Removed dead code that assumed that M_TRYWAIT can return NULL; it's not true
since the advent of MBUMA.
Reviewed by: arch
There are ongoing disputes as to whether we want to switch to directly using
UMA flags M_WAITOK/M_NOWAIT for mbuf(9) allocation.
the provided trailers. This has been broken since revision 1.240.
Submitted by: Dan Nelson
PR: kern/120948
"sounds ok to me" from: phk
MFC after: 3 days
consists of the null-terminated name and the contents of any structure
you wish to record. A new ktrstruct() function constructs and emits a
KTR_STRUCT record. It is accompanied by convenience macros for struct
stat and struct sockaddr.
In kdump(1), KTR_STRUCT records are handled by a dispatcher function
that runs stringent sanity checks on its contents before handing it
over to individual decoding funtions for each type of structure.
Currently supported structures are struct stat and struct sockaddr for
the AF_INET, AF_INET6 and AF_UNIX families; support for AF_APPLETALK
and AF_IPX is present but disabled, as I am unable to test it properly.
Since 's' was already taken, the letter 't' is used by ktrace(1) to
enable KTR_STRUCT trace points, and in kdump(1) to enable their
decoding.
Derived from patches by Andrew Li <andrew2.li@citi.com>.
PR: kern/117836
MFC after: 3 weeks
referencing the files VM pages are returned from the network stack,
making changes to the file safe.
This flag does not guarantee that the data has been transmitted to the
other end.
free function controlable, instead of passing the KVA of the buffer
storage as the first argument.
Fix all conventional users of the API to pass the KVA of the buffer
as the first argument, to make this a no-op commit.
Likely break the only non-convetional user of the API, after informing
the relevant committer.
Update the mbuf(9) manual page, which was already out of sync on
this point.
Bump __FreeBSD_version to 800016 as there is no way to tell how
many arguments a CPP macro needs any other way.
This paves the way for giving sendfile(9) a way to wait for the
passed storage to have been accessed before returning.
This does not affect the memory layout or size of mbufs.
Parental oversight by: sam and rwatson.
No MFC is anticipated.
read socket buffers in shutdown() and close():
- Call socantrcvmore() before sblock() to dislodge any threads that
might be sleeping (potentially indefinitely) while holding sblock(),
such as a thread blocked in recv().
- Flag the sblock() call as non-interruptible so that a signal
delivered to the thread calling sorflush() doesn't cause sblock() to
fail. The sblock() is required to ensure that all other socket
consumer threads have, in fact, left, and do not enter, the socket
buffer until we're done flushin it.
To implement the latter, change the 'flags' argument to sblock() to
accept two flags, SBL_WAIT and SBL_NOINTR, rather than one M_WAITOK
flag. When SBL_NOINTR is set, it forces a non-interruptible sx
acquisition, regardless of the setting of the disposition of SB_NOINTR
on the socket buffer; without this change it would be possible for
another thread to clear SB_NOINTR between when the socket buffer mutex
is released and sblock() is invoked.
Reviewed by: bz, kmacy
Reported by: Jos Backus <jos at catnook dot com>
conjuction with 'thread' argument passing which is always curthread.
Remove the unuseful extra-argument and pass explicitly curthread to lower
layer functions, when necessary.
KPI results broken by this change, which should affect several ports, so
version bumping and manpage update will be further committed.
Tested by: kris, pho, Diego Sardina <siarodx at gmail dot com>
Remove this argument and pass curthread directly to underlying
VOP_LOCK1() VFS method. This modify makes the code cleaner and in
particular remove an annoying dependence helping next lockmgr() cleanup.
KPI results, obviously, changed.
Manpage and FreeBSD_version will be updated through further commits.
As a side note, would be valuable to say that next commits will address
a similar cleanup about VFS methods, in particular vop_lock1 and
vop_unlock.
Tested by: Diego Sardina <siarodx at gmail dot com>,
Andrea Di Pasquale <whyx dot it at gmail dot com>
- Introduce a finit() which is used to initailize the fields of struct file
in such a way that the ops vector is only valid after the data, type,
and flags are valid.
- Protect f_flag and f_count with atomic operations.
- Remove the global list of all files and associated accounting.
- Rewrite the unp garbage collection such that it no longer requires
the global list of all files and instead uses a list of all unp sockets.
- Mark sockets in the accept queue so we don't incorrectly gc them.
