Commit message (Collapse) | Author | Age | Files | Lines | |
---|---|---|---|---|---|
* | testsuite: Force-enable caret diagnostics in T17786wip/windows-clang | GHC GitLab CI | 2022-01-29 | 1 | -1/+1 |
| | | | | | Otherwise GHC realizes that it's not attached to a proper tty and will disable caret diagnostics. | ||||
* | testsuite: Make T7275 more robust against CCid changes | GHC GitLab CI | 2022-01-29 | 2 | -17/+17 |
| | | | | The cost-center numbers are somewhat unstable; normalise them out. | ||||
* | rts: Don't allocate closurePtrs# pointers on C stack | GHC GitLab CI | 2022-01-28 | 1 | -2/+5 |
| | | | | | | Previously `closurePtrs#` would allocate an aray of the size of the closure being decoded on the C stack. This was ripe for overflowing the C stack overflow. This resulted in `T12492` failing on Windows. | ||||
* | testsuite/T4029: Don't depend on echo | GHC GitLab CI | 2022-01-28 | 1 | -108/+108 |
| | | | | | | On Windows the `cmd.exe` shell may be used to execute the command, which will print `ECHO is on.` instead of a newline if you give it no argument. Avoid this by rather using `printf`. | ||||
* | PATH_FMT | GHC GitLab CI | 2022-01-28 | 1 | -1/+1 |
| | |||||
* | m32 | GHC GitLab CI | 2022-01-28 | 1 | -1/+1 |
| | |||||
* | m32 | GHC GitLab CI | 2022-01-28 | 1 | -2/+3 |
| | |||||
* | rts: Generalize mmapForLinkerMarkExecutable | GHC GitLab CI | 2022-01-28 | 7 | -17/+57 |
| | | | | | Renamed to mprotectForLinker and allowed setting of arbitrary protection modes. | ||||
* | squash: m32 | GHC GitLab CI | 2022-01-28 | 1 | -13/+17 |
| | |||||
* | PATH_FMT | GHC GitLab CI | 2022-01-28 | 2 | -2/+2 |
| | |||||
* | squash: m32 consistency | GHC GitLab CI | 2022-01-28 | 1 | -29/+32 |
| | |||||
* | rts/m32: Increase size of free page pool to 256 pages | GHC GitLab CI | 2022-01-28 | 1 | -1/+1 |
| | |||||
* | rts/m32: Free large objects back to the free page pool | GHC GitLab CI | 2022-01-28 | 1 | -11/+87 |
| | | | | | Since this is a somewhat subtle thing to get right I also added quite a bit of consistency checking. | ||||
* | Kill more debug output | GHC GitLab CI | 2022-01-28 | 2 | -9/+2 |
| | |||||
* | Clean up debug output | GHC GitLab CI | 2022-01-28 | 2 | -37/+29 |
| | |||||
* | Linking | GHC GitLab CI | 2022-01-28 | 2 | -60/+21 |
| | |||||
* | rts: Don't mark object code in markCAFs unless necessary | GHC GitLab CI | 2022-01-28 | 1 | -2/+2 |
| | |||||
* | Progress | GHC GitLab CI | 2022-01-28 | 6 | -243/+380 |
| | |||||
* | rts/PEi386: Fix memory leak | GHC GitLab CI | 2022-01-27 | 1 | -1/+3 |
| | |||||
* | rts/PEi386: Fix address space probing | GHC GitLab CI | 2022-01-27 | 2 | -71/+75 |
| | |||||
* | PATH_FMT | GHC GitLab CI | 2022-01-27 | 1 | -1/+1 |
| | |||||
* | rts/PEi386: Move some debugging output to -DL | GHC GitLab CI | 2022-01-27 | 1 | -0/+4 |
| | |||||
* | rts/PEi386: Use m32 for image allocation | GHC GitLab CI | 2022-01-27 | 1 | -66/+5 |
| | |||||
* | rts: Fix missing PATH_FMTs | GHC GitLab CI | 2022-01-27 | 1 | -3/+3 |
