Commit Graph

153 Commits

Author SHA1 Message Date
whitequark a1ea208652 arch/x86_64: hack to work around ld64 shortcomings on OS X.
fix #35, close #36
2016-08-31 20:25:00 -04:00
whitequark a733290359 arch: bring back clobbers for input operands.
This reverts a part of 365e40b. While it is not legal to clobber
the output register, it is legal and necessary to do this for inputs,
since we do not preserve them, and LLVM could have decided to
reuse them.

Thanks to @Amanieu for discovering this.
2016-08-31 14:17:54 -04:00
whitequark 248d7b16b5 arch/or1k: fix typo. 2016-08-31 14:17:54 -04:00
edef 1aa5ffdcd0 fix doc comment of Stack::limit 2016-08-30 18:30:58 -04:00
edef 1994ffee27 don't drop yielded values in the generator
fix #31
2016-08-30 18:17:52 -04:00
edef e47cb809f4 use valgrind_request from crates.io 2016-08-30 13:43:51 -04:00
whitequark 2d8cefcabc Do not simultaneously clobber registers passed into assembly.
This is not actually legal (although the restriction is not
documented anywhere), and is not caught by LLVM unless the codegen
option -verify-machineinstrs is specified. This option is now used
on Travis.

While not legal, this does not seem to result in invalid output
(although it creates an unnecessary spill); however, under extremely
specific circumstances (e.g. when a register scavenger is run under
severe pressure), this will result in a codegen crash, which is
how I found it.
2016-08-30 12:20:34 -04:00
whitequark e02ef03304 Add #[derive(Debug)] for SliceStack and OwnedStack. 2016-08-30 12:20:34 -04:00
whitequark 5a77a01863 Implement OwnedStack. 2016-08-30 12:20:34 -04:00
whitequark 4b32c18f70 Rename Yielder::generate to Yielder::suspend.
This makes more sense, because "generate" implies *running*
a generator, not *returning* from a generator. That name was
a vestige of the time when libfringe only really provided a .next()
implementation, which made marginally more sense together with
.generate(), but it doesn't anymore when we have .resume().
2016-08-30 12:20:34 -04:00
edef f21b8dc936 Update the license headers for MIT/Apache 2016-08-20 23:45:01 +02:00
whitequark f8cf95f686 Implement SliceStack. 2016-08-19 15:12:42 +02:00
whitequark e0ad79ea0c Add an OR1K port. 2016-08-19 15:12:42 +02:00
whitequark d4b7c427c9 impl Iterator for Generator<Input=()>. 2016-08-13 13:38:04 +02:00
whitequark 145c18125a Explain how trampolines work. 2016-08-13 13:38:04 +02:00
whitequark b79e7c0a9b Rename Stack::top() to Stack::base(), which is correct.
Fixes #25.
2016-08-13 13:38:04 +02:00
whitequark 7d5075edc2 Add more examples to README.
This also renames some functions for a mild increase in backtrace
clarity.
2016-08-13 13:38:04 +02:00
whitequark f34ddc6805 Don't use core::intrinsics.
The core::intrinsics::unreachable() we used at the end of every naked
function is essentially pointless, as #[naked] implies that intrinsic
at the end of the function. rustc currently does not implement that
behavior (rust-lang/rust#32487), but it is a bug. On top of that,
anything except a single asm!() in naked functions is likely to be
disallowed in the future (rust-lang/rust#32490).

