the NFSv3 ACCESS RPC problems a little for busy clients that do a lot of
open/close. The nfs code could probably cache the results, but I'm not
sure whether this would be legal or useful. The problem is that with
a CPU farm, on each open there would be a lookup, getattr then access RPC
then the read/write RPC activity. Caching the access results probably
isn't going to help much if the clients access lots of files. Having the
nfs_access() routine interpret the getattr results is a bit of a hack, but
it's how NFSv2 is done and it might be OK for a mount attribute for v3.
manipulation away from the length comparison. Measurements on beast.cdrom.com
show >3X improvement over the original code on large block sizes, putting the
performance on par with the optimized assembly code in libc.
and increase the tx interrupt threshold to 4. This fixes performance
problems on slower systems.
Also fix a mind-o in the rx ring init routine: I used the TX
constant instead of the RX. This isn't a problem as long as the
rings are the same size, but if they aren't hijinx will ensue.
* Embed the stack into the bss section for loader and netboot. This
is required for netboot since otherwise the stack would be inside our
heap.
* Install loader and netboot in /boot by default.
* Fix getbootfile so that it searches for a ',' instead of a ';'
when terminating the filename.
sure that this is necessary to be a sync write here since a VOP_FSYNC()
follows and it will schedule, sort and complete the writes that the
vm_object_page_clean() started (as I think I understand things).
- Use TAILQ_* macros extensively instead of internal names
- use b_xflags instead of the NOLIST magic number hack in the next pointer
- clean bufs are inserted at the tail rather than the head.
- redo dirty buffer insert so that metadata (negative lbn) goes to the
tail directly rather than at the HEAD. This makes a difference when
inserting dirty data blocks in lbn sorted order since data block
insertion will not have to bypass all the metadata cruft. data is
lbn sorted since it makes sense for clustering and writeback ordering,
while metadata sorting doesn't help much since the lbn's are
meaningless when walking the list for writebacks.
Small systems will not notice much (if any) benefit from this, but really
busy systems with large dirty block lists should get a lot more.
I've tested this with softdep, and it doesn't seem to mind the change of
queueing of metadata.
Reviewed (in princible) by: dg
Obtained from: partly from John Dyson's work-in-progress patches in June.
Add a new flags field (we get this for free because of struct packing)
for cleaner management of tailq membership.
We had two spare b_flags slots, but they are a precious resource and may
be needed for other things that are related to other b_flags bits. The two
new flags are convenient to use in a seperate location.
Reviewed (in principle) by: dg
Obtained from: John Dyson's old work-in-progress
basic i/o functions, bit-banging mechanism is implemented by dev/iicbus/iicbb.c
immio.c: some bootverbose logs to watch zip+ connect/disconnect process
a vga.
* Fix broken logic in syscons for a failed probe.
* Fix AlphaStation 500/600 so that non-serial consoles are supported.
Submitted by: Thomas Valentino Crimi <tcrimi+@andrew.cmu.edu> (vga bits),
Andrew Gallatin <gallatin@cs.duke.edu> (AS500/AS600)
the old true/false.
While here, have vfs_msync() only call vm_object_page_clean() with
OBJPC_SYNC if called with MNT_WAIT flags. vfs_msync() is called at unmount
time (with MNT_WAIT) and from the syncer process (formerly update).
This should make dirty mmap writebacks a little less nasty.
I have tested this a little with SOFTUPDATES enabled, but I don't normally
use it since I've been badly burned too many times.
filesystems.
- New 'help' command and data in the help.* files (not yet installed),
provides topic and subtopic help, indexes, etc.
- Don't crash if the user tries to set an invalid console. Be helpful
instead.
- Expand tabs (badly) on the i386 video console.
- Some minor cosmetic changes.
installed.
Remove cpu_power_down, and replace it with an entry at the end of the
SHUTDOWN_FINAL queue in the only place it's used (APM).
Submitted by: Some ideas from Bruce Walter <walter@fortean.com>
and dies if it can't find the MFS root whereas the x86 one seems to sail
past. Looking at the code, I can't see how either one works, so I'm
confused. :)
client programs are allowed to finish up (coda_call is
forced to complete) and release their locks. Thus there
is a reasonable chance that the vflush implicit in the
unmount will not get hung on held locks.
