First, update the return types of aio_return() and aio_waitcomplete() to
ssize_t.
POSIX requires aio_return() to return a ssize_t so that it can represent
all return values from read() and write(). aio_waitcomplete() should use
ssize_t for the same reason.
aio_return() has used ssize_t in <aio.h> since r31620 but the manpage and
system call entry were not updated. aio_waitcomplete() has always
returned int.
Note that this does not require new system call stubs as this is
effectively only an API change in how the compiler interprets the return
value.
Second, allow aio_nbytes values up to IOSIZE_MAX instead of just INT_MAX.
aio_read/write should now honor the same length limits as normal read/write.
Third, use longs instead of ints in the aio_return() and aio_waitcomplete()
system call functions so that the 64-bit size_t in the in-kernel aiocb
isn't truncated to 32-bits before being copied out to userland or
being returned.
Finally, a simple test has been added to verify the bounds checking on the
maximum read size from a file.
For the !unmap case it may happen that pbuf gets called unreferenced
when vm_fault_quick_hold_pages() fails.
Initialize it so it doesn't cause trouble.
CID: 1352776
Reviewed by: jhb
MFC after: 1 week
On some architectures, u_long isn't large enough for resource definitions.
Particularly, powerpc and arm allow 36-bit (or larger) physical addresses, but
type `long' is only 32-bit. This extends rman's resources to uintmax_t. With
this change, any resource can feasibly be placed anywhere in physical memory
(within the constraints of the driver).
Why uintmax_t and not something machine dependent, or uint64_t? Though it's
possible for uintmax_t to grow, it's highly unlikely it will become 128-bit on
32-bit architectures. 64-bit architectures should have plenty of RAM to absorb
the increase on resource sizes if and when this occurs, and the number of
resources on memory-constrained systems should be sufficiently small as to not
pose a drastic overhead. That being said, uintmax_t was chosen for source
clarity. If it's specified as uint64_t, all printf()-like calls would either
need casts to uintmax_t, or be littered with PRI*64 macros. Casts to uintmax_t
aren't horrible, but it would also bake into the API for
resource_list_print_type() either a hidden assumption that entries get cast to
uintmax_t for printing, or these calls would need the PRI*64 macros. Since
source code is meant to be read more often than written, I chose the clearest
path of simply using uintmax_t.
Tested on a PowerPC p5020-based board, which places all device resources in
0xfxxxxxxxx, and has 8GB RAM.
Regression tested on qemu-system-i386
Regression tested on qemu-system-mips (malta profile)
Tested PAE and devinfo on virtualbox (live CD)
Special thanks to bz for his testing on ARM.
Reviewed By: bz, jhb (previous)
Relnotes: Yes
Sponsored by: Alex Perez/Inertial Computing
Differential Revision: https://reviews.freebsd.org/D4544
This is several year's worth of fail point upgrades done at EMC Isilon. They
are interdependent enough that it makes sense to put a single diff up for them.
Primarily, we added:
- Changing all mainline execution paths to be lockless, which lets us use fail
points in more sleep-sensitive areas, and allows more parallel execution
- A number of additional commands, including 'pause' that lets us do some
interesting deterministic repros of race conditions
- The ability to dump the stacks of all threads sleeping on a fail point
- A number of other API changes to allow marking up the fail point's context in
the code, and firing callbacks before and after execution
- A man page update
Submitted by: Matthew Bryan <matthew.bryan@isilon.com>
Reviewed by: cem (earlier version), jhb, kib, pho
With feedback from: bdrewery
Sponsored by: EMC / Isilon Storage Division
Differential Revision: https://reviews.freebsd.org/D5427
In timer2sbintime(), calculate the second and fractional second portions of
the sbintime separately. When calculating the the fractional second portion,
use a 64bit multiply to prevent excess truncation. This avoids the ~7% error
in the original conversion for ns, and smaller errors of the same type for us
and ms.
PR: 198139
Reviewed by: jhb
MFC after: 1 week
Differential Revision: https://reviews.freebsd.org/D5397
The base system libc is only used to run binaries built on FreeBSD 7.0 and
later. It does not need to include system call wrappers for system calls
only used by FreeBSD binaries built on versions older than 7.0. This was
already true for "COMPAT" system calls, but now wrappers for system calls
used on FreeBSD 4 and 6 are excluded as well.
Reviewed by: kib
Differential Revision: https://reviews.freebsd.org/D5597
192.168.1.1, with share "share". This commit fixes a problem
where "mkdir /net/192.168.1.1/share/meh" would return spurious
error instead of creating the directory if the target filesystem
wasn't mounted yet; subsequent attempts would work correctly.
The failure scenario is kind of complicated to explain, but it all
boils down to calling VOP_MKDIR() for the target filesystem (NFS)
with wrong dvp - the autofs vnode instead of the filesystem root
mounted over it.
Reviewed by: kib@
MFC after: 1 month
Sponsored by: The FreeBSD Foundation
Differential Revision: https://reviews.freebsd.org/D5442
- Mark AIO system calls as STD and remove the helpers to dynamically
register them.
- Use COMPAT6 for the old system calls with the older sigevent instead of
an 'o' prefix.
- Simplify the POSIX configuration to note that AIO is always available.
- Handle AIO in the default VOP_PATHCONF instead of special casing it in
the pathconf() system call. fpathconf() is still hackish.
- Remove freebsd32_aio_cancel() as it just called the native one directly.
Reviewed by: kib
Sponsored by: Chelsio Communications
Differential Revision: https://reviews.freebsd.org/D5589
files format into printfs and errors to caller. Some leaks of
resources are there, but the same leaks are present in other error
pathes. With the change, the kernel at least boots even when module
with unexpected or corrupted ELF structure is preloaded.
Sponsored by: The FreeBSD Foundation
MFC after: 2 weeks
as SHT_PROGBITS. This is needed after the clang 3.8 import, which
generates that type for .eh_frame section, which had SHT_PROGBITS type
before.
Reported by: Nikolai Lifanov <lifanov@mail.lifanov.com>
PR: 207729
Tested by: dim (previous version)
Sponsored by: The FreeBSD Foundation
MFC after: 2 weeks
Summary:
The idea behind this is '~0ul' is well-defined, and casting to uintmax_t, on a
32-bit platform, will leave the upper 32 bits as 0. The maximum range of a
resource is 0xFFF.... (all bits of the full type set). By dropping the 'ul'
suffix, C type promotion rules apply, and the sign extension of ~0 on 32 bit
platforms gets it to a type-independent 'unsigned max'.
Reviewed By: cem
Sponsored by: Alex Perez/Inertial Computing
Differential Revision: https://reviews.freebsd.org/D5255
but next invocation is cancelled while migrating,
sleepq_check_timeout() needs to be informed that the callout is
stopped. Otherwise the thread switches off CPU and never become
runnable, since running callout could have already raced with us,
while the migrating and cancelled callout could be one which is
expected to set TDP_TIMOFAIL flag for us. This contradicts with the
expected behaviour of callout_stop() for other callers, which
e.g. decrement references from the callout callbacks.
