CVS log for pvfs2/src/server/unstuff.sm
	Help

Request diff between arbitrary revisions / Display revisions graphically

Merged robust security code into Orange-Branch.

client/server builds, much to fix

remove extents, handle ranges, precreate pools and various other things that aren't needed in their current forms with UUIDs

changes to PVFS_handle_X

bug fix for SEGFAULT after capability creation in unstuff.sm. thanks to
mtmoore for pointing out this bug.

merge with Orange-Branch

Changed the ".perms" value for the unstuff state machine from PINT_SERVER_CHECK_WRITE to
PINT_SERVER_CHECK_ATTR.  The CHECK_WRITE perms value caused the unstuff machine to check
file permissions for the user, even when the user was the owner.  This security check
caused a problem with the "tar" command when it tried to untar a file having only
444 permissions on it.  The "tar" command used the system call
open(<filename>,O_CREAT|O_WRONLY|O_EXCL,444).  According to the Linux man page, this
combination of open parameters should be allowable.  ext3, ext4, and qfs all
allowed this combination.  NOTE:  PINT_SERVER_CHECK_ATTR allows the owner of the
file to perform an unstuff, no file permissions are checked.  If the user is not
the owner, then file permissions are checked.

Changed the ".perms" value for the unstuff state machine from PINT_SERVER_CHECK_WRITE to
PINT_SERVER_CHECK_ATTR.  The CHECK_WRITE perms value caused the unstuff machine to check
file permissions for the user, even when the user was the owner.  This security check
caused a problem with the "tar" command when it tried to untar a file having only
444 permissions on it.  The "tar" command used the system call
open(<filename>,O_CREAT|O_WRONLY|O_EXCL,444).  According to the Linux man page, this
combination of open parameters should be allowable.  ext3, ext4, and qfs all
allowed this combination.  NOTE:  PINT_SERVER_CHECK_ATTR allows the owner of the
file to perform an unstuff, no file permissions are checked.  If the user is not
the owner, then file permissions are checked.

*** empty log message ***

per-dataspace handle type precreate pools implementation

*** empty log message ***

merging Orange Branch changes in

cleaned capability generation in server unstuff sm

began cleanup of unstuff state machine

initial merge with Orange-Branch. much will be broken

lots of little fixes that have been in the works for a while

cleaned up credential verification

revert cu-security-branch to before the attempted merge with Orange-Branch

merge of B2O-Blue-Sync,HEAD,and latest of Orange.

commiting changes from merge of Orange and Blue into test branch.

merging unstuff permission check bug fix from trunk to 2-8 branch

fix bug reported by Kevin Harms: unstuff operation should use write
permission checks, as anyone who can write to a file should be able to
trigger an unstuff operation.  setattr check doesn't cover enough cases.

Merge branches

merged in changes from latest Orange-Branch

merged in changes from summer at LANL

*** empty log message ***

Setting up Orange Branch

merging hints/events code into trunk.

reverse merge from HEAD.

reverse merge of changes to HEAD since branch.  Includes small file changes.

Merging small files branch to head.  Includes server side precreation of
data files and file stuffing.

Adjust state machine functions to track error code in frame stack
following SM_RETURN.  This allows PINT_sm_pop_frame to return a valid
error code in cases where the pushed frame was just a simple nested
state machine rather than a PJMP.

removing stale comment

extra safety check

add flag field to control trove sync in get_handles() for completeness

just updating comments; punting on stuffing arbitrary layouts for now.

store complete layout struct (encoded) for stuffed files

misc cleanup

adding a comment to remember to do something here later

bug fix, using wrong attr struct to update dfile_count

unstuff works now for basic test cases

just checking in some comments to record train of thought

some sanity checking on layout type before unstuffing; simplifying the
get_handles process a bit if we are using the normal round robin layout.

cutting down on keyvals read by unstuff by relying on what getattr work
nested sm found where possible

if unstuff is called for a file that has already been unstuffed, then we
need it to behave exactly like a getattr (retrieving dfiles etc. that are
necessary for sys-io.sm).  Tinkering with using getattr nested machine to
serve that purpose.  Also added attr mask to unstuff request.

temporarily disable unstuff operation, make sure that server side unstuff
errors are propigated to client

commit of unstuff code that got missed before.

a few extras

file unstuff.sm was initially added on branch small-file-branch.