They checked for the magic major number for the "device" behind mfs
mount points. Use a more obvious check for this device.
Debugged by: Andrew Gallatin <gallatin@cs.duke.edu>
clear if the check is necessary, but vfs_object_create() is called
for all vnodes and it was silly to create objects for VBLK vnodes
that don't even have a driver.
when bdevsw[] became sparse. We still depend on magic to avoid having to
check that (v_rdev) device numbers in vnodes are not NODEV.
Removed a redundant `major(dev) < nblkdev' test instead of updating it.
Don't follow a garbage bdevsw pointer for attempts to swap on empty
regular files. This case currently can't happen. Swapping on regular
files is ifdefed out in swapon() and isn't attempted for empty files
in nfs_mountroot().
when bdevsw[] became sparse. We still depend on magic to avoid having to
check that (v_rdev) device numbers in vnodes are not NODEV.
Removed redundant `major(dev) < nblkdev' tests instead of updating them.
- dev != NODEV was checked for, but 0 was returned on failure. This was
fixed in Lite2 (except the return code was still slightly wrong (ENODEV
instead of ENXIO)) but the changes were not merged. This case probably
doesn't actually occur under FreeBSD.
- major(dev) was not checked to have a valid non-NULL bdevsw entry. This
caused panics when the driver for the root device didn't exist.
Fixed minor misformattings in bdevvp(). Rev.1.14 consisted mainly of
gratuitous reformattings that seem to have caused many Lite2 merge
errors.
PR: 8417
minor devices.
Improve PLL/OCXO DAC dithering.
General remodeling.
Performance is now 2.5e-11 in frequency and +/- 100 nsec in time, both
of which are actually the limits of the transmitted signal.
Interrupt handlers are now configured in drivers.
Didn't update config/SMM.doc. It doesn't have any i386 examples (not
even `isa').
Bumped CONFIGVERS. This is not necessary for -current yet, but using
the new config with old system sources gives null pointers for all
vectors.
If you have problems with the "calcru" messages and processes being
killed for excessive cpu time, try to increase the NTIMECOUNTER
#define and report your findings.
needs to be called prior to freeing remaining pages in the object so that
the device pager has an opportunity to grab its "fake" pages. Also, in
the case of wired pages, the page must be made busy prior to calling
vm_page_remove. This is a difference from 2.2.x that I overlooked when
I brought these changes forward.
to a device failed.
In theory, the same steps that happen when we get an AC_LOST_DEVICE async
notification should have been taken when a driver fails to attach. In
practice, that wasn't the case.
This only affected the da, cd and ch drivers, but the fix affects all
peripheral drivers.
There were several possible problems:
- In the da driver, we didn't remove the peripheral's softc from the da
driver's linked list of softcs. Once the peripheral and softc got
removed, we'd get a kernel panic the next time the timeout routine
called dasendorderedtag().
- In the da, cd and possibly ch drivers, we didn't remove the
peripheral's devstat structure from the devstat queue. Once the
peripheral and softc were removed, this could cause a panic if anyone
tried to access device statistics. (one component of the linked list
wouldn't exist anymore)
- In the cd driver, we didn't take the peripheral off the changer run
queue if it was scheduled to run. In practice, it's highly unlikely,
and maybe impossible that the peripheral would have been on the
changer run queue at that stage of the probe process.
The fix is:
- Add a new peripheral callback function (the "oninvalidate" function)
that is called the first time cam_periph_invalidate() is called for a
peripheral.
- Create new foooninvalidate() routines for each peripheral driver. This
routine is always called at splsoftcam(), and contains all the stuff
that used to be in the AC_LOST_DEVICE case of the async callback
handler.
- Move the devstat cleanup call to the destructor/cleanup routines, since
some of the drivers do I/O in their close routines.
- Make sure that when we're flushing the buffer queue, we traverse it at
splbio().
- Add a check for the invalid flag in the pt driver's open routine.
Reviewed by: gibbs
Revert the transmission packet queueing strategy changes. Clearly I missed
something while debugging this, although I never encountered any problems
on my test machines.
Also make one other minor change: jack up the TX reclaim threshold for
3c90xB adapters in order to stave off 'transmission error: 82' errors.