Add a new flag CS_MIGRBLOCK requesting report of the situation as
'successfully stopped'.
Reviewed by: jhb (previous version)
Tested by: cognet, pho
PR: 200992
Sponsored by: The FreeBSD Foundation
MFC after: 2 weeks
Differential revision: https://reviews.freebsd.org/D5221
EXT_MOD_TYPE, EXT_DISPOSABLE types we should first execute the free
callback, then free the mbuf, otherwise we will derefernce memory that
was just freed.
Reported and tested: jhibbits
improve cancellation robustness.
Introduce a new file operation, fo_aio_queue, which is responsible for
queueing and completing an asynchronous I/O request for a given file.
The AIO subystem now exports library of routines to manipulate AIO
requests as well as the ability to run a handler function in the
"default" pool of AIO daemons to service a request.
A default implementation for file types which do not include an
fo_aio_queue method queues requests to the "default" pool invoking the
fo_read or fo_write methods as before.
The AIO subsystem permits file types to install a private "cancel"
routine when a request is queued to permit safe dequeueing and cleanup
of cancelled requests.
Sockets now use their own pool of AIO daemons and service per-socket
requests in FIFO order. Socket requests will not block indefinitely
permitting timely cancellation of all requests.
Due to the now-tight coupling of the AIO subsystem with file types,
the AIO subsystem is now a standard part of all kernels. The VFS_AIO
kernel option and aio.ko module are gone.
Many file types may block indefinitely in their fo_read or fo_write
callbacks resulting in a hung AIO daemon. This can result in hung
user processes (when processes attempt to cancel all outstanding
requests during exit) or a hung system. To protect against this, AIO
requests are only permitted for known "safe" files by default. AIO
requests for all file types can be enabled by setting the new
vfs.aio.enable_usafe sysctl to a non-zero value. The AIO tests have
been updated to skip operations on unsafe file types if the sysctl is
zero.
Currently, AIO requests on sockets and raw disks are considered safe
and are enabled by default. aio_mlock() is also enabled by default.
Reviewed by: cem, jilles
Discussed with: kib (earlier version)
Sponsored by: Chelsio Communications
Differential Revision: https://reviews.freebsd.org/D5289
taskqueue_enqueue() was changed to support both fast and non-fast
taskqueues 10 years ago in r154167. It has been a compat shim ever
since. It's time for the compat shim to go.
Submitted by: Howard Su <howard0su@gmail.com>
Reviewed by: sephe
Differential Revision: https://reviews.freebsd.org/D5131
Note that isrc_arg member of struct intr_irqsrc is used only for
INTR_SOLO and IPI filter. This should be remembered if IPI filters
and their arguments will be stored on another place.
This option could be unusable very soon, if interrupt controllers
implementations will not be implemented considering it.
The m_ext.ext_cnt pointer becomes a union. It can now hold the refcount
value itself. To tell that m_ext.ext_flags flag EXT_FLAG_EMBREF is used.
The first mbuf to attach a cluster stores the refcount. The further mbufs
to reference the cluster point at refcount in the first mbuf. The first
mbuf is freed only when the last reference is freed.
The benefit over refcounts stored in separate slabs is that now refcounts
of different, unrelated mbufs do not share a cache line.
For EXT_EXTREF mbufs the zone_ext_refcnt is no longer needed, and m_extadd()
becomes void, making widely used M_EXTADD macro safe.
For EXT_SFBUF mbufs the sf_ext_ref() is removed, which was an optimization
exactly against the cache aliasing problem with regular refcounting.
Discussed with: rrs, rwatson, gnn, hiren, sbruno, np
Reviewed by: rrs
Differential Revision: https://reviews.freebsd.org/D5396
Sponsored by: Netflix
breaking the ABI. Special value is stored in the lock pointer to
indicate shared lock, and offline page in the shared memory is
allocated to store the actual lock.
Reviewed by: vangyzen (previous version)
Discussed with: deischen, emaste, jhb, rwatson,
Martin Simmons <martin@lispworks.com>
Tested by: pho
Sponsored by: The FreeBSD Foundation
the interrupt framework is also going to be used by another (MIPS)
architecture. IPI implementations may vary much across different
architectures.
An IPI implementation should still define INTR_IPI_COUNT and use
intr_ipi_setup_counters() to setup IPI counters which are inside of
intrcnt[] and intrnames[] arrays. Those are used for sysctl and ddb.
Then, intr_ipi_increment_count() should be used to increment obtained
counter.
Reviewed by: imp
Differential Revision: https://reviews.freebsd.org/D5459
- Set td_errno so that ktrace and dtrace can obtain the syscall error
number in the usual way.
- Pass negative error numbers directly to the syscall layer, as they're
not intended to be returned to userland.
Reviewed by: kib
Sponsored by: EMC / Isilon Storage Division
Differential Revision: https://reviews.freebsd.org/D5425
While making CloudABI work well on Linux, I discovered that I had a
FreeBSD-ism in one of my unit tests. The test did the following:
- Create UNIX socket 1, bind it, make it listen.
- Create UNIX socket 2, connect it to UNIX socket 1.
- Close UNIX socket 1.
- Obtain SO_ERROR from socket 2.
On FreeBSD this returns ECONNABORTED, while on Linux it returns
ECONNRESET. I dug through some of the relevant specifications[1] and it
looks like Linux is all right here. ECONNABORTED should only be returned
when the local connection (socket 2) is aborted; not the peer (socket 1).
It is of course slightly misleading: the function in which we set this
error is called uipc_abort(), but keep in mind that we're aborting the
peer, thus resetting the local socket.
[1] http://pubs.opengroup.org/onlinepubs/9699919799/functions/connect.html
Reviewed by: cem
Sponsored by: Nuxi, the Netherlands
Differential Revision: https://reviews.freebsd.org/D5419
the virtvnodes calculation. Include the size of fs-specific v_data as
the nfs nclnode inline, the NFS nclnode is bigger than either ZFS
znode or UFS inode. Include the size of namecache_ts and short cache
path element, multiplied by the name cache population factor, again
inline.
Inline defines are used to avoid pollution of the vnode.h with the
subsystem-private objects. Non-significant unsynchronized changes of
the definitions are fine, we do not care about that precision, and
e.g. ZFS consumes much malloced memory per vnode for reasons
unaccounted in the formula.
Lower the partition of kmem dedicated to vnodes, from 1/7 to 1/10.
The measures reduce vnode cache pressure on kmem and bring the vnode
cache memory use below some apparent thresholds that were exceeded by
r291244 due to more robust vnode reuse.
Reported and tested by: marius (i386, previous version)
Reviewed by: bde
Sponsored by: The FreeBSD Foundation
MFC after: 2 weeks
This simplifies checking for default resource range for bus_alloc_resource(),
and improves readability.
This is part of, and related to, the migration of rman_res_t from u_long to
uintmax_t.
Discussed with: jhb
Suggested by: marcel
r274560 modified kqueue_register() to only test the event condition if the
corresponding knote is not disabled. However, this check takes place before
the EV_ENABLE flag is used to clear the KN_DISABLED flag on the knote, so
enabling a previously-disabled kevent would not result in a notification for
a triggered event. This change fixes the problem by testing for EV_ENABLED
before possibly checking the event condition.
