Commit message (Collapse)AuthorAgeFilesLines
* Add FIXME: watch out for excessive fragmentation when shrinking anto-tytso-20011015H. Peter Anvin2001-10-151-0/+4
| | | | allocation with realloc().
* Correct the computation of current_size in realloc(). It now passes theH. Peter Anvin2001-10-151-1/+1
| | | | stress tester all the way through.
* Add code to verify that none of the allocated memory regions overlap.H. Peter Anvin2001-10-151-2/+27
* Now realloc() mostly works... checkpoint the work here.H. Peter Anvin2001-10-153-53/+122
* Add realloc; clean up prototypes and comments.H. Peter Anvin2001-10-151-7/+8
* Implement realloc(). Not tested yet.H. Peter Anvin2001-10-151-7/+170
* Start with an arena size of 128K. We can extend it at our leisure.to-tytso-20011013H. Peter Anvin2001-10-131-1/+1
* Now we have functioning extension of the arena.H. Peter Anvin2001-10-131-27/+147
* Replace debugging printf()'s with DPRINTF()H. Peter Anvin2001-10-131-19/+24
* Modify the bitmap definition so that we use positive offsets, which makeH. Peter Anvin2001-10-133-42/+48
| | | | life *so* much easier to deal with. Fix a couple of bogus (c) notices.
* Add back pointers so we can reclaim SLAB pages easier; make work betterH. Peter Anvin2001-10-131-24/+25
| | | | | | on 64-bit systems. Still need: realloc(), arena extension.
* Working allocation and freeing; still need reverse list pointerH. Peter Anvin2001-10-133-52/+171
| | | | to return completely freed slab page to the buddy system.
* Mostly working allocator (slab and buddy), although extension of theH. Peter Anvin2001-10-135-77/+578
| | | | arena currently doesn't work. No deallocator yet, though.
* Create a Makefile; fix memory-overwrite bug when file_len > requested_lenH. Peter Anvin2001-10-093-11/+57
* Split the internals into a separate header file.H. Peter Anvin2001-07-084-41/+59
* Initial version under CVS controlH. Peter Anvin2001-07-087-0/+955