Currently, all servers and drivers run as root as they are forks of
RS. srv_fork now tells PM with which credentials to run the resulting
fork. Subsequently, PM lets VFS now as well.
This patch also fixes the following bugs:
- RS doesn't initialize the setugid variable during exec, causing the
servers and drivers to run setuid rendering the srv_fork extension
useless.
- PM erroneously tells VFS to run processes setuid. This doesn't
actually lead to setuid processes as VFS sets {r,e}uid and {r,e}gid
properly before checking PM's approval.
When an FS crashes, VFS will clean up resources tied to that FS:
- Pending requests to the FS are canceled (i.e., fail with EIO)
- Threads waiting for a reply are stopped (i.e., fail with EIO)
- Open files are marked invalid. Future operations on a file descriptor
will cause EBADF errors.
- vmnt entry is cleared, so in-flight system calls that got past the
file descriptor check but not yet talking to the crashed FS, will
fail with EIO.
- The reference counter of the mount point is decreased, effectively
removing the crashed FS from the file system tree. Descendants of
this part of the tree are unreachable by means of a path, but can
still be unmounted by feeding the block special file to unmount(2).
This patch also gets rid of the "not a known driver endpoint" messages
during shutdown.
Last_dir didn't consider paths that end in a symlink and hence didn't
actually return the last_dir when provided with one. For example,
/var/log is a symlink to /usr/log. Issuing `>/var/log' would trigger
an assert in AVFS, because /var/ is not the actual last directory; /usr/
is.
Last_dir now verifies the final component is not a symlink. If it is, it
follows the symlink and restarts finding of the last the directory.
When a lock has read-serialized and read-only locks, releasing the read-
serialized lock would not set the state to read-only when no other locks
were pending.
Some code relies on having the file descriptor in m_in.fd. Consequently,
m_in is not only used to provide syscall parameters from user space to
VFS, but also as a global variable to store temporary data within VFS.
This has the ugly side effect that m_in gets overwritten during core
dumping.*
To work around this problem VFS now uses a so called "scratchpad" to
store temporary data that has to be globally accessible. This is a simple
table indexed by process number, just like fproc. The scratchpad allows
us to store the buffer pointer and buffer size for suspended system calls
(i.e., read, write, open, lock) instead of using fproc. This makes fproc
a bit smaller and fproc iterators a bit faster. Moreover, suspension of
processes becomes simpler altogether and suspended operations on pipes
are now less of a special case.
* This patch fixes a bug where due to unexpected m_in overwriting a
coredump would fail, and consequently resources are leaked. The coredump
was triggered with:
$ a() { a; }
$ a
The nbyte in read(int fildes, void *buf, size_t nbyte) is unsigned,
so although technically we're doing the same comparison, this is more
in line with POSIX.
The comparison was moved to read_write as that routine is used within
VFS to let it VFS write out coredumps.
When a process wants something done from VFS, but VFS has no worker
threads available, the request is stored and executed later. However,
when PM also sends a request for that process at the same time, discard
the pending request from the process and give priority to PM. The request
PM sends is either an EXIT or a DUMPCORE request, so we're not interested
in executing the pending request anyway.
This patch provides basic protection against damage resulting from
differently compiled servers blindly copying tables to one another.
In every getsysinfo() call, the caller is provided with the expected
size of the requested data structure. The callee fails the call if
the expected size does not match the data structure's actual size.
Using sendrec directly only results in problems. While it is not
clear whether using fs_sendrec is the best option, it is at least
an improvement.
Also remove some legacy cruft.
This patch separates the character and block driver communication
protocols. The old character protocol remains the same, but a new
block protocol is introduced. The libdriver library is replaced by
two new libraries: libchardriver and libblockdriver. Their exposed
API, and drivers that use them, have been updated accordingly.
Together, libbdev and libblockdriver now completely abstract away
the message format used by the block protocol. As the memory driver
is both a character and a block device driver, it now implements its
own message loop.
The most important semantic change made to the block protocol is that
it is no longer possible to return both partial results and an error
for a single transfer. This simplifies the interaction between the
caller and the driver, as the I/O vector no longer needs to be copied
back. Also, drivers are now no longer supposed to decide based on the
layout of the I/O vector when a transfer should be cut short. Put
simply, transfers are now supposed to either succeed completely, or
result in an error.
After this patch, the state of the various pieces is as follows:
- block protocol: stable
- libbdev API: stable for synchronous communication
- libblockdriver API: needs slight revision (the drvlib/partition API
in particular; the threading API will also change shortly)
- character protocol: needs cleanup
- libchardriver API: needs cleanup accordingly
- driver restarts: largely unsupported until endpoint changes are
reintroduced
As a side effect, this patch eliminates several bugs, hacks, and gcc
-Wall and -W warnings all over the place. It probably introduces a
few new ones, too.
Update warning: this patch changes the protocol between MFS and disk
drivers, so in order to use old/new images, the MFS from the ramdisk
must be used to mount all file systems.
PUFFS file systems need to make back calls for every operation we
send to them. Consequently, they cannot handle block reads and writes
themselves. Instead, the root file system has to do it (for now).
When the mount operation causes an FS to make a back call, AVFS now
concludes that every block read and write for that FS has to go
through the root file system.
In some places it was assumed that PATH_MAX does not include a
terminating null character.
Increases PATH_MAX to 1024 to get in sync with NetBSD. Required some
rewriting in AVFS to keep memory usage low (the stack in use by a thread
is very small).
During shutdown all processes are semi-exited and FSes are unmounted.
This semi-exit causes trouble for FUSE mounts as they still need access
to file descriptors and working directory in order to unmount.