This change also updates a kqueue regression test to exercise this case.
PR: 206368
Reviewed by: kib
Sponsored by: EMC / Isilon Storage Division
Differential Revision: https://reviews.freebsd.org/D5307
Provide bus_get_bus_tag() for sparc64, powerpc, arm, arm64 and mips
nexus and its children in order to return a platform specific default tag.
This is required to ensure generic correctness of the bus_space tag.
It is especially needed for arches where child bus tag does not match
the parent bus tag. This solves the problem with ppc architecture
where the PCI bus tag differs from parent bus tag which is big-endian.
This commit is a part of the following patch:
https://reviews.freebsd.org/D4879
Submitted by: Marcin Mazurek <mma@semihalf.com>
Obtained from: Semihalf
Sponsored by: Annapurna Labs
Reviewed by: jhibbits, mmel
Differential Revision: https://reviews.freebsd.org/D4879
numbers range.
This effectively skips indirect and extdata blocks on the buffer
queue. Since their logical block numbers are negative, bnoreuselist()
could loop infinitely.
Reported and tested by: pho
Sponsored by: The FreeBSD Foundation
MFC after: 1 week
for all struct bio you get back from g_{new,alloc}_bio. Temporary
bios that you create on the stack or elsewhere should use this before
first use of the bio, and between uses of the bio. At the moment, it
is nothing more than a wrapper around bzero, but that may change in
the future. The wrapper also removes one place where we encode the
size of struct bio in the KBI.
This patch allows the newly imported INTRNG code to be built without necessarily
having FDT support in the kernel. This may be useful for some MIPS platforms
that wish to move to INTRNG, but not to FDT at the same time.
Basically all the code is already within ifdef's where FDT is concerned,
it's just the headers that aren't.
Submitted by: Stanislav Galabov <sgalabov@gmail.com>
Differential Revision: https://reviews.freebsd.org/D5249
mbuf(9) manipulation functions into uipc_mbuf.c. This looks like
the initial intent, but had diffused in the last decade.
o Gather all declarations in mbuf.h in one place and sort them.
o Uninline m_clget() and m_cljget().
There are no functional changes in this patch.
The patch comes from a larger version, where all mbuf(9) allocation was
uninlined, which allowed to make mbuf(9) UMA zones private to kern_mbuf.c.
The performance impact of the total uninlining is still unclear, so we
are holding on now with larger version.
Together with: melifaro, olivier
Kernel threads (and processes) are supposed to call kthread_exit() (or
kproc_exit()) to terminate. However, the kernel includes a fallback in
fork_exit() to force a kthread exit if a kernel thread's "main" routine
returns. This fallback was added back when the kernel only had processes
and was not updated to call kthread_exit() instead of kproc_exit() when
threads were added to the kernel.
This mistake was particular exciting when the errant thread belonged to
proc0. Due to the missing P_KTHREAD flag the fallback did not kick in
and instead tried to return to userland via whatever garbage was in the
trapframe. With P_KTHREAD set it tried to terminate proc0 resulting in
other amusements.
PR: 204999
MFC after: 1 week
All other kernel processes have this flag set and all threads in proc0
(including thread0) have the similar TDP_KTHREAD flag set.
PR: 204999
Submitted by: Oliver Pinter @ HardenedBSD
Reviewed by: kib
MFC after: 1 week
disable compilation of the code which made it possible to call
stop_all_proc() from usermode at all.
Move the comment to the preamble of stop_all_proc() and reword it to
give overview of the function intent.
proc0 has P_HADTHREADS flag set due to kthread_add(), but no
P_KTHREAD, which triggered the assert, which does not serve a purpose
now.
Reported by: Oliver Pinter
Sponsored by: The FreeBSD Foundation
MFC after: 2 weeks
Although POSIX literally permits failing with [EINVAL] if IPC_CREAT and
IPC_EXCL were both passed, the semaphore set already exists and has fewer
semaphores than nsems, this does not allow an application to retry safely:
if the [EINVAL] is actually because of the semmsl limit, an infinite loop
would result.
PR: 206927
Typically <foo>list is used for a structure that holds a list head in
FreeBSD, not for members of a list. As such, rename 'struct aiocblist'
to 'struct kaiocb' (the kernel version of 'struct aiocb').
While here, use more consistent variable names for AIO control blocks:
- Use 'job' instead of 'aiocbe', 'cb', 'cbe', or 'iocb' for kernel job
objects.
- Use 'jobn' instead of 'cbn' for use with TAILQ_FOREACH_SAFE().
- Use 'sjob' and 'sjobn' instead of 'scb' and 'scbn' for fsync jobs.
- Use 'ujob' instead of 'aiocbp', 'job', 'uaiocb', or 'uuaiocb' to hold
a user pointer to a 'struct aiocb'.
- Use 'ujobp' instead of 'aiocbp' for a user pointer to a 'struct aiocb *'.
Reviewed by: kib
Sponsored by: Chelsio Communications
Differential Revision: https://reviews.freebsd.org/D5125
with interpreter name exactly matching one wanted by the binary. If
no such brand exists, return first brand which accepted the binary by
note.
The change fixes a regression after r292749, where e.g. our two ia32
compat brands, ia32_brand_info and ia32_brand_oinfo, only differ by
the interpeter path and binary matches to a brand by linkage order.
Then old binaries which require /usr/libexec/ld-elf.so.1 but matched
against ia32_brand_info with interp_path /libexec/ld-elf.so.1, were
considered requiring non-standard interpreter name, and magic to force
ld-elf32.so.1 did not happen.
Note that it might make sense to apply the same selection of brands
for other matching criteria, SCO EI_OSABI and 3.x string.
Reported and tested by: dwmalone
Sponsored by: The FreeBSD Foundation
MFC after: 3 days
export_args on mount update, bzero() is consistent with
vfs_oexport_conv().
Make the code structure more explicit by using switch.
Return EINVAL if export option layout (deduced from size) is unknown.
Based on the submission by: bde
Sponsored by: The FreeBSD Foundation
fork1 required its callers to pass a pointer to struct proc * which would
be set to the new process (if any). procdesc and racct manipulation also
used said pointer.
However, the process could have exited prior to do_fork return and be
automatically reaped, thus making this a use-after-free.
Fix the problem by letting callers indicate whether they want the pid or
the struct proc, return the process in stopped state for the latter case.
Reviewed by: kib
print until cncheckc returned a non -1, i.e. a character had been entered.
After this change it would print only if cncheckc returned a character.
As this was before each call to db_mach_vtrace the normal outcome was
nothing was printed.
With this change 'show ktr /a' will now keep printing until the user stops
the command with a key press, or there is no more entries to print.
NOTE_CHILD and NOTE_EXIT return something in kevent.data: the parent
pid (ppid) for NOTE_CHILD and the exit status for NOTE_EXIT.
Do not let the two events be combined, since one would overwrite
the other's data.
PR: 180385
Submitted by: David A. Bright <david_a_bright@dell.com>
Reviewed by: jhb
MFC after: 1 month
Sponsored by: Dell Inc.
