libk  History of mod/kmem/heapa.fn.c of acb4a9944e5a48cbab044b5ec26e822496624e9a7e7f74abf8cb9ace70b86e81

History of the file that is called mod/kmem/heapa.fn.c at checkin acb4a9944e5a48cbab044b5ec26e822496624e9a7e7f74abf8cb9ace70b86e81

2019-08-24
23:02
stop attempting to auto-detect system constants during build process, and maintain os/arch-specific tables instead file: [14fc9d1a27] check-in: [e42b590b45] user: lexi, branch: trunk, size: 3686 [annotate] [blame] [check-ins using] [diff]
00:20
improve comments file: [99ad7a3351] check-in: [858eba7607] user: lexi, branch: trunk, size: 3692 [annotate] [blame] [check-ins using] [diff]
2019-08-22
08:44
add kmlini() and kmlina() functions; restructure allocation functions to work more reasonably (returning a tuple struct instead of making a user pass in a void**); update docs accordingly file: [c08b42ee14] check-in: [acb4a9944e] user: lexi, branch: trunk, size: 2996 [annotate] [blame] [check-ins using] [diff]
04:31
finish moving heap allocation/free functions to the posix syscall apparatus and deprecate the direct assembly implementations of platform_mmap; update the kmem docs to match new function signatures (and remove typos) file: [c8ccec879e] check-in: [709ffb094d] user: lexi, branch: trunk, size: 2871 [annotate] [blame] [check-ins using] [diff]
02:52
removed sneaky segfault in x86-64 syscall fn where %r8 (the register that contains the pointer to the syscall arguments from the C syscall wrapper, which need to be copied into the correct registers before the kernel is invoked) gets overwritten if the syscall valency > 5, because of overlapping ccall and syscall ABI argument registers - r8 is clobbered by argument 5 and any further attempts to use it as a ptr segfault at best. also modified the report function so that it immediate cancels compilation if a sub-process reports failure. changed allocator function signatures so they can return a condition code if the kernel reports an error; updated example code so it compiles and runs without fault. file: [e093792a2c] check-in: [e50a476efe] user: lexi, branch: trunk, size: 3183 [annotate] [blame] [check-ins using] [diff]
2019-08-21
06:00
Renamed kmem/heapa.fn.c → mod/kmem/heapa.fn.c. move modules to a subdirectory in order to keep the directory tree organized and make room for OS-specific build files file: [641176b71f] check-in: [14172a910a] user: lexi, branch: trunk, size: 2565 [annotate] [blame] [check-ins using] [diff]
2019-08-20
22:54
reorganize posix assembly code file: [641176b71f] check-in: [f85e6a07dd] user: lexi, branch: trunk, size: 2565 [annotate] [blame] [check-ins using] [diff]
2019-08-18
11:34
add memory functions file: [8818444ab8] check-in: [5393623a84] user: lexi, branch: trunk, size: 2413 [annotate] [blame] [check-ins using] [diff]
10:20
fix kmheapa() and add kmheapf() file: [aa6d2182a2] check-in: [5279674525] user: lexi, branch: trunk, size: 2203 [annotate] [blame] [check-ins using] [diff]
2019-07-26
09:51
Added: major update. fix ridiculous old type size determination mechanism. mmap is still broken and i'm not sure why; the syscall does not appear to be going through correctly - see posix_mmap, kmheapa, and kcore/testbin.exe.fn file: [f32eb209ee] check-in: [6479e060a3] user: lexi, branch: trunk, size: 1500 [annotate] [blame] [check-ins using]