Document the existence of the tx reclaim register (for inspecting the
current reclaim threshold) in register window 5 (if_xlreg.h).
{ port_name = "IO_TIMER", port_number = 1 } and only worked because
it was reassembled to "IO_TIMER1". Trailing digits always work, but
this is too magic to depend on.
Don't quote port names that don't have a digit in them.
bug sound was not played if the total amount of data written to
the device was less than one blocksize
Noticed by: NABETANI Masaki and FreeBSD-users-jp
legitimately wired pages. Currently we print a diagnostic when this
happens, but this will be removed soon when it will be common for this
to occur with zero-copy TCP/IP buffers.
- Use the ISA PnP enumerator.
- Use the new linker set code, throw out the gensetdefs stuff.
- Produce an intermediate loader image that has symbols stripped, to aid
- in debugging.
- Supply ISA port access functions required for ISA PnP
can fit into my test machine.
- Move to using STAILQs rather than ad-hoc singly-linked lists.
- Use a mostly procedural interface to the PnP information. This
improves data-hiding.
Implement a new linker-set technique (currently on i386 only but should work
on Alpha as well). This is a good candidate for replacing the current
gensetdefs cruft completely.
config_drive:
Catch an instance of anonymous drives. Doubtless many remain.
interrupt.c:
complete_rqe:
Call logrq to log iodone events if DEBUG_LASTREQS is set.
Call set_sd_state with setstate_noupdate to avoid buffered I/O out
of interrupt context.
Use define DEBUG_RESID instead of constant.
memory.c:
Remove dead expandrq() function
Malloc:
Remove directory component of file names in malloc table.
Add function vinum_rqinfo (part of the request tracing stuff).
request.c:
Add function logrq (part of the request tracing stuff).
vinumstrategy:
Check whether config needs to be written to disk, do it if so.
This is a stopgap until the Vinum daemon (bacchusd? oenologistd?)
is written.
If DEBUG_LASTREQS is set, call logrq to log user buffer headers.
launch_requests:
Correct format of debug output to console.
If DEBUG_LASTREQS is set, call logrq to log request elements.
request.h:
Add definitions for request trace.
state.c:
set_sd_state:
Check flags for setstate_noupdate. If set, don't write the config
to disk, just set global VF_DIRTYCONFIG flag. This is part of the
kludge to avoid writing config from an interrupt context.
vinumext.h:
Add declaration for vinum_rqinfo, put inside #ifdef DEBUG
Remove dead macro expandrq
vinumio.h:
Increase maximum ioctl reply length to 4 kB if DEBUG is set.
Define VINUM_RQINFO ioctl if DEBUG is set.
vinumioctl.c:
vinumioctl:
Change implementation of VINUM_DEBUG ioctl: use a debug flag
(DEBUG_REMOTEGDB) to decide whether to go into remote debugging or
not.
Implement VINUM_RQINFO.
vinumkw.h:
Define kw_info even when not debugging.
vinumvar.h:
Define VF_DIRTYCONFIG
Add pointers to request info to vinum_info if DEBUG is set.
Define setstate_noupdate
Define additional debug bits DEBUG_RESID, DEBUG_LASTREQS and
DEBUG_REMOTEGDB.
agressive. With the old code, if a descriptor chain was already on its
way to the chip, xl_start() would try to splice new chains onto the end
of the current chain by stopping the transmitter, modifying the tail
pointer of the current chain to point to the head of the new chain, then
restart the transmitter. The manual says you're allowed to do this and
it works, but I'm not too keen on it anymore.
The new code waits until the eixsting chain has been sent and then
queues the next waiting chain in the 'transmit ok' handler.
Performance still looks good one way or the other.
RealTek 8129/8139 chipset like I've been threatening. Update kernel
configs, userconfig.c, relnotes and sysinstall. No man page yet;
comming soon.
I consider this driver stable enough that I want to give it some
exposure in -current.
- Use the system headers method for Elf32/Elf64 symbol compatability
- get rid of the UPRINTF debugging.
- check the ELF header for compatability much more completely
- optimize the section mapper. Use the same direct VM interfaces that
imgact_aout.c and kern_exec.c use.