Differential Revision: https://reviews.freebsd.org/D4900
a buffer pointer in the event of an error (for some errors it would
return a buffer pointer and for other errors it would not return a
buffer pointer). The cluster_read() function was similarly inconsistent.
Clients of these functions were inconsistent in handling errors.
Some would assume that no buffer was returned after an error and
would thus lose buffers under certain error conditions. Others would
assume that brelse() should always be called after an error and
would thus panic the system under certain error conditions.
To correct both of these problems with minimal code churn, bread()
and cluster_write() now always free the buffer when returning an
error thus ensuring that buffers will never be lost. The brelse()
routine checks for being passed a NULL buffer pointer and silently
returns to avoid panics. Thus both approaches to handling error
returns from bread() and cluster_read() will work correctly.
Future code should be written assuming that bread() and cluster_read()
will never return a buffer with an error, so should not attempt to
brelse() the buffer when an error is returned.
Reviewed by: kib
Summary:
Migrate to using the semi-opaque type rman_res_t to specify rman resources. For
now, this is still compatible with u_long.
This is step one in migrating rman to use uintmax_t for resources instead of
u_long.
Going forward, this could feasibly be used to specify architecture-specific
definitions of resource ranges, rather than baking a specific integer type into
the API.
This change has been broken out to facilitate MFC'ing drivers back to 10 without
breaking ABI.
Reviewed By: jhb
Sponsored by: Alex Perez/Inertial Computing
Differential Revision: https://reviews.freebsd.org/D5075
close or assert the bug that it is clear when leaving.
Remove an unrelated rotted comment that was attached to the buggy
clearing.
Since draining is not done in more cases, flushing is needed in more
cases, so start fixing flushing:
- do a full flush in ttydisc_close(). State what POSIX requires more
clearly. This was missing ttydevsw_pktnotify() calls to tell the
devsw layer to flush. Hardware tty drivers don't actually flush
since they don't understand this API.
- fix 2 missing wakeups in tty_flush(). Most of the wakeups here are
unnecessary for last close. But ttydisc_close() did one of the
missing ones.
This flow control bug ameliorated the design bug of requiring
potentially unbounded waits in draining. Software flow control is the
easiest way to get an unbounded wait, and a long wait is sometimes
actually useful. Users can type the xoff character on the receiver
and (if ixon is set on the sender) expect the output to be held until
the user is ready for more.
Hardware flow control can also give the unbounded wait, and this bug
didn't affect hardware flow control. Unbounded waits from hardware
flow control take a more unusual configuration. E.g., a terminal
program that controls the modem status lines, or unplugging the cable
in a configuration where this doesn't break the connection.
The design bug is still ameliorated by a newer bug in draining for
last close -- the 1 second timeout. E.g., if the user types the
xoff character and the sender reaches last close, then output is
not resumed and the wait times out after just 1 second. This is
broken, but preferable to an unbounded wait. Before this change,
the output was resumed immediately and usually completed.
Submitted by: bde
MFC after: 2 weeks
intended behaviour in its man page. Simplify tty_drain() to match.
Don't call ttydevsw methods in tty_flush() if the device is gone
since we now sometimes call it then.
The flushing was supposed to be implemented by passing the FNONBLOCK
flag to VOP_CLOSE() for revoke(). The tty driver is one of the few
that can block in close and was one of the fewer that knew about this.
This almost worked in FreeBSD-1 and similarly in Net/2. These
versions only almost worked because there was and is considerable
confusion between IO_NDELAY and FNONBLOCK (aka O_NONBLOCK). IO_NDELAY
is only valid for VOP_READ() and VOP_WRITE(). For other VOPs it has
the same value as O_SHLOCK. But since vfs_subr.c and tty.c
consistently used the wrong flag and the O_SHLOCK flag is rarely set,
this mostly worked. It also gave the feature than applications could
get the non-blocking close by abusing O_SHLOCK.
This was first broken then fixed in 1995. I changed only the tty
driver to use FNONBLOCK, as a hack to get non-blocking via the normal
flag FNONBLOCK for last closes. I didn't know about revoke()'s use
of IO_NDELAY or change it to be consistent, so revoke() was broken.
Then I changed revoke() to match.
This was next broken in 1997 then fixed in 1998. Importing Lite2 made
the flags inconsistent again by undoing the fix only in vfs_subr.c.
This was next broken in 2008 by replacing everything in tty.c and not
checking any flags in last close. Other bugs in draining limited the
resulting unbounded waits to drain in some cases.
It is now possible to fix this better using the new FREVOKE flag.
Just restore flushing for revoke() for now. Don't restore or undo any
hacks for ordinary last closes yet. But remove dead code in the
1-second relative timeout (r272789). This did extra work to extend
the buggy draining for revoke() for as long as possible. The 1-second
timeout made this not very long by usually flushing after 1 second.
Submitted by: bde
MFC after: 2 weeks
The fail point handler may sleep, but this is not permitted while holding a
rm read lock.
MFC after: 1 week
Sponsored by: EMC / Isilon Storage Division
open (in disguise as the console device). The only allowed combination
was supposed to be the callin device with the console.
Fix the assertion in ttydev_close() that was meant to detect this (it
only detected all 3 devices being open). Assert this in ttydev_open()
too.
Submitted by: bde
MFC after: 2 weeks
(TF_OPENED_CONS) were broken in r188147 by adding TF_OPENED_CONS
without updating the string. It was especially confusing to display
OPENED_CONS as GONE and BYPASS as ZOMBIE. 2 flags at the end were
not updated in r188487.
Don't print an extra 0x prefix for %p in a ddb command. In the rest
of the kernel there are more than 6000 lines with %p and only about
40 with this bug.
Print a non-extra 0x prefix for %b in a ddb command. In the rest
of the kernel, there are approx. 180 lines with %b and 2/3 of them
have this bug.
Submitted by: bde
MFC after: 2 weeks
There are number of radix consumers in kernel land (pf,ipfw,nfs,route)
with different requirements. In fact, first 3 don't have _any_ requirements
and first 2 does not use radix locking. On the other hand, routing
structure do have these requirements (rnh_gen, multipath, custom
to-be-added control plane functions, different locking).
Additionally, radix should not known anything about its consumers internals.
So, radix code now uses tiny 'struct radix_head' structure along with
internal 'struct radix_mask_head' instead of 'struct radix_node_head'.
Existing consumers still uses the same 'struct radix_node_head' with
slight modifications: they need to pass pointer to (embedded)
'struct radix_head' to all radix callbacks.
Routing code now uses new 'struct rib_head' with different locking macro:
RADIX_NODE_HEAD prefix was renamed to RIB_ (which stands for routing
information base).
New net/route_var.h header was added to hold routing subsystem internal
data. 'struct rib_head' was placed there. 'struct rtentry' will also
be moved there soon.
data. If vnode bypass for devfs file failed, vn_read/vn_write are
called and might try to dereference f_advice. Limit the accesses to
f_advice to VREG vnodes only, which is the type ensured by
posix_fadvise().