Tested by: kris, pho
from Mac OS X Leopard--rationalize naming for entry points to
the following general forms:
mac_<object>_<method/action>
mac_<object>_check_<method/action>
The previous naming scheme was inconsistent and mostly
reversed from the new scheme. Also, make object types more
consistent and remove spaces from object types that contain
multiple parts ("posix_sem" -> "posixsem") to make mechanical
parsing easier. Introduce a new "netinet" object type for
certain IPv4/IPv6-related methods. Also simplify, slightly,
some entry point names.
All MAC policy modules will need to be recompiled, and modules
not updates as part of this commit will need to be modified to
conform to the new KPI.
Sponsored by: SPARTA (original patches against Mac OS X)
Obtained from: TrustedBSD Project, Apple Computer
the last message on the send stream was "null" but still
there, a state we allow, we could get hung and not clean
it up and wait for the shutdown guard timer to clear the
association without a graceful close. Fix this so that
that we properly clean up.
- Added support for Multiple ASCONF per new RFC. We only
(so far) accept input of these and cannot yet generate
a multi-asconf.
- Sysctl'd support for experimental Fast Handover feature. Always
disabled unless sysctl or socket option changes to enable.
- Error case in add-ip where the peer supports AUTH and ADD-IP
but does NOT require AUTH of ASCONF/ASCONF-ACK. We need to
ABORT in this case.
- According to the Kyoto summit of socket api developers
(Solaris, Linux, BSD). We need to have:
o non-eeor mode messages be atomic - Fixed
o Allow implicit setup of an assoc in 1-2-1 model if
using the sctp_**() send calls - Fixed
o Get rid of HAVE_XXX declarations - Done
o add a sctp_pr_policy in hole in sndrcvinfo structure - Done
o add a PR_SCTP_POLICY_VALID type flag - yet to-do in a future patch!
- Optimize sctp6 calls to reuse code in sctp_usrreq. Also optimize
when we close sending out the data and disabling Nagle.
- Change key concatenation order to match the auth RFC
- When sending OOTB shutdown_complete always do csum.
- Don't send PKT-DROP to a PKT-DROP
- For abort chunks just always checksums same for
shutdown-complete.
- inpcb_free front state had a bug where in queue
data could wedge an assoc. We need to just abandon
ones in front states (free_assoc).
- If a peer sends us a 64k abort, we would try to
assemble a response packet which may be larger than
64k. This then would be dropped by IP. Instead make
a "minimum" size for us 64k-2k (we want at least
2k for our initack). If we receive such an init
discard it early without all the processing.
- When we peel off we must increment the tcb ref count
to keep it from being freed from underneath us.
- handling fwd-tsn had bugs that caused memory overwrites
when given faulty data, fixed so can't happen and we
also stop at the first bad stream no.
- Fixed so comm-up generates the adaption indication.
- peeloff did not get the hmac params copied.
- fix it so we lock the addr list when doing src-addr selection
(in future we need to use a multi-reader/one writer lock here)
- During lowlevel output, we could end up with a _l_addr set
to null if the iterator is calling the output routine. This
means we would possibly crash when we gather the MTU info.
Fix so we only do the gather where we have a src address
cached.
- we need to be sure to set abort flag on conn state when
we receive an abort.
- peeloff could leak a socket. Moved code so the close will
find the socket if the peeloff fails (uipc_syscalls.c)
Approved by: re@freebsd.org(Ken Smith)
previously conditionally acquired Giant based on debug.mpsafenet. As that
has now been removed, they are no longer required. Removing them
significantly simplifies error-handling in the socket layer, eliminated
quite a bit of unwinding of locking in error cases.
While here clean up the now unneeded opt_net.h, which previously was used
for the NET_WITH_GIANT kernel option. Clean up some related gotos for
consistency.
Reviewed by: bz, csjp
Tested by: kris
Approved by: re (kensmith)
ignore the size of any headers that were passed with the sendfile(2)
system call. Otherwise the file sent will be truncated by the header
size if the nbytes parameter was provided. The bug doesn't show up
when either nbytes is zero, meaning send the whole file, or no header
iovec is provided.
Resolve a potential error aliasing of errors from the VM and sf_buf
parts and the protocol send parts where an error of the latter over-
writes one of the former.
Update comments.
The byte accounting bug wasn't seen in earlier because none of the popular
sendfile(2) consumers, Apache, lighttpd and our ftpd(8) use it in modes
that trigger it. The varnish HTTP proxy makes full use of it and exposed
the problem.
Bug found by: phk
Tested by: phk
on each socket buffer with the socket buffer's mutex. This sleep lock is
used to serialize I/O on sockets in order to prevent I/O interlacing.
This change replaces the custom sleep lock with an sx(9) lock, which
results in marginally better performance, better handling of contention
during simultaneous socket I/O across multiple threads, and a cleaner
separation between the different layers of locking in socket buffers.