- Check the return codes from the vm_* functions better. Some return
KERN_* results, not an errno.
- prefault the page tables to reduce startup faults on page tables like
a.out does.
- reset the segment protection to zero for each loop, otherwise each
segment could get progressively more privs. (eg: if the first was
read/write/execute, and the second was meant to be read/execute, the
bug would make the second r/w/x too. In practice this was not a
problem because executables are normally laid out with text first.)
- Don't impose arbitary limits. Use the limits on headers imposed by
the need to fit them into one page.
- Remove unused switch() cases now that the verbose debugging is gone.
I've been using an earlier version of this for a month or so.
This sped up ELF exec speed a bit for me but I found it hard to get
consistant benchmarks when I tested it last (a few weeks ago).
I'm still bothered by the page read out of order caused by the
transition from data to bss. This which requires either part filling the
transition page or clearing the remainder.
Move the initialization before isa_configure() and npx, in case npx does
something to initialize the state of the emulator somehow.
I do not have any machines without a FPU so that I can test this with -
except an old 386sx motherboard in a box somewhere that might work...
Supply the moduledata handle rather than the event dispatcher function.
This should explain the panic on boot problem that's been discussed in
-current at the moment. Both machines had GNU_MATH_EMULATE.
the boot.config settings are persistent, this seems to provide a
useful override capability, and should break only on broken
boot.config's.
Output a cosmetic newline if booting with no input.
a raw partition at a nonzero offset (EINVAL should have been EXDEV;
DIOCSDINFO was broken, and DIOCWDINFO was broken because it depended
on DIOCSDINFO).
A zero offset for the raw partition should probably be enforced in
setdisklabel(), and DIOCWDINFO should probably always be handled by
first calling setdisklabel() so that writedisklabel() doesn't need to
enforce it, but this has never been done; dsioctl() has a special
check. Changes in this commit are limited to dsioctl() to preserve
bug for bug compatibility in drivers that don't use the slice code
(notably the ccd driver, which allows setting a bogus label in
DIOCWDINFO and doesn't undo the setting when writedisklabel() fails).
Read boot.help before parsing boot.config. We were parsing
boot.config first, which could result in boot.help being read
from a different location (or not found), which would probably
just cause surprise, without being useful.
partition that the label ioctl is being done on just because it has
offset 0, since there is no guarantee that such a partition is large
enough to contain the label. Don't use the wrong raw partition (0
instead of RAW_PART).
This fixes problems rewriting bizarre labels (with a nonzero offset
for the 'a' partition) in newfs(8). Such labels shouldn't normally
be used, but creating them was allowed if the ioctl was done on the
raw partition, and sysinstall creates them if the root partition isn't
allocated first.
Note that allowing write access to a partition other than the one that
has been checked for write access doesn't increase security holes
significantly, since write access to any partition already allows
changing the in-core label.
This fix should be in 3.0R. Rev.1.26 of newfs/newfs.c shouldn't be
in 3.0R.
This is the bulk of the support for doing kld modules. Two linker_sets
were replaced by SYSINIT()'s. VFS's and exec handlers are self registered.
kld is now a superset of lkm. I have converted most of them, they will
follow as a seperate commit as samples.
This all still works as a static a.out kernel using LKM's.
can seek back to the first PT_LOAD and doing a close/reopen if it cannot.
This is because the first PT_LOAD section includes the ELF headers.
This fixes gzipped kernels on the i386, it should solve mike's problem
for the Alpha.
Correct a problem where an external bus reset on the 'background' channel of
a Twin Channel EISA controller could put the driver into an infinite loop.
Noticed by: Twin Channel bug, Joerg Wunsch <joerg@FreeBSD.org>
Submitted by: -Wunused, Poul-Henning Kamp <phk@freebsd.org>
Fixed bug in advansys.c where we did not properly clean up ccb_info data
structures in softc teardown.
Submitted by: Poul-Henning Kamp <phk@freebsd.org>
currently operating in a tagged mode. The SIM driver should determine
if a device is in tag mode by looking at the CAM_TAG_ACTION_VALID flag
in the ccb header. If the flag is set, the tag_action field is either
a SCSI II tag message (simple, ordered, head) or CAM_TAG_ACTION_NONE
to specify that no tagging should be performed.
bogus comment to proper place.