The f_advice for regular files is protected by mtxpool lock. Recheck
that f_advice is not NULL after lock is taken.
Reported and tested by: bde
Sponsored by: The FreeBSD Foundation
MFC after: 3 weeks
separate file. Claim my copyright.
- Provide more comments, better function and structure names.
- Sort out unneeded includes from resulting two files.
No functional changes.
user VM spaces while servicing jobs. Update various comments and data
structures that refer to AIO daemons as threads to refer to processes
instead.
Reviewed by: kib
Sponsored by: Chelsio Communications
Differential Revision: https://reviews.freebsd.org/D4999
In r55943, a per-process queue of pending socket AIO requests (requests
waiting for the socket to become ready) was added so that they could be
cancelled during process rundown. In r154765, the rundown code was
changed to handle jobs in this state (JOBST_JOBQSOCK) directly removing
the need for the extra queue. However, the per-process queue head and
global lock were never removed.
Reviewed by: kib
Sponsored by: Chelsio Communications
Differential Revision: https://reviews.freebsd.org/D4997
1. vhold and zap immediately instead of postponing few lines later
2. increment numneg after new entry is added
No functional changes.
No objections: kib
Previously the code would just increment statistics while only holding a
shared lock, in effect losing updates.
Separate tracking for nchstats is removed as values can be obtained from
existing counters. Note that some fields are updated by external
consumers and are left unfixed. This should not be a serious issue as
this structure looks quite obsolete.
No strong objections: kib
to actually wait until the TX FIFOs of UARTs have be drained before
returning. This is done by bringing the equivalent of the TS_BUSY flag
found in the previous implementation back in an ABI-preserving way.
Reported and tested by: Patrick Powell
Most likely, drivers for USB-serial-adapters likewise incorporating
TX FIFOs as well as other terminal devices that buffer output in some
form should also provide implementations of tsw_busy.
MFC after: 3 days
- Pull the vmspace logic out into helper functions and reduce duplication.
Operations on the vmspace are all isolated to vm_map.c, but it now exports
a new 'vmspace_switch_aio' for use by AIO kernel processes.
- When an AIO kernel process wants to exit, break out of the main loop and
perform cleanup after the loop end. This reduces a lot of indentation and
allows cleanup to more closely mirror setup actions before the loop starts.
- Convert a DIAGNOSTIC to KASSERT().
- Replace mycp with more typical 'p'.
Reviewed by: kib
Sponsored by: Chelsio Communications
Differential Revision: https://reviews.freebsd.org/D4990
This code dates back to the initial AIO support and the commit log does
not explain why it is needed. However, I cannot find anything in the
AIO code or the various file methods (fo_read/fo_write) that would change
behavior due to using a private session instead of proc0's session.
Reviewed by: kib
Sponsored by: Chelsio Communications
Differential Revision: https://reviews.freebsd.org/D4988
This API has no in-tree consumers at the moment but is useful to at least
one out-of-tree consumer, and naturally complements existing vnode refcount
functions (vholdl(9), vdropl(9)).
Obtained from: kib (sys/ portion)
Sponsored by: EMC / Isilon Storage Division
Differential Revision: https://reviews.freebsd.org/D4947
Differential Revision: https://reviews.freebsd.org/D4953
not call VOP_CLOSE() manually. Instead, delegate the close to
fo_close() performed as part of the fdrop() on the file failed to
open. For this, finish constructing file on error, in particular, set
f_vnode and f_ops.
Forcibly resetting f_ops to badfileops disabled additional cleanups
performed by fo_close() for some file types, in this case it was noted
that cdevpriv data was corrupted. Since fo_close() call must be
enabled for some file types, it makes more sense to enable it for all
files opened through vn_open_cred().
In collaboration with: pho
Sponsored by: The FreeBSD Foundation
MFC after: 2 weeks
It hasn't been used since the AIO code was made MPSAFE 10 years ago.
Reviewed by: kib
Sponsored by: Chelsio Communications
Differential Revision: https://reviews.freebsd.org/D4946
This taskqueue is not used to handle bio requests. It is only used to
run aio_kick_nowait() to spin up new aio daemon processes.
Reviewed by: kib
Differential Revision: https://reviews.freebsd.org/D4904
and replace crcopysafe by crcopy as crcopysafe is is not intended to be
safe in a threaded environment, it drops PROC_LOCK() in while() that
can lead to unexpected results, such as overwrite kernel memory.
In my POV crcopysafe() needs special attention. For now I do not see
any problems with this function, but who knows.
Submitted by: dchagin
Found by: trinity
Security: SA-16:04.linux
"The problem with the approach taken both in _bus_dmamap_load_pages and
bus_dmamap_load_ma_triv is that they split the request buffer into
arbitrary chunks based on page boundaries, creating segments that no
longer have a size that's a multiple of the sector size. This breaks
drivers like blkfront (and probably other stuff)." [1]
This was most easily triggered by running `fsck /` on a system running
in Xen (e.g. Amazon EC2) but also showed up via growfs(8) and probably
many other userland tools which access the disk directly.
Patch by: royger [1]
"Thinks this should be fine" by: ken
r282971 attempted to fix this problem by decrementing cv_waiters after
waking up from sleeping on a condition variable, but this can result in
a use-after-free if the CV is freed before all woken threads have had a
chance to run. Instead, avoid incrementing cv_waiters past INT_MAX, and
have cv_signal() explicitly check for sleeping threads once cv_waiters has
reached this bound.
Reviewed by: jhb
MFC after: 2 weeks
Sponsored by: EMC / Isilon Storage Division
Differential Revision: https://reviews.freebsd.org/D4822
up to now.
The new sendfile is the code that Netflix uses to send their multiple tens
of gigabits of data per second. The new implementation features asynchronous
I/O, when I/O operations are launched, but not awaited to be complete. An
explanation of why such behavior is beneficial compared to old one is
going to be too long for a commit message, so we will skip it here.
Additional features of new syscall are extra flags, which provide an
application more control over data sent. The SF_NOCACHE flag tells
kernel that data shouldn't be cached after it was sent. The SF_READAHEAD()
macro allows to specify readahead size in pages.
The new syscalls is a drop in replacement. No modifications are required
to applications. One can take nginx binary for stable/10 and run it
successfully on head. Although SF_NODISKIO lost its original sense, as now
sendfile doesn't block, and now means something completely different (tm),
using the new sendfile the old way is absolutely safe.
Celebrates: Netflix global launch!
Sponsored by: Nginx, Inc.
Sponsored by: Netflix
Relnotes: yes
sbappendstream() does. Although, M_NOTREADY may appear only on SOCK_STREAM
sockets, due to sendfile(2) supporting only the latter, there is a corner
case of AF_UNIX/SOCK_STREAM socket, that still uses records for the sake
of control data, albeit being stream socket.
Provide private version of m_clrprotoflags(), which understands PRUS_NOTREADY,
similar to m_demote().
Tty.c was untypical in that it handled the si_drv1 issue consistently
and correctly, by always checking for si_drv1 being non-NULL and
sleeping if NULL. The removed code also illustrated unneeded
complications in drivers which are eliminated by the use of new KPI.