Specifically, the socket buffer mutex is now solely responsible for
serializing simultaneous operation on the socket buffer data structure,
and not for I/O serialization.
While here, fix two historic bugs:
(1) a bug allowing I/O to be occasionally interlaced during long I/O
operations (discovere by Isilon).
(2) a bug in which failed non-blocking acquisition of the socket buffer
I/O serialization lock might be ignored (discovered by sam).
SCTP portion of this patch submitted by rrs.
When nbytes=0, sendfile(2) should use file size. Because of the bug, it
was sending half of a file. The bug is that 'off' variable can't be used
for size calculation, because it changes inside the loop, so we should
use uap->offset instead.
and flags with an sxlock. This leads to a significant and measurable
performance improvement as a result of access to shared locking for
frequent lookup operations, reduced general overhead, and reduced overhead
in the event of contention. All of these are imported for threaded
applications where simultaneous access to a shared file descriptor array
occurs frequently. Kris has reported 2x-4x transaction rate improvements
on 8-core MySQL benchmarks; smaller improvements can be expected for many
workloads as a result of reduced overhead.
- Generally eliminate the distinction between "fast" and regular
acquisisition of the filedesc lock; the plan is that they will now all
be fast. Change all locking instances to either shared or exclusive
locks.
- Correct a bug (pointed out by kib) in fdfree() where previously msleep()
was called without the mutex held; sx_sleep() is now always called with
the sxlock held exclusively.
- Universally hold the struct file lock over changes to struct file,
rather than the filedesc lock or no lock. Always update the f_ops
field last. A further memory barrier is required here in the future
(discussed with jhb).
- Improve locking and reference management in linux_at(), which fails to
properly acquire vnode references before using vnode pointers. Annotate
improper use of vn_fullpath(), which will be replaced at a future date.
In fcntl(), we conservatively acquire an exclusive lock, even though in
some cases a shared lock may be sufficient, which should be revisited.
The dropping of the filedesc lock in fdgrowtable() is no longer required
as the sxlock can be held over the sleep operation; we should consider
removing that (pointed out by attilio).
Tested by: kris
Discussed with: jhb, kris, attilio, jeff
- Close the new file objects created during socketpair() if the copyout of
the new file descriptors fails.
- Add a test to the socketpair regression test for this edge case.
- Remove also "MP SAFE" after prior "MPSAFE" pass. (suggested by bde)
- Remove extra blank lines in some cases.
- Add extra blank lines in some cases.
- Remove no-op comments consisting solely of the function name, the word
"syscall", or the system call name.
- Add punctuation.
- Re-wrap some comments.
system calls now enter without Giant held, and then in some cases, acquire
Giant explicitly.
Remove a number of other MPSAFE annotations in the credential code and
tweak one or two other adjacent comments.
written to the socket). The rewrite in revision 1.240 got confused by the
FreeBSD 4.x bug compatibility code.
For some reason lighttpd, that was used for testing the new sendfile code,
was not affected by the problem but apache and others using headers/trailers
in the sendfile call received incorrect sbytes values after return from non-
blocking sockets. This then lead to restarts with wrong offsets and thus
mixed up file contents when the socket was writeable again. All programs
not using headers/trailers, like ftpd, were not affected by the bug.
Reported by: Pawel Worach <pawel.worach-at-gmail.com>
Tested by: Pawel Worach <pawel.worach-at-gmail.com>
label after the sbunlock() part.
This correctly handles calls to sendfile(2) without valid parameters
that was broken in rev. 1.240.
Coverity error: 272162
work is not just mine, but it is also the works of Peter Lei
and Michael Tuexen. They both are my two key other developers
working on the project.. and they need ata-boy's too:
****
peterlei@cisco.comtuexen@fh-muenster.de
****
I did do a make sysent which updated the
syscall's and sysproto.. I hope that is correct... without
it you don't build since we have new syscalls for SCTP :-0
So go out and look at the NOTES, add
option SCTP (make sure inet and inet6 are present too)
and play with SCTP.
I will see about comitting some test tools I have after I
figure out where I should place them. I also have a
lib (libsctp.a) that adds some of the missing socketapi
functions that I need to put into lib's.. I will talk
to George about this :-)
There may still be some 64 bit issues in here, none of
us have a 64 bit processor to test with yet.. Michael
may have a MAC but thats another beast too..
If you have a mac and want to use SCTP contact Michael
he maintains a web site with a loadable module with
this code :-)
Reviewed by: gnn
Approved by: gnn
mbuf clusters. Add a flags parameter to accept M_PKTHDR and M_EOR mbuf
chain flags. Provide compatibility macro for m_getm() calling m_getm2()
with M_PKTHDR set.
