Set up a test framework. This lays out a common test framework structure that should make it simpler to extend tests in the future. Signed-off-by: Charles Manning <cdhmanning@gmail.com>
Clean up obsolete CVS version stamps and some other messiness. Just obsolete and benign messiness being cleaned up. Signed-off-by: Charles Manning <cdhmanning@gmail.com>
yaffs: Refactor drivers WIP - stress tests passing Both yaffs1 and yaffs2 modes seem to be doing the right thing under stress testing. Signed-off-by: Charles Manning <cdhmanning@gmail.com>
yaffs for u-boot: Cleanups to make patches acceptable for u-boot The u-boot patches were not acceptable to the u-booters. Hopefully yhey will be happy now. Signed-off-by: Charles Manning <cdhmanning@gmail.com>
Set up u-boot glue code and patching scripts. This should patch yaffs2 into a u-boot system. Signed-off-by: Charles Manning <cdhmanning@gmail.com>
Change Makefiles to support new way to structure yaffs_strxxx renaming for yaffs direct Signed-off-by: Charles Manning <cdhmanning@gmail.com>
Set up new version of case insensitive code using sed This version uses sed to change strncmp->yaffs_strncmp etc. This emnables us to keep the core code free of yaffs_strxxx macros. Signed-off-by: Charles Manning <cdhmanning@gmail.com>
yaffs: Add first cut of block summary code Some issues still need to be checked. In particular: * Is free space handling correct? * Is gc doing The Right Thing? Signed-off-by: Charles Manning <cdhmanning@gmail.com>
yaffs: Remove tags validity checking code. It was only really useful during development and no longer serves any purpose. Signed-off-by: Charles Manning <cdhmanning@gmail.com>
yaffs: Refactor bit counting code For Linux kernel we use hweight8() and hweight32(). This allows use of special bit-counting code instructions in processors that do this. The yaffs direct we now only use one set of code. There were two. NB On x86 the yaffs hweight8 code is actually a lot faster than the sofwtare implementation in Linux kernel. Signed-off-by: Charles Manning <cdhmanning@gmail.com>
yaffs direct: Partition off os glue code This makes the porting process a bit cleaner. Signed-off-by: Charles Manning <cdhmanning@gmail.com>
yaffs: Remove SHORT_NAMES_IN_RAM and WIDE_TNODE_DISABLE configs These have been removed from Linux as their use is discouraged except for debugging purposes. The logic has also changed for SHORT_NAMES_IN_RAM. This is now negative logic. Set CONFIG_YAFFS_NO_SHORT_NAMES to **disable** short names. Signed-off-by: Charles Manning <cdhmanning@gmail.com>
yaffs: More clean up Signed-off-by: Charles Manning <cdhmanning@gmail.com>
yaffs: More clean up. Signed-off-by: Charles Manning <cdhmanning@gmail.com>
yaffs: More clean up Signed-off-by: Charles Manning <cdhmanning@gmail.com>
yaffs Further name changes and file reorganisation Signed-off-by: Charles Manning <cdhmanning@gmail.com>
yaffs Move yaffscfg.h to where it should be Signed-off-by: Charles Manning <cdhmanning@gmail.com>
yaffs Refactor yaffs direct device list management, add nand abstraction, divide up test code This all makes for a far cleaner layout of test code. The new yaffs direct list management allows partitions to be added synamically rather than statically. This is better for OSs that install partitions at runtime (eg. WInCE). The nand abstraction makes it tidier to add a new nand driver to yaffs direct and test code. The test code has been repartitioned to make porting and file management cleaner. Signed-off-by: Charles Manning <cdhmanning@gmail.com>
yaffs Refactorisation first phase Partitoning code. More to happen. Signed-off-by: Charles Manning <cdhmanning@gmail.com>
yaffs Reactoring WIP Signed-off-by: Charles Manning <cdhmanning@gmail.com>