freebsd-nq/sys
Poul-Henning Kamp b15a966ec6 1. Add a {pointer, v_id} pair to the vnode to store the reference to the
".." vnode.  This is cheaper storagewise than keeping it in the
    namecache, and it makes more sense since it's a 1:1 mapping.

2.  Also handle the case of "." more intelligently rather than stuff
    the namecache with pointless entries.

3.  Add two lists to the vnode and hang namecache entries which go from
    or to this vnode.  When cleaning a vnode, delete all namecache
    entries it invalidates.

4.  Never reuse namecache enties, malloc new ones when we need it, free
    old ones when they die.  No longer a hard limit on how many we can
    have.

5.  Remove the upper limit on namelength of namecache entries.

6.  Make a global list for negative namecache entries, limit their number
    to a sysctl'able (debug.ncnegfactor) fraction of the total namecache.
    Currently the default fraction is 1/16th.  (Suggestions for better
    default wanted!)

7.  Assign v_id correctly in the face of 32bit rollover.

8.  Remove the LRU list for namecache entries, not needed.  Remove the
    #ifdef NCH_STATISTICS stuff, it's not needed either.

9.  Use the vnode freelist as a true LRU list, also for namecache accesses.

10. Reuse vnodes more aggresively but also more selectively, if we can't
    reuse, malloc a new one.  There is no longer a hard limit on their
    number, they grow to the point where we don't reuse potentially
    usable vnodes.  A vnode will not get recycled if still has pages in
    core or if it is the source of namecache entries (Yes, this does
    indeed work :-)  "." and ".." are not namecache entries any longer...)

11. Do not overload the v_id field in namecache entries with whiteout
    information, use a char sized flags field instead, so we can get
    rid of the vpid and v_id fields from the namecache struct.  Since
    we're linked to the vnodes and purged when they're cleaned, we don't
    have to check the v_id any more.

12. NFS knew about the limitation on name length in the namecache, it
    shouldn't and doesn't now.

Bugs:
        The namecache statistics no longer includes the hits for ".."
        and "." hits.

Performance impact:
        Generally in the +/- 0.5% for "normal" workstations, but
        I hope this will allow the system to be selftuning over a
        bigger range of "special" applications.  The case where
        RAM is available but unused for cache because we don't have
        any vnodes should be gone.

Future work:
        Straighten out the namecache statistics.

        "desiredvnodes" is still used to (bogusly ?) size hash
        tables in the filesystems.

        I have still to find a way to safely free unused vnodes
        back so their number can shrink when not needed.

        There is a few uses of the v_id field left in the filesystems,
        scheduled for demolition at a later time.

        Maybe a one slot cache for unused namecache entries should
        be implemented to decrease the malloc/free frequency.
1997-05-04 09:17:38 +00:00
..
alpha Removed unused #includes. 1997-04-09 15:34:09 +00:00
amd64 added declaration for get_isa_apic_mask(). 1997-05-03 18:05:31 +00:00
compat/linux Always include PROT_READ for Linux mmap operations. 1997-04-28 02:53:17 +00:00
compile
conf add net/if_media.c 1997-05-03 21:06:00 +00:00
ddb Everyone's favorite, i think: make DDB understand the arrow keys for the 1997-04-12 17:35:02 +00:00
dev Initialize PCI/CardBus bridges. 1997-05-03 13:52:29 +00:00
fs Fixed panic message in union_lock(): union_link --> union_lock. 1997-05-03 01:55:19 +00:00
geom Make ccd use the maxsecsize sector size as denominator, this 1997-05-01 19:18:40 +00:00
gnu Make the necessary changes so that an ELF kernel can be built. I 1997-04-22 06:55:47 +00:00
i386 disabled checks for smp_active == 0. 1997-05-03 19:24:16 +00:00
isa Restore CRTC start address check back after looking into more sources 1997-04-30 15:31:22 +00:00
isofs/cd9660 For multi-session CD-ROMs, we have to account for previous sessions as 1997-04-29 17:11:51 +00:00
kern 1. Add a {pointer, v_id} pair to the vnode to store the reference to the 1997-05-04 09:17:38 +00:00
libkern Fix a bug that caused some false mismatches when both FNM_PATHNAME 1997-04-29 03:24:57 +00:00
miscfs Fix the umount problems for DEVFS. 1997-05-03 21:19:54 +00:00
modules Also quiet down the ibcs2 startup (same reason as linux lkm - it screws 1997-05-01 06:09:01 +00:00
msdosfs Get the declaration of `struct dirent' from <sys/dirent.h>, not from 1997-04-10 14:56:49 +00:00
net add SIOC{S,G}IFMEDIA ioctl support 1997-05-03 21:07:13 +00:00
netatalk Fixed the type of timeout functions and removed casts that hid the 1997-04-20 15:36:12 +00:00
netinet The long-awaited mega-massive-network-code- cleanup. Part I. 1997-04-27 20:01:29 +00:00
netipx Oops I missed one of the pr_usrreqs changes. 1997-05-01 12:24:20 +00:00
netkey This is the `netkey' kernel key-management service (the PF_KEY analogue 1996-06-14 17:22:18 +00:00
netns Remove unneeded member nsp_head from struct nspcb. 1997-04-14 18:26:51 +00:00
nfs 1. Add a {pointer, v_id} pair to the vnode to store the reference to the 1997-05-04 09:17:38 +00:00
nfsclient 1. Add a {pointer, v_id} pair to the vnode to store the reference to the 1997-05-04 09:17:38 +00:00
nfsserver Allow NULL rpcs on non-privileged ports at all times to work around broken 1997-04-30 09:51:37 +00:00
pc98 Synchronize with sys/i386/isa/syscons.c up to revision 1.213. 1997-05-01 11:28:49 +00:00
pccard Recognize, but don't actually do much with a TI PCI1130 PcCard/CardBus bridge. 1997-05-03 13:57:26 +00:00
pci Initialize PCI/CardBus bridges. 1997-05-03 13:52:29 +00:00
powerpc/include Back out part 1 of the MCFH that changed $Id$ to $FreeBSD$. We are not 1997-02-22 09:48:43 +00:00
rpc Revert $FreeBSD$ to $Id$ 1997-02-23 09:21:14 +00:00
scsi Add the Archive 4586NP tape changer. 1997-05-03 22:23:13 +00:00
sys 1. Add a {pointer, v_id} pair to the vnode to store the reference to the 1997-05-04 09:17:38 +00:00
tools Back out part 1 of the MCFH that changed $Id$ to $FreeBSD$. We are not 1997-02-22 09:48:43 +00:00
ufs Mount MFS read/write as in days of yore. 1997-04-28 09:15:18 +00:00
vm Check the correct queue for waking up the pageout daemon. Specifically, 1997-05-01 14:36:01 +00:00
Makefile Back out part 1 of the MCFH that changed $Id$ to $FreeBSD$. We are not 1997-02-22 09:48:43 +00:00