Rewrite m_uiotombuf() to use m_getm2() for mbuf allocation and do the
uiomove() in a tight loop over the mbuf chain. Add a flags parameter to
accept mbuf flags to be passed to m_getm2(). Adjust all callers for the
extra parameter.
Sponsored by: TCP/IP Optimization Fundraise 2005
MFC after: 3 month
VM pages into mbufs as it can -- up to the free send socket buffer space.
The outer loop then drops the whole mbuf chain into the send socket buffer,
calls tcp_output() on it and then waits until 50% of the socket buffer are
free again to repeat the cycle. This way tcp_output() gets the full amount
of data to work with and can issue up to 64K sends for TSO to chop up in
the network adapter without using any CPU cycles. Thus it gets very efficient
especially with the readahead the VM and I/O system do.
The previous sendfile(2) code simply looped over the file, turned each 4K
page into an mbuf and sent it off. This had the effect that TSO could only
generate 2 packets per send instead of up to 44 at its maximum of 64K.
Add experimental SF_MNOWAIT flag to sendfile(2) to return ENOMEM instead of
sleeping on mbuf allocation failures.
Benchmarking shows significant improvements (95% confidence):
45% less cpu (or 1.81 times better) with new sendfile vs. old sendfile (non-TSO)
83% less cpu (or 5.7 times better) with new sendfile vs. old sendfile (TSO)
(Sender AMD Opteron 852 (2.6GHz) with em(4) PCI-X-133 interface and receiver
DELL Poweredge SC1425 P-IV Xeon 3.2GHz with em(4) LOM connected back to back
at 1000Base-TX full duplex.)
Sponsored by: TCP/IP Optimization Fundraise 2005
MFC after: 3 month
begun with a repo-copy of mac.h to mac_framework.h. sys/mac.h now
contains the userspace and user<->kernel API and definitions, with all
in-kernel interfaces moved to mac_framework.h, which is now included
across most of the kernel instead.
This change is the first step in a larger cleanup and sweep of MAC
Framework interfaces in the kernel, and will not be MFC'd.
Obtained from: TrustedBSD Project
Sponsored by: SPARTA
synchronized by the lock on the object containing the page.
Transition PG_WANTED and PG_SWAPINPROG to use the new field,
eliminating the need for holding the page queues lock when setting
or clearing these flags. Rename PG_WANTED and PG_SWAPINPROG to
VPO_WANTED and VPO_SWAPINPROG, respectively.
Eliminate the assertion that the page queues lock is held in
vm_page_io_finish().
Eliminate the acquisition and release of the page queues lock
around calls to vm_page_io_finish() in kern_sendfile() and
vfs_unbusy_pages().
kern_accept() and accept1(). If another thread closed the new file
descriptor and the first thread later got an error trying to copyout the
socket address, then it would attempt to close the wrong file object. To
fix, add a struct file ** argument to kern_accept(). If it is non-NULL,
then on success kern_accept() will store a pointer to the new file object
there and not release any of the references. It is up to the calling code
to drop the references appropriately (including a call to fdclose() in case
of error to safely handle the aforementioned race). While I'm at it, go
ahead and fix the svr4 streams code to not leak the accept fd if it gets an
error trying to copyout the streams structures.
soreceive(), and sopoll(), which are wrappers for pru_sosend,
pru_soreceive, and pru_sopoll, and are now used univerally by socket
consumers rather than either directly invoking the old so*() functions
or directly invoking the protocol switch method (about an even split
prior to this commit).
This completes an architectural change that was begun in 1996 to permit
protocols to provide substitute implementations, as now used by UDP.
Consumers now uniformly invoke sosend(), soreceive(), and sopoll() to
perform these operations on sockets -- in particular, distributed file
systems and socket system calls.
Architectural head nod: sam, gnn, wollman
use by ABI emulators.
- Alter the interface of kern_recvit() somewhat. Specifically, go ahead
and hard code UIO_USERSPACE in the uio as that's what all the callers
specify. In place, add a new uioseg to indicate what type of pointer
is in mp->msg_name. Previously it was always a userland address, but
ABI emulators may pass in kernel-side sockaddrs. Also, remove the
namelenp field and instead require the two places that used it to
explicitly copy mp->msg_namelen out to userland.
- Use the patched kern_recvit() to replace svr4_recvit() and the stock
kern_sendit() to replace svr4_sendit().
- Use kern_bind() instead of stackgap use in ti_bind().
- Use kern_getpeername() and kern_getsockname() instead of stackgap in
svr4_stream_ti_ioctl().
