Aren/NetBSD
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity
156,307 Commits 606 Branches 0 Tags 3.1 GiB
eebb2036b8
Commit Graph

175 Commits

Author SHA1 Message Date
pooka
27f51110f8 lookup in-core nodes in readdir() and provide the real inode for 
each node.  This makes pwd work.
 2007-02-16 16:39:56 +00:00
pooka
34faca38b1 * introduce path comparison method for the sysctl MIB paths 
* use puffs_path_walkcmp() instead of a homegrown routine
 2007-02-15 17:07:31 +00:00
pooka
3f1bccc16c Use new libpuffs dotdot features to get rid of embedding the directory 
hierarchy in the in-memory nodes.  This allows us to easily do
reclaim, since we don't have to worry about reclaiming a node which
has alive children which might want to dotdot lookup later on.
 2007-02-15 12:59:22 +00:00
pooka
5f82fb0670 support reclaim 2007-02-15 12:54:52 +00:00
pooka
70d42dbba1 do reclaim 2007-02-15 12:54:24 +00:00
pooka
9c41498b0b PUFFS_ISDOTDOT -> PCNISDOTDOT 2007-02-15 12:52:30 +00:00
gdt
5ff979921c Even though the parent doesn't descend here, it's still convenient to 
have a makefile to build the 4 examples at once.
 2007-01-30 01:26:58 +00:00
pooka
c1b9179682 Avoid doing a ludicrous number of reallocs when the file grows by only 
lazy extending the file in setattr and really extending it in write.
 2007-01-28 10:47:36 +00:00
pooka
f235a8d60c Add very uncorrect "suspend fs if server gets SIGUSR1" test for fs 
suspension (hey, it's a test file system after all .... )
 2007-01-26 23:02:05 +00:00
elad
a78693aa19 Kill KAUTH_PROCESS_RESOURCE and just replace it with two actions for 
nice and rlimit.
 2007-01-20 16:47:38 +00:00
elad
e52ee73ea8 arg0 is always 'struct proc *' for the process scope. 2007-01-15 17:05:01 +00:00
pooka
b5da39fea6 Gur Zbba vf Jnavat Perfprag (19% bs Shyy) 
Add a little layering experiment, which both exercises the
componentname modification feature of libpuffs and tries to do
layering (just a little too hardcoded for comfort, fow now).  This
simply uses the puffs nullfs to present all data and paths in rot13.
You can even mount it twice for extra fun.
 2007-01-15 00:46:29 +00:00
pooka
e6ea11c9c0 use new style to indicate rootpath 2007-01-15 00:43:07 +00:00
pooka
9a25a94ef7 adapt to libpuffs changes 2007-01-15 00:41:09 +00:00
pooka
14995912ad use the libpuffs path management routines to help build the sysctl 
mib names
 2007-01-15 00:40:37 +00:00
pooka
dd15744c39 * register fsync operation 
* test for -s, which causes no detach (don't need to remember to
  do follow-dances in gdb because of fork() used in daemon())
 2007-01-11 17:50:35 +00:00
pooka
764f2241da * initialize pflags properly 
* make 'dump' imply 'don't detach'
 2007-01-11 11:52:53 +00:00
pooka
cf4b2a379f example null file system built on top of the null layer in libpuffs 2007-01-11 01:03:03 +00:00
pooka
a67e4ce37f * do mount as a forward operation instead of a callback 
* process -o args for mount
 2007-01-06 18:25:19 +00:00
elad
a75d82b0bc We no longer have 'enum kauth_machdep_req'. 2007-01-05 13:23:22 +00:00
elad
a13160f423 Make mount(2) and unmount(2) use kauth(9) for security policy. 
Okay yamt@.
 2007-01-02 10:47:28 +00:00
pooka
d584a8d23a SUSv3 says: should the implementation not support directory unlinking, 
unlink shall fail with EPERM and the errno of the failure shall be EPERM
 2007-01-01 21:32:12 +00:00
pooka
6a42c9c838 apply wide-angle rotollation beam to match libpuffs changes 2006-12-29 15:37:06 +00:00
pooka
fb7890b666 ssshfs is obsoleted by mount_psshfs 2006-12-29 15:36:18 +00:00
elad
504c71d9fe Make machdep scope architecture-agnostic by removing all arch-specific 
requests and centralizing them all. The result is that some of these
are not used on some architectures, but the documentation was updated
to reflect that.
 2006-12-26 10:43:43 +00:00
elad
3d11477c94 Add requests indicating access to unmanaged memory for arm, pc532, powerpc, 
sh3, sh5, and vax, and use them instead of KAUTH_GENERIC_ISSUSER.

