Commit Graph

59 Commits

Author SHA1 Message Date
Ralf Jung
454fc854ab Rename value accessors to "into_*" so the three of them are better aligned 2017-07-12 17:52:57 -07:00
Ralf Jung
c149c3fc6a Re-do packed memory accesses
We now track in the lvalue whether what we computed is expected to be aligend or not, and then set some state in the memory system accordingly to make it (not) do alignment checks
2017-07-12 17:46:56 -07:00
Oliver Schneider
03f0a88002
Use a wrapper type to differentiate between PrimVal and pointers 2017-07-11 12:28:11 +02:00
Ralf Jung
1cbf5e8962 leave notes regarding possible alignment checks 2017-07-10 18:09:46 -07:00
Oliver Schneider
9a04be974c Merge pull request #232 from oli-obk/master
Only check pointers when dereferencing
2017-07-04 09:50:54 +02:00
Oliver Schneider
10ec543b9e Merge pull request #235 from RalfJung/intrinsics
Implement some missing intrinsics
2017-07-04 08:38:22 +02:00
Ralf Jung
ffd482e2f5 check the assumptions made by the unchecked_ and copy_nonoverlapping intrinsics 2017-07-03 20:37:38 -07:00
Ralf Jung
f118ff43e7 implement the unchecked_ intrinsics; add all the doctests from one of the integer modules 2017-07-03 20:37:33 -07:00
Ralf Jung
030166757c Fix transmute on ByValPair
Fixes #227
2017-07-03 13:58:50 -07:00
Oliver Schneider
823b952ef2 Only check pointers when dereferencing
Before we also checked whether pointers had alloc_ids when we created rvalue references
2017-07-03 14:16:11 +02:00
Oliver Schneider
5dfaacf310 Simplify all the code 2017-06-27 13:36:41 +02:00
Ralf Jung
269667e152 implement _nonzero intrinsics 2017-06-22 14:46:53 -07:00
Oliver Schneider
8de111018d Merge pull request #201 from oli-obk/test_suite_fails
"Support" more libc functions
2017-06-22 14:16:45 +02:00
Ralf Jung
7b2b0dd56c test HashMap creation in libstd-MIR, and make it work again 2017-06-21 23:09:34 -07:00
Oliver Schneider
5414825f09
Simplify numeric intrinsics 2017-06-21 15:08:19 +02:00
Oliver Schneider
ea6f6079ca
Use PrimVal instead of Pointer where applicable 2017-06-20 14:28:19 +02:00
Oliver Schneider
fcf495821e
Get rid of the integer allocation 2017-06-20 14:28:18 +02:00
Ralf Jung
e1562fbe71 comments 2017-06-06 10:15:54 -07:00
David Renshaw
49fb43f293 use type_align() and type_size() instaed of size_and_align_of_dst() 2017-06-06 10:29:53 -04:00
David Renshaw
c8be312933 fix issue 184 by marking the destination as a packed struct 2017-06-06 09:49:34 -04:00
Ralf Jung
f39e015163 check for overflow when doing pointer arithmetic 2017-06-05 14:23:35 -07:00
Ralf Jung
70227c87bf fix arith_offset not taking the size of the type into account; test for offset 2017-06-05 14:23:35 -07:00
Ralf Jung
cfff91ba3e write_bytes intrinsic: if the write count is 0, the pointer does not have to be valid 2017-06-05 14:23:34 -07:00
Ralf Jung
1b5f77e4c1 Implement Offset like the other binary operators, share code with the intrinsic
Also improve drop glue tests
2017-06-05 14:19:10 -07:00
Ralf Jung
31cf66d0e8 remove our array drop glue and use rustc's instead; implement the new Offset and SizeOf operators 2017-06-05 10:06:40 -07:00
David Renshaw
ca8347a1ff update for upstream ParamEnv changes 2017-06-02 21:00:35 -04:00
David Renshaw
48662d5199 update for upstream changes with ty::ParamEnv 2017-05-23 23:40:39 -04:00
David Renshaw
d666bd7e62 update for latest rustc nightly: type_needs_drop_given_env() and type_contents() are gone 2017-04-23 13:45:04 -04:00
Oliver Schneider
fdeee8fb59 Cleanup the diff 2017-03-29 09:10:05 +02:00
Oliver Schneider
dc1b0fb436 Compiles again 2017-03-21 13:53:55 +01:00
Scott Olson
a82924b3e2 Merge pull request #152 from oli-obk/intrinsics
Intrinsics and overflow
2017-03-14 13:35:51 -07:00
Oliver Schneider
bb33830c60 Implement more atomic intrinsics 2017-03-14 13:05:51 +01:00
Oliver Schneider
adddde7cba Implement more float intrinsics 2017-03-14 12:35:38 +01:00
David Renshaw
f0e2247f4d fix argument order on check_align 2017-03-13 21:07:25 -04:00
David Renshaw
41c2aa677a simplify write_bytes() by using memory.write_repeat() 2017-03-13 20:47:08 -04:00
David Renshaw
e79ee140ee write_bytes(): fix handling of types that are larger than u128 2017-03-13 20:32:43 -04:00
David Renshaw
2f3440d213 implement write_bytes intrinsic 2017-03-13 19:46:44 -04:00
Oliver Schneider
1a697f9bba
move all code accessing vtable internals into the trait module
fixes #124
2017-02-24 11:04:13 +01:00
Scott Olson
f73f001ce5 Fix fabsf{32,64} intrinsics. 2017-02-10 14:53:56 -08:00
Scott Olson
98d28b688c Merge pull request #139 from oli-obk/master
lvalue references into fields of ValPair locals
2017-02-10 22:29:54 +01:00
Oliver Schneider
0b86d30594 enable Lvalue::Local to refer to a ValPair field 2017-02-10 16:14:59 +01:00
Scott Olson
5a2cdc2689 Implement the fast float math intrinsics. 2017-02-10 05:58:34 -08:00
Scott Olson
14d7501ca4 Merge pull request #128 from oli-obk/align_thyself
fix size of dst in size_of_val intrinsic
2017-02-09 15:35:15 +01:00
Oliver Schneider
e6006c35fb don't unnecessarily convert Align -> u64 -> Align 2017-02-09 13:00:30 +01:00
Oliver Schneider
0d3cee2db3 fix size of dst in size_of_val intrinsic 2017-02-09 11:15:56 +01:00
Oliver Schneider
1844381ad0 compute the offset of dst fields by checking the vtable 2017-02-09 08:38:01 +01:00
Scott Olson
7acbf7ef74 Cleanup: EvalResult<'a, ()> becomes EvalResult<'a>. 2017-02-04 13:09:10 -08:00
Oliver Schneider
24870428a7
more intrinsics 2017-01-12 12:43:37 +01:00
Oliver Schneider
3a658e09e8
rustup (i128) 2017-01-12 08:28:42 +01:00
Scott Olson
b233ada529 Change Option<Value> to Value, using ByVal(Undef).
This job isn't quite finished because it caused me to discover bugs
related to reading `ByVal(Undef)` when a `ByValPair` is expected, e.g.
for a fat pointer. This wasn't a problem with the `None` of
`Option<Value>`, but I realized an equivalent bug existed even then,
since you could transmute a `u64` like `ByVal(Bytes(42))` to a fat
pointer type on 32-bit targets.

Likewise, you could transmute a fat pointer to `u64` and get panics
related to expecting `ByVal` but finding `ByValPair`, so the problem
goes both ways.
2016-12-18 20:59:01 -08:00