- Use kern_connect() instead of stackgap in svr4_do_putmsg().
- Use kern_getpeername() and kern_accept() instead of stackgap in
svr4_do_getmsg().
- Retire the stackgap from SVR4 compat as it is no longer used.
in setsockopt so that they can be compared correctly against negative
values. Passing in a negative value had a rather negative effect
on our socket code, making it impossible to open new sockets.
PR: 98858
Submitted by: James.Juran@baesystems.com
MFC after: 1 week
- Move sonewconn(), which creates new sockets for incoming connections on
listen sockets, so that all socket allocate code is together in
uipc_socket.c.
- Move 'maxsockets' and associated sysctls to uipc_socket.c with the
socket allocation code.
- Move kern.ipc sysctl node to uipc_socket.c, add a SYSCTL_DECL() for it
to sysctl.h and remove lots of scattered implementations in various
IPC modules.
- Sort sodealloc() after soalloc() in uipc_socket.c for dependency order
reasons. Statisticize soalloc() and sodealloc() as they are now
required only in uipc_socket.c, and are internal to the socket
implementation.
After this change, socket allocation and deallocation is entirely
centralized in one file, and uipc_socket2.c consists entirely of socket
buffer manipulation and default protocol switch functions.
MFC after: 1 month
sendfile(). This causes sendfile() to use the file descriptor
reference to the socket instead of bumping the socket reference
count, which avoids an additional refcount operation, as well as a
potential expensive socket refcount drop, which can lead to
contention on the accept mutex. This change also has the side
effect of further reducing the number of cases where an in-progress
I/O operation can occur on a socket after close, as using the file
descriptor refcount prevents the socket from closing while in use.
MFC after: 3 months
file lock, in the style of fgetsock().
Modify accept1() to use getsock() instead of fgetsock(), relying on the
file descriptor reference rather than an acquired socket reference to
prevent the listen socket from being destroyed during accept(). This
avoids additional reference count operations, which should improve
performance, and also avoids accept1() operating on a socket whose file
descriptor has been torn down, which may have resulted in protocol
shutdown starting.
MFC after: 3 months
acquiring Giant in kern_sendfile().
Guard against the forced reclamation of a vnode in kern_sendfile().
Discussed with: jeff
Reviewed by: tegge
MFC after: 3 weeks
which is invoked from socket() and socketpair(), permitting MAC
policy modules to control the creation of sockets by domain, type, and
protocol.
Obtained from: TrustedBSD Project
Sponsored by: SPARTA, SPAWAR
Approved by: re (scottl)
Requested by: SCC
to the mbuf. Offset cannot exceed MHLEN bytes. This is currently used to
fix Ethernet header alignment problem on alpha and sparc64. Also change all
users of m_uiotombuf to pass proper offset.
Reviewed by: jmg, sam
Tested by: Sten Spans "sten AT blinkenlights DOT nl"
MFC after: 1 week
control socket poll() (select()), fstat(), and accept() operations,
required for some policies:
poll() mac_check_socket_poll()
fstat() mac_check_socket_stat()
accept() mac_check_socket_accept()
Update mac_stub and mac_test policies to be aware of these entry points.
While here, add missing entry point implementations for:
mac_stub.c stub_check_socket_receive()
mac_stub.c stub_check_socket_send()
mac_test.c mac_test_check_socket_send()
mac_test.c mac_test_check_socket_visible()
Obtained from: TrustedBSD Project
Sponsored by: SPAWAR, SPARTA
SIGPIPE signal for the duration of the sento-family syscalls. Use it to
replace previously added hack in Linux layer based on temporarily setting
SO_NOSIGPIPE flag.
Suggested by: alfred
soref() to also covering the update of so_state. While no other user
threads can update the socket state here as it's not yet hooked up to
the file descriptor array yet, the protocol could also frob the
socket state here, leading to a lost update to the so_state field.
No reported instances of this bug (as yet).
MFC after: 3 days
Use this in all the places where sleeping with the lock held is not
an issue.
The distinction will become significant once we finalize the exact
lock-type to use for this kind of case.
Change the spelling of the "catch" option to be consistent with the new
options. Implement the "no wait" option. An implementation of the "CPU
private" for i386 will be committed at a later date.
need for most calls to vm_page_busy(). Specifically, most calls to
vm_page_busy() occur immediately prior to a call to vm_page_remove().
In such cases, the containing vm object is locked across both calls.
Consequently, the setting of the vm page's PG_BUSY flag is not even
visible to other threads that are following the synchronization
protocol.