Reviewed by: hps, jhb
Discussed with: bde
Sponsored by: The FreeBSD Foundation
MFC after: 3 weeks
Differential revision: https://reviews.freebsd.org/D4746
Immediate problem fixed by the new KPI is the long-standing race
between device creation and assignments to cdev->si_drv1 and
cdev->si_drv2, which allows the window where cdevsw methods might be
called with si_drv1,2 fields not yet set. Devices typically checked
for NULL and returned spurious errors to usermode, and often left some
methods unchecked.
The new function interface is designed to be extensible, which should
allow to add more features to make_dev_s(9) without inventing yet
another name for function to create devices, while maintaining KPI and
even KBI backward-compatibility.
Reviewed by: hps, jhb
Sponsored by: The FreeBSD Foundation
MFC after: 3 weeks
Differential revision: https://reviews.freebsd.org/D4746
Advance the logical block number to the lblkno of the found block plus
one, instead of incrementing the block number which was used for
lookup. This change skips sparcely populated buffer ranges, similar
to r292325, instead of doing useless lookups.
Do not restart the bnoreuselist() from the start of the range if
buffer lock cannot be obtained without sleep. Only retry lookup and
lock for the same queue and same logical block number.
Reported by: benno
Tested by: pho
Sponsored by: The FreeBSD Foundation
MFC after: 3 days
providing compiled-in static environment data that is used instead of any
data passed in from a boot loader.
Previously 'env' worked only on i386 and arm xscale systems, because it
required the MD startup code to examine the global envmode variable and
decide whether to use static_env or an environment obtained from the boot
loader, and set the global kern_envp accordingly. Most startup code wasn't
doing so. Making things even more complex, some mips startup code uses an
alternate scheme that involves calling init_static_kenv() to pass an empty
buffer and its size, then uses a series of kern_setenv() calls to populate
that buffer.
Now all MD startup code calls init_static_kenv(), and that routine provides
a single point where envmode is checked and the decision is made whether to
use the compiled-in static_kenv or the values provided by the MD code.
The routine also continues to serve its original purpose for mips; if a
non-zero buffer size is passed the routine installs the empty buffer ready
to accept kern_setenv() values. Now if the size is zero, the provided buffer
full of existing env data is installed. A NULL pointer can be passed if the
boot loader provides no env data; this allows the static env to be installed
if envmode is set to do so.
Most of the work here is a near-mechanical change to call the init function
instead of directly setting kern_envp. A notable exception is in xen/pv.c;
that code was originally installing a buffer full of preformatted env data
along with its non-zero size (like mips code does), which would have allowed
kern_setenv() calls to wipe out the preformatted data. Now it passes a zero
for the size so that the buffer of data it installs is treated as
non-writeable.
CPU_ISSET(), CPU_SET etc. in sparc64 asm. This approach has the
benefit of not clobbering %y, allowing to revert r222827 and
partially r222828.
- In r222828, CATR() already was changed to use the equivalent of
PCPU_GET(cpuid) instead of the MD module ID for KTR_CPU, so
belatedly also catch up with the C side of ktr(9). Originally,
in r203838 CATR() was moved away from directly reading the
module ID or equivalent as that became impractical with other
CPU types than USI/II supported. With r222828 in place, per-CPU
data generally is set up soon enough, though, that employing
PCPU things in ktr(9) also for use during early stages works.
- Unfortunately, an exception to the latter is the ktr(9) use
in pmap_bootstrap(), which actually is run so early that even
checking for bootverbose being set via the loader doesn't work.
Consequently, replace the ktr(9) use in pmap_bootstrap() with
OF_printf(9) and put it under #ifdef DIAGNOSTIC instead.
MFC after: 3 days
Add two new LWPINFO flags: PL_FLAG_BORN and PL_FLAG_EXITED for reporting
thread creation and destruction. Newly created threads will stop to report
PL_FLAG_BORN before returning to userland and exiting threads will stop to
report PL_FLAG_EXIT before exiting completely. Both of these events are
only enabled and reported if PT_LWP_EVENTS is enabled on a process.
This code is missing the racct_subr() call from kern_thr_exit() and would
require further code duplication in future changes.
Reviewed by: kib
MFC after: 1 week
(CLOCK_REALTIME case) system calls is non negative.
This commit hides a kernel panic in atrtc_settime() as the clock_ts_to_ct()
does not properly convert negative tv_sec.
ps. in my opinion clock_ts_to_ct() should be rewritten to properly handle
negative tv_sec values.
Differential Revision: https://reviews.freebsd.org/D4714
Reviewed by: kib
MFC after: 1 week
different from the interpreter path requested by the binary.
Before this change, it is impossible to activate non-default
interpreter for 32bit image on amd64, when /libexec/ld-elf32.so.1 file
exists.
Sponsored by: The FreeBSD Foundation
MFC after: 2 weeks
vim overzealously removed some trailing `+' and I didn't check the
diff
MFC after: 1 week
X-MFC with: r292640
Pointyhat to: ngie
Sponsored by: EMC / Isilon Storage Division
exhausted.
It is possible for a bug in the code (or, theoretically, even unusual
network conditions) to exhaust all possible mbufs or mbuf clusters.
When this occurs, things can grind to a halt fairly quickly. However,
we currently do not call mb_reclaim() unless the entire system is
experiencing a low-memory condition.
While it is best to try to prevent exhaustion of one of the mbuf zones,
it would also be useful to have a mechanism to attempt to recover from
these situations by freeing "expendable" mbufs.
This patch makes two changes:
a) The patch adds a generic API to the UMA zone allocator to set a
function that should be called when an allocation fails because the
zone limit has been reached. Because of the way this function can be
called, it really should do minimal work.
b) The patch uses this API to try to free mbufs when an allocation
fails from one of the mbuf zones because the zone limit has been
reached. The function schedules a callout to run mb_reclaim().
Differential Revision: https://reviews.freebsd.org/D3864
Reviewed by: gnn
Comments by: rrs, glebius
MFC after: 2 weeks
Sponsored by: Juniper Networks
During fork p_starcopy - p_endcopy area of a process is populated with bcopy
with only proc lock held. Another forking thread can find such a process and
proceed to access p_pgrp included in said area.
Fix the problem by moving the field outside. It is being properly assigned
later.
Reviewed by: kib
Diagnosed by: kib
Tested by: Fabian Keil <freebsd-listen fabiankeil.de>
MFC after: 10 days
The ci20 port (by kan@) is going to reuse almost all of the intrng code
since the SoC in question looks suspiciously like someone took an ARM
SoC design and replaced the ARM core with a MIPS core.
* migrate out the code;
* rename ARM_ -> INTR_;
* rename arm_ -> intr_;
* move the interrupt flush routine from intr.c / intrng.c into
arm/machdep_intr.c - removing the code duplication and removing
the ARM specific bits from here.
Thanks to the Star Wars: The Force Awakens premiere line for allowing
me a couple hours of quiet time to finish the universe builds.
Tested:
* make universe
TODO:
* The structure definitions in subr_intr.c still includes machine/intr.h
which requires one duplicates all of the intrng definitions in
the platform code (which kan has done, and I think we don't have to.)