| | |||||
* | linker: add allocation to high mem. | Tamar Christina | 2022-01-27 | 3 | -17/+151 |
| | |||||
* | rts: Mark anything that might have an info table as data | GHC GitLab CI | 2022-01-21 | 1 | -268/+272 |
| | |||||
* | Bump text submodule | GHC GitLab CI | 2022-01-21 | 1 | -0/+0 |
| | |||||
* | mim32 fixes | GHC GitLab CI | 2022-01-21 | 1 | -3/+3 |
| | |||||
* | indirect symbol type | GHC GitLab CI | 2022-01-21 | 4 | -6/+35 |
| | |||||
* | rts/PEi386: Fix handling of overflowing data relocations | GHC GitLab CI | 2022-01-21 | 7 | -41/+78 |
| | | | | | | | | | | As noted in #20978, the linker would previously handle overflowed relocations by creating a trampoline. While this is fine in the case of code symbols, it's very much not okay in the case of data symbols. To fix this we keep track of whether each symbol is code or data and relocate them appropriately. Fixes #20978. | ||||
* | m32: Don't truncate next pointer | GHC GitLab CI | 2022-01-21 | 1 | -3/+2 |
| | |||||
* | movabs | GHC GitLab CI | 2022-01-21 | 1 | -0/+20 |
| | | | | Thanks to Sylvain | ||||
* | rts/m32: Accept any address within 4GB displacement | GHC GitLab CI | 2022-01-20 | 1 | -3/+10 |
| | |||||
* | Build system hacks for clang | GHC GitLab CI | 2022-01-20 | 2 | -5/+3 |
| | |||||
* | rts: Eliminate use of nested functions | GHC GitLab CI | 2022-01-20 | 1 | -9/+11 |
| | | | | This is a gcc-specific extension. | ||||
* | Drop ld.bfd-specific linker flags | GHC GitLab CI | 2022-01-20 | 1 | -10/+3 |
| | |||||
* | Fix jmp table addressing | GHC GitLab CI | 2022-01-20 | 1 | -2/+6 |
| | | | | | On Windows assemblers appear to assume that indirect JMP operands can be relocated with a 32-bit relocation. | ||||
* | Generate LEA for label expressions | GHC GitLab CI | 2022-01-20 | 1 | -0/+14 |
| | |||||
* | Don't assume that labels are 32-bit literals | GHC GitLab CI | 2022-01-20 | 1 | -3/+3 |
| | |||||
* | Bump win32-tarballs to 0.5 | Ben Gamari | 2022-01-18 | 2 | -2/+2 |
| | |||||
* | Add test for #20938. | Andreas Klebinger | 2022-01-14 | 3 | -1/+11 |
| | |||||
* | reinsert target | Eric Lindblad | 2022-01-14 | 1 | -0/+1 |
| | |||||
* | update URLs | Eric Lindblad | 2022-01-14 | 1 | -16/+16 |
| | |||||
* | Revert "add NUMJOBS" | Eric Lindblad | 2022-01-14 | 1 | -3/+1 |
| | | | This reverts commit c0b854e929f82c680530e944e12fad24f9e14f8e | ||||
* | add NUMJOBS | Eric Lindblad | 2022-01-14 | 1 | -1/+3 |
| | |||||
* | release notes: Changes to Demand analysis | Sebastian Graf | 2022-01-13 | 1 | -0/+11 |
| | |||||
* | release notes: Changes to CPR analysis | Sebastian Graf | 2022-01-13 | 2 | -3/+69 |
| | |||||
* | hadrian: Include bash completion script in bindist | Ben Gamari | 2022-01-13 | 1 | -0/+5 |
| | | | | See #20802. | ||||
* | testsuite: Fix import on python 3.10 | Zubin Duggal | 2022-01-12 | 1 | -1/+2 |
| | |||||
* | Untangled GHC.Types.Id.Make from the driver | Sylvain Henry | 2022-01-12 | 7 | -35/+44 |
| |