This change (1) eliminates the calls to vm_page_busy() that
immediately precede a call to vm_page_remove() or functions, such as
vm_page_free() and vm_page_rename(), that call it and (2) relaxes the
requirement in vm_page_remove() that the vm page's PG_BUSY flag is
set. Now, the vm page's PG_BUSY flag is set only when the vm object
lock is released while the vm page is still in transition. Typically,
this is when it is undergoing I/O.
count to prevent sockets from being garbage collected during
socket-specific system calls. This is the same approach used in
most VFS-specific system calls, as well as generic file descriptor
system calls such as read() and write().
To do this, add a utility function getsock(), which is logically
identical to getvnode() used for the same purpose in VFS. Unlike
fgetsock(), it returns with the file reference count elevated, but
no bump of the socket reference count. Replace matching calls to
fputsock() with fdrop().
This change is made to all socket system calls other than
sendfile() and accept(), but the approach should be applicable to
those system calls also.
This shaves about four mutex operations off of each of these
system calls, including send() and recv() variants, adding about
1% to pps on minimal UDP packets for UP using netblast, and 4% on
SMP.
Reviewed by: pjd
is locked when vm_page_io_finish() is called on a page. This is to satisfy
a new, post-RELENG_5 assertion in vm_page_io_finish(). (I am in the
process of transitioning the responsibility for synchronizing access to
various fields/flags on the page from the global page queues lock to the
per-object lock.)
Tripped over by: obrien@
a more complete subsystem, and removes the knowlege of how things are
implemented from the drivers. Include locking around filter ops, so a
module like aio will know when not to be unloaded if there are outstanding
knotes using it's filter ops.
Currently, it uses the MTX_DUPOK even though it is not always safe to
aquire duplicate locks. Witness currently doesn't support the ability
to discover if a dup lock is ok (in some cases).
Reviewed by: green, rwatson (both earlier versions)
values from either user land or from the kernel. Use them for
[gs]etsockopt and to clean up some calls to [gs]etsockopt in the
Linux emulation code that uses the stackgap.
Add copyiniov() which copies a struct iovec array in from userland into
a malloc'ed struct iovec. Caller frees.
Change uiofromiov() to malloc the uio (caller frees) and name it
copyinuio() which is more appropriate.
Add cloneuio() which returns a malloc'ed copy. Caller frees.
Use them throughout.
the SS_NBIO flag from the parent socket to the child socket during an
accept() operation.
The file descriptor O_NONBLOCK flag would have been propagated already
by the fflag assignment, and therefore would have been inconsistent
with the underlying socket's so_state member.
This makes accept() more closely adhere to the API contract we effectively
outline in the manual page. Note also that Linux continues to differ here;
O_NONBLOCK is not propagated. The other BSDs do propagate the flag, as
does Solaris. The Single UNIX Specification does not offer specific
advice on this issue.
PR: kern/45733
Requested by: Jayanth Vijayaraghavan
Reviewed by: rwatson
lock state. Convert tsleep() into msleep() with socket buffer mutex
as argument. Hold socket buffer lock over sbunlock() to protect sleep
lock state.
Assert socket buffer lock in sbwait() to protect the socket buffer
wait state. Convert tsleep() into msleep() with socket buffer mutex
as argument.
Modify sofree(), sosend(), and soreceive() to acquire SOCKBUF_LOCK()
in order to call into these functions with the lock, as well as to
start protecting other socket buffer use in their implementation. Drop
the socket buffer mutexes around calls into the protocol layer, around
potentially blocking operations, for copying to/from user space, and
VM operations relating to zero-copy. Assert the socket buffer mutex
strategically after code sections or at the beginning of loops. In
some cases, modify return code to ensure locks are properly dropped.
Convert the potentially blocking allocation of storage for the remote
address in soreceive() into a non-blocking allocation; we may wish to
move the allocation earlier so that it can block prior to acquisition
of the socket buffer lock.
Drop some spl use.
NOTE: Some races exist in the current structuring of sosend() and
soreceive(). This commit only merges basic socket locking in this
code; follow-up commits will close additional races. As merged,
these changes are not sufficient to run without Giant safely.
Reviewed by: juli, tjr
flags relating to several aspects of socket functionality. This change
breaks out several bits relating to send and receive operation into a
new per-socket buffer field, sb_state, in order to facilitate locking.