This should fix the hangs people are seeing under very heavy load, at
least it does for me. Please let me know if you continue to have
problems.
tested both in the kernel and in userland. Also, fix a couple of printf
warnings that show up when CAMDEBUG is defined.
Reviewed by: imp
Partially submitted by: imp
hung up when you send tags to them too quickly. (CAM is able to recover
from the problem, but this just avoids it altogether.)
Reviewed by: gibbs
Reported by: Bret Ford <bford@uop.cs.uop.edu>
and: Martin Renters <martin@tdc.on.ca>
release goes out the door. We know there's a bug in the devstat
implementation in the wd driver, but bde and msmith haven't been able to
fix it yet.
So, disable the printf to avoid confusing/worrying people.
Suggested by: msmith
independent elf loader and have access to kld modules. Jordan and I were
not sure how to create boot floppies, and the things we tried just made
SRM laugh in our faces - but it was upset at boot1 which was not touched
by these changes. Essentially this has been untested. :-(
What this does is to steal the last three slots from the nine spare longs
in the bootinfo_v1 struct to pass the module base pointer through.
The startup code now to set up and fills in the module and environment
structures, hopefully close enough to the i386 layout to be able to use
the same kernel code. We now pass though the updated end of the kernel
space used, rather than _end. (like the i386).
If this does not work, it needs to be beaten into shape pronto. Otherwise
it should be backed out before 3.0.
Pre-approved in principle by: dfr
one error recovery action oustanding for a given peripheral.
This is bad for several reasons. The first problem is that the error
recovery actions would likely be to fix the same problem. (e.g., we
queue 5 CCBs to a disk, and the first one comes back with 0x04,0x02. We
start error recovery, and the second one comes back with the same status.
Then the third one comes back, and so on. Each one causes the drive to get
nailed with a start unit, when we really only need one.)
The other problem is that we only have space to store one CCB while we're
doing error recovery. The subsequent error recovery actions that got
started were over-writing the CCBs from previous error recovery actions,
but we still tried to call the done routine N times for N error recovery
actions. Each call to dadone() was done with the same CCB, though. So on
the second one, we got a "biodone: buffer not busy" panic, since the buffer
in question had already been through biodone().
In any case, this fixes things so that any any given time, there's only one
error recovery action outstanding for any given peripheral driver.
Reviewed by: gibbs
Reported by: Philippe Regnauld <regnauld@deepo.prosa.dk>
[ Philippe wins the "bug finder of the week" award ]
sequence of things:
- spin up a disk
- send an async event to refresh the inquiry data
- run through xpt_scan_lun() to re-probe the device
- eventually finish the probe, but panic in xpt_done() because the
periph pointer wasn't set.
Reviewed by: gibbs
Reported by: Philippe Regnauld <regnauld@deepo.prosa.dk>
but when i_effnlink was added to support soft updates, there was only
room for 4 spares. The number of spares was not reduced, so the inode
size became 260 (on i386's), or 512 after rounding up by malloc().
Use one spare field in `struct dinode' instead of the 5th spare field
in the inode and reduced to 4 spares in the inode so that the size is
256 again.
Changed the types of the spares in the inode from int to u_int32_t
so that the inode size has more chance of being <= 256 under other
arches, and downdated ext2fs to match (it was broken to use ints
before rev.1.1).
1) The vnode pager wasn't properly tracking the file size due to
"size" being page rounded in some cases and not in others.
This sometimes resulted in corrupted files. First noticed by
Terry Lambert.
Fixed by changing the "size" pager_alloc parameter to be a 64bit
byte value (as opposed to a 32bit page index) and changing the
pagers and their callers to deal with this properly.
2) Fixed a bogus type cast in round_page() and trunc_page() that
caused some 64bit offsets and sizes to be scrambled. Removing
the cast required adding casts at a few dozen callers.
There may be problems with other bogus casts in close-by
macros. A quick check seemed to indicate that those were okay,
however.
have the passthrough device configured in their kernel.
This will hopefully reduce the number of people complaining that they can't
get {camcontrol, xmcd, tosha, cdrecord, etc.} to work.
Reviewed by: gibbs
things, like msdosfs, do not work (panic) on devices with VMIO enabled.