Instead I should break out the generic things (function declarations,
common intr structures, etc) into a separate header.
* Kan has requested I make the PIC based IPI stuff optional.
Rather than pushing all eight possible arguments into dtrace_probe()'s
stack frame, make the syscall_args struct for the current syscall available
via the current thread. Using a custom getargval method for the systrace
provider, this allows any syscall argument to be fetched, even in kernels
that have modified the maximum number of system call arguments.
Sponsored by: EMC / Isilon Storage Division
- Use SDT_PROBE<N>() instead of SDT_PROBE(). This has no functional effect
at the moment, but will be needed for some future changes.
- Don't hardcode the module component of the probe identifier. This is
set automatically by the SDT framework.
MFC after: 1 week
o With new KPI consumers can request contiguous ranges of pages, and
unlike before, all pages will be kept busied on return, like it was
done before with the 'reqpage' only. Now the reqpage goes away. With
new interface it is easier to implement code protected from race
conditions.
Such arrayed requests for now should be preceeded by a call to
vm_pager_haspage() to make sure that request is possible. This
could be improved later, making vm_pager_haspage() obsolete.
Strenghtening the promises on the business of the array of pages
allows us to remove such hacks as swp_pager_free_nrpage() and
vm_pager_free_nonreq().
o New KPI accepts two integer pointers that may optionally point at
values for read ahead and read behind, that a pager may do, if it
can. These pages are completely owned by pager, and not controlled
by the caller.
This shifts the UFS-specific readahead logic from vm_fault.c, which
should be file system agnostic, into vnode_pager.c. It also removes
one VOP_BMAP() request per hard fault.
Discussed with: kib, alc, jeff, scottl
Sponsored by: Nginx, Inc.
Sponsored by: Netflix
buffer for each block number in the range with gbincore(), look up the
next instantiated buffer with the logical block number which is
greater or equal to the next lblkno. This significantly speeds up the
iteration for sparce-populated range.
Move the iteration into new helper bnoreuselist(), which is structured
similarly to flushbuflist().
Reported and tested by: pho
Reviewed by: markj
Sponsored by: The FreeBSD Foundation
the type stability of the buffers memory. Instead of memoizing
pointer to the next buffer and validating it, remember the next
logical block number in the bo list and re-lookup.
Reviewed by: markj
Tested by: pho
Sponsored by: The FreeBSD Foundation
The intention was to just limit leading zeroes on numeric names. That
check is now improved to also catch the leading spaces and '+' that
strtoul can pass through.
PR: 204897
MFC after: 3 days
A panicking thread always executes with a critical section held, so any
attempt to allocate or free memory while dumping will otherwise cause a
second panic. This can occur, for example, if xpt_polled_action() completes
non-dump I/O that was pending at the time of the panic. The fact that this
can occur is itself a bug, but asserting in this case does little but
reduce the reliability of kernel dumps.
Suggested by: kib
Reported by: pho
These helper functions can be used to read in or write a buffer from or to
an arbitrary process' address space. Without them, this can only be done
using proc_rwmem(), which requires the caller to fill out a uio. This is
onerous and results in code duplication; the new functions provide a simpler
interface which is sufficient for most existing callers of proc_rwmem().
This change also adds a manual page for proc_rwmem() and the new functions.
Reviewed by: jhb, kib
Differential Revision: https://reviews.freebsd.org/D4245
clock_gettime(2) on ARMv7 and ARMv8 systems which have architectural
generic timer hardware. It is similar how the RDTSC timer is used in
userspace on x86.
Fix a permission problem where generic timer access from EL0 (or
userspace on v7) was not properly initialized on APs.
For ARMv7, mark the stack non-executable. The shared page is added for
all arms (including ARMv8 64bit), and the signal trampoline code is
moved to the page.
Reviewed by: andrew
Discussed with: emaste, mmel
Sponsored by: The FreeBSD Foundation
Differential revision: https://reviews.freebsd.org/D4209
For completeness add a VNASSERT that there are no threads waiting on a
range lock (this was previously checked on every vnode free).
Reported by; Rick Macklem
Fix from: Mateusz Guzik
PR: 204949
camdd(8) utility.
CCBs may be queued to the driver via the new CAMIOQUEUE ioctl, and
completed CCBs may be retrieved via the CAMIOGET ioctl. User
processes can use poll(2) or kevent(2) to get notification when
I/O has completed.
While the existing CAMIOCOMMAND blocking ioctl interface only
supports user virtual data pointers in a CCB (generally only
one per CCB), the new CAMIOQUEUE ioctl supports user virtual and
physical address pointers, as well as user virtual and physical
scatter/gather lists. This allows user applications to have more
flexibility in their data handling operations.
Kernel memory for data transferred via the queued interface is
allocated from the zone allocator in MAXPHYS sized chunks, and user
data is copied in and out. This is likely faster than the
vmapbuf()/vunmapbuf() method used by the CAMIOCOMMAND ioctl in
configurations with many processors (there are more TLB shootdowns
caused by the mapping/unmapping operation) but may not be as fast
as running with unmapped I/O.
The new memory handling model for user requests also allows
applications to send CCBs with request sizes that are larger than
MAXPHYS. The pass(4) driver now limits queued requests to the I/O
size listed by the SIM driver in the maxio field in the Path
Inquiry (XPT_PATH_INQ) CCB.
There are some things things would be good to add:
1. Come up with a way to do unmapped I/O on multiple buffers.
Currently the unmapped I/O interface operates on a struct bio,
which includes only one address and length. It would be nice
to be able to send an unmapped scatter/gather list down to
busdma. This would allow eliminating the copy we currently do
for data.
2. Add an ioctl to list currently outstanding CCBs in the various
queues.
3. Add an ioctl to cancel a request, or use the XPT_ABORT CCB to do
that.
4. Test physical address support. Virtual pointers and scatter
gather lists have been tested, but I have not yet tested
physical addresses or scatter/gather lists.
5. Investigate multiple queue support. At the moment there is one
queue of commands per pass(4) device. If multiple processes
open the device, they will submit I/O into the same queue and
get events for the same completions. This is probably the right
model for most applications, but it is something that could be
changed later on.
Also, add a new utility, camdd(8) that uses the asynchronous pass(4)
driver interface.
This utility is intended to be a basic data transfer/copy utility,
a simple benchmark utility, and an example of how to use the
asynchronous pass(4) interface.
It can copy data to and from pass(4) devices using any target queue
depth, starting offset and blocksize for the input and ouptut devices.
It currently only supports SCSI devices, but could be easily extended
to support ATA devices.
It can also copy data to and from regular files, block devices, tape
devices, pipes, stdin, and stdout. It does not support queueing
multiple commands to any of those targets, since it uses the standard
read(2)/write(2)/writev(2)/readv(2) system calls.
The I/O is done by two threads, one for the reader and one for the
writer. The reader thread sends completed read requests to the
writer thread in strictly sequential order, even if they complete
out of order. That could be modified later on for random I/O patterns
or slightly out of order I/O.
camdd(8) uses kqueue(2)/kevent(2) to get I/O completion events from
the pass(4) driver and also to send request notifications internally.