This is required because, in order to provide more granular locking of
sockets, different state fields have different locking properties. The
following fields are moved to sb_state:
SS_CANTRCVMORE (so_state)
SS_CANTSENDMORE (so_state)
SS_RCVATMARK (so_state)
Rename respectively to:
SBS_CANTRCVMORE (so_rcv.sb_state)
SBS_CANTSENDMORE (so_snd.sb_state)
SBS_RCVATMARK (so_rcv.sb_state)
This facilitates locking by isolating fields to be located with other
identically locked fields, and permits greater granularity in socket
locking by avoiding storing fields with different locking semantics in
the same short (avoiding locking conflicts). In the future, we may
wish to coallesce sb_state and sb_flags; for the time being I leave
them separate and there is no additional memory overhead due to the
packing/alignment of shorts in the socket buffer structure.
SOCK_LOCK(so):
- Hold socket lock over calls to MAC entry points reading or
manipulating socket labels.
- Assert socket lock in MAC entry point implementations.
- When externalizing the socket label, first make a thread-local
copy while holding the socket lock, then release the socket lock
to externalize to userspace.
reference count:
- Assert SOCK_LOCK(so) macros that directly manipulate so_count:
soref(), sorele().
- Assert SOCK_LOCK(so) in macros/functions that rely on the state of
so_count: sofree(), sotryfree().
- Acquire SOCK_LOCK(so) before calling these functions or macros in
various contexts in the stack, both at the socket and protocol
layers.
- In some cases, perform soisdisconnected() before sotryfree(), as
this could result in frobbing of a non-present socket if
sotryfree() actually frees the socket.
- Note that sofree()/sotryfree() will release the socket lock even if
they don't free the socket.
Submitted by: sam
Sponsored by: FreeBSD Foundation
Obtained from: BSD/OS
when I reordered events in accept1() to allocate a file descriptor
earlier, I didn't properly update use of goto on exit to unwind for
cases where the file descriptor is now held, but wasn't previously.
The result was that, in the event of accept() on a non-blocking socket,
or in the event of a socket error, a file descriptor would be leaked.
This ended up being non-fatal in many cases, as the file descriptor
would be properly GC'd on process exit, so only showed up for processes
that do a lot of non-blocking accept() calls, and also live for a long
time (such as qmail).
This change updates the use of goto targets to do additional unwinding.
Eyes provided by: Brian Feldman <green@freebsd.org>
Feet, hands provided by: Stefan Ehmann <shoesoft@gmx.net>,
Dimitry Andric <dimitry@andric.com>
Arjan van Leeuwen <avleeuwen@piwebs.com>
global mutex, accept_mtx, which serializes access to the following
fields across all sockets:
so_qlen so_incqlen so_qstate
so_comp so_incomp so_list
so_head
While providing only coarse granularity, this approach avoids lock
order issues between sockets by avoiding ownership of the fields
by a specific socket and its per-socket mutexes.
While here, rewrite soclose(), sofree(), soaccept(), and
sonewconn() to add assertions, close additional races and address
lock order concerns. In particular:
- Reorganize the optimistic concurrency behavior in accept1() to
always allocate a file descriptor with falloc() so that if we do
find a socket, we don't have to encounter the "Oh, there wasn't
a socket" race that can occur if falloc() sleeps in the current
code, which broke inbound accept() ordering, not to mention
requiring backing out socket state changes in a way that raced
with the protocol level. We may want to add a lockless read of
the queue state if polling of empty queues proves to be important
to optimize.
- In accept1(), soref() the socket while holding the accept lock
so that the socket cannot be free'd in a race with the protocol
layer. Likewise in netgraph equivilents of the accept1() code.
- In sonewconn(), loop waiting for the queue to be small enough to
insert our new socket once we've committed to inserting it, or
races can occur that cause the incomplete socket queue to
overfill. In the previously implementation, it was sufficient
to simply tested once since calling soabort() didn't release
synchronization permitting another thread to insert a socket as
we discard a previous one.
- In soclose()/sofree()/et al, it is the responsibility of the
caller to remove a socket from the incomplete connection queue
before calling soabort(), which prevents soabort() from having
to walk into the accept socket to release the socket from its
queue, and avoids races when releasing the accept mutex to enter
soabort(), permitting soabort() to avoid lock ordering issues
with the caller.
- Generally cluster accept queue related operations together
throughout these functions in order to facilitate locking.
Annotate new locking in socketvar.h.
the socket is on an accept queue of a listen socket. This change
renames the flags to SQ_COMP and SQ_INCOMP, and moves them to a new
state field on the socket, so_qstate, as the locking for these flags
is substantially different for the locking on the remainder of the
flags in so_state.
mbuma is an Mbuf & Cluster allocator built on top of a number of
extensions to the UMA framework, all included herein.
Extensions to UMA worth noting:
- Better layering between slab <-> zone caches; introduce
Keg structure which splits off slab cache away from the
zone structure and allows multiple zones to be stacked
on top of a single Keg (single type of slab cache);
perhaps we should look into defining a subset API on
top of the Keg for special use by malloc(9),
for example.