Update documentation and example secmodel code.
 2006-12-22 11:13:21 +00:00
elad
238ad51d2d - moves 'nice' access semantics to secmodel code, 
- makes sysctl_proc_find() just lookup the process,
- use KAUTH_PROCESS_CANSEE requests to determine if the caller is
  allowed to view the target process' corename, stop flags, and
  rlimits,
- use explicit kauth(9) calls with KAUTH_PROCESS_CORENAME,
  KAUTH_REQ_PROCESS_RESOURCE_NICE, KAUTH_REQ_PROCESS_RESOURCE_RLIMIT,
  and KAUTH_PROCESS_STOPFLAG when modifying the aforementioned.
- sync man-page and example skeleton secmodel with reality.

okay yamt@

this is a pullup candidate.
 2006-12-14 11:45:08 +00:00
manu
098ea45eb4 New README.plainrsa sample file 2006-12-10 05:52:10 +00:00
pooka
fe53828f70 use command line flags to set various options 2006-12-07 22:49:04 +00:00
pooka
7e3489d64a write requires setattr, so do a dummy version to make write here work 
again after the checkops kernel changes
 2006-12-07 10:59:03 +00:00
pooka
fb5e346800 counter-rototill for libpuffs change: unify all callback ops under 
a single structure
 2006-12-07 10:54:29 +00:00
pooka
44576f90f6 fix IO_APPEND for write 2006-12-05 23:14:13 +00:00
pooka
05b1899256 only count directories in the number of hard links to a directory 2006-12-05 14:32:03 +00:00
pooka
69dbc08863 update flag names 2006-12-01 12:50:52 +00:00
pooka
496f59a99f adapt to flags parameter for puffs_mainloop() 2006-11-30 05:38:54 +00:00
wiz
fa34b615d2 Correct spelling of "immediate(ly)". From Zafer. 2006-11-24 21:23:07 +00:00
hubertf
1ee989e6be Don't try to handle symlinks on versions of the SSH File Transfer Protocol 
that don't support it, e.g. the Solaris 9/x86 sshd (SSH-2.0-Sun_SSH_1.0;
..._1.1 from Solaris 10/x86 is fine)

OK'd by pooka@
 2006-11-23 01:51:15 +00:00
elad
4b316db1d1 Introduce KAUTH_REQ_MACHDEP_{ALPHA,X86}_UNMANAGEDMEM to handle access 
to unmanaged memory.

These are the last two securelevel references in the MD code.
 2006-11-22 12:12:51 +00:00
pooka
118c65eb6d * update comments 
* puffs_nextdent() takes a directory type, not vtype, so translate
  that before calling it.  fixes find & other fts-stuff.
 2006-11-21 23:19:45 +00:00
pooka
5676e026f8 * change usage to ssshfs user@host:path mountpath 
* assorted little fixes

ssshfs now works for practical uses
 2006-11-21 23:09:23 +00:00
pooka
2773d88a59 do better directory link accounting, fix bug in readdir(). find, rm -rf, 
etcetc. now work
 2006-11-21 15:35:58 +00:00
pooka
a9772ae2fb better accounting. e.g. du works now 2006-11-21 13:05:47 +00:00
pooka
e7a63c481b don't recache directory listing in all operations that use a directory 2006-11-21 12:42:15 +00:00
pooka
8f2fd4cb28 reclaim needs some more work, so don't do anything there for now. 
This is good especially for not crashing in unmount, which in turn
would cause unpleasantries without the very latest & greatest kernel
 2006-11-21 02:31:48 +00:00
pooka
1d9515c070 today's fs-in-a-day: quick glue to tie puffs together with openssh 
sftp code to accomplish something close to sshfs.  this basically
works, but there are some issues with directory grovelers still,
e.g. so untar won't work.  but it works for browsing mailboxes,
transferring large files, etcetc.
 2006-11-21 00:54:06 +00:00
pooka
66e8fee83b implement puffs_write. It's now possible to change the values of 
existing sysctl nodes (except for CTLTYPE_STRUCT).
 2006-11-20 00:04:05 +00:00
pooka
4d33479aef * no more fsid 
* no more start()
 2006-11-18 12:41:06 +00:00
pooka
a2646bee20 mount with PUFFSFLAG_NOCACHE 2006-11-17 17:48:51 +00:00
pooka
48697c774d do correct size calculation in getattr(). accidentally replaced that 
with a dummy before commit.
 2006-11-17 00:27:16 +00:00
pooka
2f8a6afe2a fs du jour: sysctlfs, which mounts an equivalent of "sysctl -a" as 
a fs hierarchy.  This is currently ro, lacking features, and very
much unfinished in every other possible sense.  Should investigate
if this can share code with sbin/sysctl so that everything must
not be rewritten.
 2006-11-14 21:37:36 +00:00