For pass(4) devcies, camdd(8) uses a single buffer (CAM_DATA_VADDR)
per CAM CCB on the reading side, and a scatter/gather list
(CAM_DATA_SG) on the writing side. In addition to testing both
interfaces, this makes any potential reblocking of I/O easier. No
data is copied between the reader and the writer, but rather the
reader's buffers are split into multiple I/O requests or combined
into a single I/O request depending on the input and output blocksize.
For the file I/O path, camdd(8) also uses a single buffer (read(2),
write(2), pread(2) or pwrite(2)) on reads, and a scatter/gather list
(readv(2), writev(2), preadv(2), pwritev(2)) on writes.
Things that would be nice to do for camdd(8) eventually:
1. Add support for I/O pattern generation. Patterns like all
zeros, all ones, LBA-based patterns, random patterns, etc. Right
Now you can always use /dev/zero, /dev/random, etc.
2. Add support for a "sink" mode, so we do only reads with no
writes. Right now, you can use /dev/null.
3. Add support for automatic queue depth probing, so that we can
figure out the right queue depth on the input and output side
for maximum throughput. At the moment it defaults to 6.
4. Add support for SATA device passthrough I/O.
5. Add support for random LBAs and/or lengths on the input and
output sides.
6. Track average per-I/O latency and busy time. The busy time
and latency could also feed in to the automatic queue depth
determination.
sys/cam/scsi/scsi_pass.h:
Define two new ioctls, CAMIOQUEUE and CAMIOGET, that queue
and fetch asynchronous CAM CCBs respectively.
Although these ioctls do not have a declared argument, they
both take a union ccb pointer. If we declare a size here,
the ioctl code in sys/kern/sys_generic.c will malloc and free
a buffer for either the CCB or the CCB pointer (depending on
how it is declared). Since we have to keep a copy of the
CCB (which is fairly large) anyway, having the ioctl malloc
and free a CCB for each call is wasteful.
sys/cam/scsi/scsi_pass.c:
Add asynchronous CCB support.
Add two new ioctls, CAMIOQUEUE and CAMIOGET.
CAMIOQUEUE adds a CCB to the incoming queue. The CCB is
executed immediately (and moved to the active queue) if it
is an immediate CCB, but otherwise it will be executed
in passstart() when a CCB is available from the transport layer.
When CCBs are completed (because they are immediate or
passdone() if they are queued), they are put on the done
queue.
If we get the final close on the device before all pending
I/O is complete, all active I/O is moved to the abandoned
queue and we increment the peripheral reference count so
that the peripheral driver instance doesn't go away before
all pending I/O is done.
The new passcreatezone() function is called on the first
call to the CAMIOQUEUE ioctl on a given device to allocate
the UMA zones for I/O requests and S/G list buffers. This
may be good to move off to a taskqueue at some point.
The new passmemsetup() function allocates memory and
scatter/gather lists to hold the user's data, and copies
in any data that needs to be written. For virtual pointers
(CAM_DATA_VADDR), the kernel buffer is malloced from the
new pass(4) driver malloc bucket. For virtual
scatter/gather lists (CAM_DATA_SG), buffers are allocated
from a new per-pass(9) UMA zone in MAXPHYS-sized chunks.
Physical pointers are passed in unchanged. We have support
for up to 16 scatter/gather segments (for the user and
kernel S/G lists) in the default struct pass_io_req, so
requests with longer S/G lists require an extra kernel malloc.
The new passcopysglist() function copies a user scatter/gather
list to a kernel scatter/gather list. The number of elements
in each list may be different, but (obviously) the amount of data
stored has to be identical.
The new passmemdone() function copies data out for the
CAM_DATA_VADDR and CAM_DATA_SG cases.
The new passiocleanup() function restores data pointers in
user CCBs and frees memory.
Add new functions to support kqueue(2)/kevent(2):
passreadfilt() tells kevent whether or not the done
queue is empty.
passkqfilter() adds a knote to our list.
passreadfiltdetach() removes a knote from our list.
Add a new function, passpoll(), for poll(2)/select(2)
to use.
Add devstat(9) support for the queued CCB path.
sys/cam/ata/ata_da.c:
Add support for the BIO_VLIST bio type.
sys/cam/cam_ccb.h:
Add a new enumeration for the xflags field in the CCB header.
(This doesn't change the CCB header, just adds an enumeration to
use.)
sys/cam/cam_xpt.c:
Add a new function, xpt_setup_ccb_flags(), that allows specifying
CCB flags.
sys/cam/cam_xpt.h:
Add a prototype for xpt_setup_ccb_flags().
sys/cam/scsi/scsi_da.c:
Add support for BIO_VLIST.
sys/dev/md/md.c:
Add BIO_VLIST support to md(4).
sys/geom/geom_disk.c:
Add BIO_VLIST support to the GEOM disk class. Re-factor the I/O size
limiting code in g_disk_start() a bit.
sys/kern/subr_bus_dma.c:
Change _bus_dmamap_load_vlist() to take a starting offset and
length.
Add a new function, _bus_dmamap_load_pages(), that will load a list
of physical pages starting at an offset.
Update _bus_dmamap_load_bio() to allow loading BIO_VLIST bios.
Allow unmapped I/O to start at an offset.
sys/kern/subr_uio.c:
Add two new functions, physcopyin_vlist() and physcopyout_vlist().
sys/pc98/include/bus.h:
Guard kernel-only parts of the pc98 machine/bus.h header with
#ifdef _KERNEL.
This allows userland programs to include <machine/bus.h> to get the
definition of bus_addr_t and bus_size_t.
sys/sys/bio.h:
Add a new bio flag, BIO_VLIST.
sys/sys/uio.h:
Add prototypes for physcopyin_vlist() and physcopyout_vlist().
share/man/man4/pass.4:
Document the CAMIOQUEUE and CAMIOGET ioctls.
usr.sbin/Makefile:
Add camdd.
usr.sbin/camdd/Makefile:
Add a makefile for camdd(8).
usr.sbin/camdd/camdd.8:
Man page for camdd(8).
usr.sbin/camdd/camdd.c:
The new camdd(8) utility.
Sponsored by: Spectra Logic
MFC after: 1 week
on every allocation and fully releases them on every free. These
are not trivial costs: it starts by zeroing a large structure then
initializes a mutex, a lock manager lock, an rw lock, four lists,
and six pointers. And looking at vfs.vnodes_created, these operations
are being done millions of times an hour on a busy machine.
As a performance optimization, this code update uses the uma_init
and uma_fini routines to do these initializations and cleanups only
as the vnodes enter and leave the vnode_zone. With this change the
initializations are only done kern.maxvnodes times at system startup
and then only rarely again. The frees are done only if the vnode_zone
shrinks which never happens in practice. For those curious about the
avoided work, look at the vnode_init() and vnode_fini() functions in
kern/vfs_subr.c to see the code that has been removed from the main
vnode allocation/free path.
Reviewed by: kib
Tested by: Peter Holm