- UMA_ZONE_REFCNT zones can now be added, and reference
counters automagically allocated for them within the end
of the associated slab structures. uma_find_refcnt()
does a kextract to fetch the slab struct reference from
the underlying page, and lookup the corresponding refcnt.
mbuma things worth noting:
- integrates mbuf & cluster allocations with extended UMA
and provides caches for commonly-allocated items; defines
several zones (two primary, one secondary) and two kegs.
- change up certain code paths that always used to do:
m_get() + m_clget() to instead just use m_getcl() and
try to take advantage of the newly defined secondary
Packet zone.
- netstat(1) and systat(1) quickly hacked up to do basic
stat reporting but additional stats work needs to be
done once some other details within UMA have been taken
care of and it becomes clearer to how stats will work
within the modified framework.
From the user perspective, one implication is that the
NMBCLUSTERS compile-time option is no longer used. The
maximum number of clusters is still capped off according
to maxusers, but it can be made unlimited by setting
the kern.ipc.nmbclusters boot-time tunable to zero.
Work should be done to write an appropriate sysctl
handler allowing dynamic tuning of kern.ipc.nmbclusters
at runtime.
Additional things worth noting/known issues (READ):
- One report of 'ips' (ServeRAID) driver acting really
slow in conjunction with mbuma. Need more data.
Latest report is that ips is equally sucking with
and without mbuma.
- Giant leak in NFS code sometimes occurs, can't
reproduce but currently analyzing; brueffer is
able to reproduce but THIS IS NOT an mbuma-specific
problem and currently occurs even WITHOUT mbuma.
- Issues in network locking: there is at least one
code path in the rip code where one or more locks
are acquired and we end up in m_prepend() with
M_WAITOK, which causes WITNESS to whine from within
UMA. Current temporary solution: force all UMA
allocations to be M_NOWAIT from within UMA for now
to avoid deadlocks unless WITNESS is defined and we
can determine with certainty that we're not holding
any locks when we're M_WAITOK.
- I've seen at least one weird socketbuffer empty-but-
mbuf-still-attached panic. I don't believe this
to be related to mbuma but please keep your eyes
open, turn on debugging, and capture crash dumps.
This change removes more code than it adds.
A paper is available detailing the change and considering
various performance issues, it was presented at BSDCan2004:
http://www.unixdaemons.com/~bmilekic/netbuf_bmilekic.pdf
Please read the paper for Future Work and implementation
details, as well as credits.
Testing and Debugging:
rwatson,
brueffer,
Ketrien I. Saihr-Kesenchedra,
...
Reviewed by: Lots of people (for different parts)
allocation and deallocation. This flag's principal use is shortly after
allocation. For such cases, clearing the flag is pointless. The only
unusual use of PG_ZERO is in vfs_bio_clrbuf(). However, allocbuf() never
requests a prezeroed page. So, vfs_bio_clrbuf() never sees a prezeroed
page.
Reviewed by: tegge@
side effect of that change caused headers to not be sent if a 0 byte
file was passed to sendfile. This change fixes that behavior, allowing
sendfile to send out the headers even with a 0 byte file again.
Noticed by: Dirk Engling
dependent function by the same name and a machine-independent function,
sf_buf_mext(). Aside from the virtue of making more of the code machine-
independent, this change also makes the interface more logical. Before,
sf_buf_free() did more than simply undo an sf_buf_alloc(); it also
unwired and if necessary freed the page. That is now the purpose of
sf_buf_mext(). Thus, sf_buf_alloc() and sf_buf_free() can now be used
as a general-purpose emphemeral map cache.
mindful of blocking on disk I/O and instead return EBUSY when such
blocking would occur.
Results from the DeBox project indicate that blocking on disk I/O
can slow the performance of a kqueue/poll based webserver. Using
a flag such as SF_NODISKIO and throwing connections that would block
to helper processes/threads helped increase performance.
Currently, only the Flash webserver uses this flag, although it could
probably be applied to thttpd with relative ease.
Idea by: Yaoping Ruan & Vivek Pai
packet along with data, instead of in their own packet. When serving files
of size (packetsize - headersize) or smaller, this will result in one less
packet crossing the network. Quick testing with thttpd and http_load has
shown a noticeable performance improvement in this case (350 vs 330 fetches
per second.)
Included in this commit are two support routines, iov_to_uio, and m_uiotombuf;
these routines are used by sendfile to construct the header mbuf chain that
will be linked to the rest of the data in the socket buffer.
