, true}.
This pollset only handles ready_input events,
ready_output is still handled by the poll threads.
During overload, polling the scheduler poll-set is done
on a 10ms timer.
|
|
|
|
* 'kenneth/asn1/ERIERL-278/OTP-15470' of git-server:kenneth_otp:
Handle erroneous length during decode (BER only) without crashing
|
|
* raimo/test-cuddling:
Rewrite to work for stream connection
Update Linux version for pktoptions TTL
|
|
On a 32-bit machine, `beam_makeops` would generate incorrect
pack instructions for instructions such as:
i_plus x x j? d
See the added comment block for a detailed description of the
problem and its fix.
|
|
* maint:
Fix unsafe optimization of stack trace building
|
|
beam_ssa_pre_codegen: Fix an internal consistency failure
|
|
Fix unsafe optimization of stack trace building
|
|
Symptom:
Test cases with small key ranges sometimes failed on debug VM with:
"No routing nodes in table?
Debug feature 'ets_force_split' does not seem to work."
Solution:
Don't provoke randomly joins when force_split is set.
|
|
|
|
|
|
* bjorn/compiler/fragility:
beam_validator: Don't discard fragility
|
|
The `sys_core_fold` pass of the compiler would optimize
away the building of the stacktrace in code such as:
try
...
catch
C:R:Stk ->
erlang:raise(C, {R,Stk}, Stk)
end
That optimization is unsafe and would cause a crash in a later compiler
pass.
|
|
The parallel configure implemented in d4b4567 would rm -f the
SKIP-APPLICATIONS file before it was consumed at the end of the
configure script. This reverses the solution by instead first building
the list of skip applications in memory and only then writing them to
the SKIP-APPLICATIONS file.
|
|
The following function:
is_two_tuple(Arg) ->
case is_tuple(Arg) of
false -> false;
true -> tuple_size(Arg) == 2
end.
would cause an internal consistency failure:
Internal consistency check failed - please report this bug.
Instruction: {bif,tuple_size,{f,0},[{x,0}],{z,0}}
Error: {invalid_store,{z,0},{integer,[]}}:
|
|
|
|
garazdawi/lukas/erts/fix_lit_msg_copy_in_gc/ERL-741/OTP-15478
erts: Fix copy of literal msg during gc
|
|
Conflicts:
lib/ssl/test/ssl_dist_bench_SUITE.erl
|
|
* raimo/ssl/tls_dist-optimization:
Handle socket close in state downgrade
Handle dead sender at terminate
Handle tls_sender exit properly
Optimize split_bin
Improve dist send throughput
|
|
* raimo/ssl/tls_dist-profiling:
Improve benchmark
Implement print on other node
Fix compiler warnings
|
|
|
|
* ingela/ssl/test-cuddle-ssl_basic_SUITE:
ssl: Correct test case
|