FFS enable VMIO on mounted devices, and nothing previously disabled it, so,
after you mounted FFS floppy, you could not mount msdosfs floppy anymore...)
This is mostly a quick before-release fix.
Reviewed by: bde
hasseen_isadev so this will be less noisy when conflicts do exist.
Also eliminate redundant warnings about conflicts.
Requested by: bde
Reviewed by: gibbs
2217's (reported by Matthew Jacob in NetBSD PR kern/6027) and Fujitsu
M2954's (reported by Tom Jackson).
Some of the Fujitsus at least hang when they get a cache sync command.
(Others just return illegal request.)
Also, make error printing in dashutdown() a little more selective. Don't
print any error when the sense key is illegal request. Drives that don't
support the synchronize cache command usually return illegal request.
Also, make sure the scsi status is check condition before going into
scsi_sense_print().
Reviewed by: gibbs
specified. This makes haveseen_isadev() useful for searching for a
free resource. This increases the bitrot in the pci RESOURCE_CHECK
code.
Fixed the pre-attach conflict message. The flag for distinguishing
pre-attach conflict checks from pre-probe ones was never set.
Drastically quieten down the verbose load progress messages. They were
more useful for debugging than anything, but are beyond a joke when loading
a few dozen modules.
Simplify the ELF extended symbol table load format. Just take the main
symbol table and the string table that corresponds. This is what we will
be getting local symbols from. (needed for the alpha stack tracebacks).
Use the (optional) full symbol tables in lookups. This means we have to
furhter distinguish between symbols that can come from the dynamic linking
table and the complete table.
The alpha boot code now needs to be adapted as ddb/db_elf.c cannot use
the simpler format.
I have not implemented loading the extended symbol tables from the syscall
interface yet, just for preloaded modules.
I am not sure about the symbol resolution. I *think* it's possible that
a local symbol can be found in preference to a global, depending on the
search sequence and dependency tree.
Change the ELF registration/unregistration scheme to be less error prone.
Adding a new brand requires a single addition to linux_brandlist instead of
modifying linux_load(), linux_unload(), and linux_elf_init().
Approved by: jkh
Reviewed by: msmith
Formerly, the heuristic involving the interpreter path took
precedence.
Also, print a better error message if the brand is missing or not
recognized. If there is no brand at all, give the user a hint that
"brandelf" needs to be run.
Remove /sys/boot from legacy-build.
Add btxld to build-tools.
In src/sys/Makefile:
Add /sys/boot for i386 ELF.
I'm still not sure why the new boot code was being built along with the
legacy stuff, which meant a completely wrong default environment for it.
This may well still be the wrong way to go about this, but it can't work
all that much worse than it has been.
difference, but might later on when we implement some sort of multi-head
console mode. Select a console after probing them all.
Don't strdup a potentially NULL return from getenv().
If we don't select an active console, choose the first regardless.
Call the console init function, at startup time and on a manual change.
The env_setenv() function needs EV_VOLATILE because it's pointing to
data that isn't malloc'ed and will cause a fault if it's freed later.
Cosmetic change to the init-time character eater (like, make it increment
the index counter - if there's a problem, it would sit there in an infinite
loop instead of only running 10 times).
Also, make sure we set %dx each time around otherwise the commands
suddenly start trying to work on things like com92 instead of com1.
Make sure comc_init() is only run once.
Cosmetic change to init-time character eater.
fxp_stop is called as the first thing in fxp_init, and if the tx desc
list has junk in it, the system may panic. This bug showed up as a side
effect of the changes in rev 1.56, but has been in the code since the
beginning.
Linux and FreeBSD signal numbers. Also, check signal numbers passed
in from application programs for validity. Without these checks,
it is trivial to panic the system from a Linux program.
be converted to fixed-sized integers when they are passed across the
binary interface to the kernel.
Didn't fix rotted bits (including not passing dosdev to the kernel and
serious out of dateness when initially committed).
this myself for ages, but wasn't able to get any feedback from the people
that I sent it to for testing.
Guy Helmer <ghelmer@scl.ameslab.gov> has given it a shot (before getting on
a plane, thanks!) and it appears to stop his reproducable page fault panic
in the testing he was able to do.
