Difference between revisions of "Mainline fast symbol resolution for kernel module loading"
(Mainlined) |
|||
Line 22: | Line 22: | ||
** [[Media:C_AMOROSO_Fast_lkm_loader_ELC-E_2009.pdf]] | ** [[Media:C_AMOROSO_Fast_lkm_loader_ELC-E_2009.pdf]] | ||
− | == | + | == Status == |
− | |||
− | + | Mainlined in v3.0 | |
− | |||
− | |||
[[Category:Project proposals 2011]] | [[Category:Project proposals 2011]] |
Latest revision as of 03:34, 24 July 2011
- Summary
- Mainline fast symbol resolution for kernel module loading
- Proposer
- Tim Bird
Description
Symbol resolution for loading kernel modules can take a long time, since the current symbol lookup mechanism in the kernel is a linear lookup.
Various projects have demonstrated improved symbol resolution time for kernel modules, using either binary trees or hash tables. Each approach has its pros and cons, but as of January 2010 neither approach has been mainlined.
This project would consist of mainlining one of these approaches to reducing symbol resolution time for kernel loadable modules.
Related work
- Alan Jenkins binary search patches
- Carmelo Amoroso hash table approach
Status
Mainlined in v3.0