A nice side effect is that we avoid the core_intrinsics feature,
which will be never stabilized, though neither asm nor
naked_functions are likely to be stabilized soon.
2016-08-13 13:38:04 +02:00
whitequark ddbf9a5afe Reimplement Generator to pass both input and output values around. 2016-08-13 13:38:04 +02:00
whitequark 308df32ac2 Use a more semantically correct PhantomData field in Generator. 2016-08-13 13:38:04 +02:00
whitequark 302ceef10a Clarify the Windows situation.
Also, remove the Windows OsStack implementation. It will not ever
be useful in libfringe, as even if we add support for Windows
using fibers, the OS allocates stacks for fibers and we needn't
do it manually.
2016-08-13 13:38:04 +02:00
whitequark a5d3430e63 Make Generator safe in presence of destructors. 2016-08-13 13:38:04 +02:00
whitequark 23cf17865d Remove workaround for rust-lang/rust#25544. 2016-08-13 13:38:04 +02:00
edef ca43ba0498 Make fringe::Context private. 2016-08-13 13:38:04 +02:00
edef b1a6b17d0a Move the Context tests and benches into src/
This concludes preparations for making Context private again.
2016-08-13 13:38:04 +02:00
whitequark 892a7696ec Allow unwinding to propagate across a context swap.
The main purpose of this is having nice backtraces in gdb, although
it also slightly simplifies poisoning state of the API consumers
after a panic.
2016-08-13 13:38:04 +02:00
whitequark 40fbfdde0c Kill impl StackPointer.
That was a silly idea for a silly abstraction, and it brought no
improvement, only confusion. So, revert it.
2016-08-13 13:38:04 +02:00
edef bb699c0f6e Allow inlining Generator::state 2016-08-13 12:52:12 +02:00
edef 8c761d944f Fix Windows x64 ABI 2016-07-17 16:57:23 -04:00
whitequark 15d2fbaf51 Add an inline hint on Generator::next().
This improves the generate benchmark by about 1ns.
2016-07-17 13:56:03 -04:00
whitequark 16c674b4f4 Implement the GuardedStack marker trait. 2016-07-17 13:56:03 -04:00
whitequark 7ffad26cfd Implement the safe Generator abstraction.
close #24
2016-07-17 13:55:56 -04:00
whitequark 758088c673 Fix Stack::new(0). 2016-07-17 13:55:43 -04:00
whitequark aa364056fd Add more documentation, and rephrase it elsewhere. 2016-07-16 20:54:47 -04:00
whitequark ebd9ca8dec Allocate guard page under existing stack, not in it.
This fixes a segfault when the allocated stack is just one page long.

This also refactors the fringe::os module to use Result consistently.

close #22
2016-07-16 20:54:33 -04:00
John Ericson 75ca6f66e1 Move x86 trampoline to naked function
Close #19
2016-07-16 16:39:19 -04:00
John Ericson 1ea4b01eda Use mostly the same registers for consistency
The exception to this is %eax is used not %edi on x86 because its
caller-save
2016-07-16 16:39:14 -04:00
whitequark cbe136b762 Completely rework fringe::Context and fringe::arch.
The new design concerns itself with one thing and exactly one thing:
passing values back and forth with an extern "C" function.
This allows to simplify fringe::arch into a single primitive, swap.

Close #21
2016-07-16 15:04:15 -04:00
edef 0d7b3d7091 add Windows support to OsStack
Fix #18
2016-04-16 07:51:10 +02:00
edef 6284a22caa add fpsr, eflags for good measure
courtesy of @whitequark
2016-04-16 07:41:00 +02:00
John Ericson 35fb046fb5 Factor out common code between x86 and x86_64
Fix #17
2016-04-14 13:37:15 +02:00
John Ericson 368fa9d3bf Remove unnecessary unsafety 2016-04-12 14:40:12 +02:00
edef 32fdbbb474 move OutStack's bound to a where clause 2016-04-03 20:07:45 +02:00
edef daece1a8d3 Registers::{swap2 => swap}, to match Context::swap 2016-04-03 19:56:15 +02:00
edef 976b971436 get rid of the old Context::swap
The two-parameter version is the only necessary API.
2016-04-03 19:55:25 +02:00
edef a50e6771b4 coerce rust_trampoline to a function pointer properly 2016-04-03 12:00:21 +02:00
edef 644cd81f9e fix parameter passing to arch/x86_64/swap.s 2016-04-03 11:59:14 +02:00
edef 3397e0bcf1 at least allow different lifetimes in Context::swap2
Really we should allow swapping between different stack types easily, so
we can easily swap between contexts with different stack types.
2016-03-22 08:46:27 +01:00
edef b986931817 remove copyright years
They're obnoxious to update, and I'm told they have no legal value.
2016-03-22 08:25:23 +01:00
edef 709dad1c4a add Context::swap2, to allow distinct save/restore contexts
Fix #5

Context::swap is now implemented in terms of Context::swap2, and it
might make sense to remove Context::swap entirely at some point.
2016-03-22 07:50:50 +01:00