Messages in this thread | | | From | "Daniel J Blueman" <> | Subject | kflushd rewrite... | Date | Mon, 13 Sep 1999 11:23:55 +0100 |
| |
Having looked at the source for kflushd, I did a complete re-write offering significantly less overhead. I am preparing a patch against 2.3.18.
Just a few questions to you all for a few refinements, before I send the patch to Linus:
- is the pointer to the buffer head from 'bh = lru_list[n]' a valid buffer_head? ie can it contain dirty data?
- is the first buffer_head in the chain, ie bh = lru_list[n] static, and can we assume no one will move it?
- is it better to check is a buffer is ready to be written, and write it before checking if it's on the 'wrong' list and refiling it? (I think this may avoid a refile if the buffer's status has recently changed)
- the waking up seems to be hardwired to 5s. Should this delay be pulled from the bdflush settings? (bdf_prm.b_un.<foo>)
- what other process fields could benefit from being filled in (eg current->comm etc) ?
Any feedback on these will help, since I'm going to get this patch out in the next couple of days, and bury the hatchet.
All replies CC'd to this email address.
Thankx, Dan __________________________ Daniel J Blueman - daniel.j.blueman@stud.umist.ac.uk Undergraduate - BSc Computing Science UMIST university - Manchester
- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.rutgers.edu Please read the FAQ at http://www.tux.org/lkml/
| |