ed [Fri, 19 Sep 2008 10:21:30 +0000 (10:21 +0000)]
Even though snp(4) in SVN is still broken, remove the unneeded D_NEEDMINOR.
kib@ and I have decided we will MFC the bpf(4)/snp(4) fixes after we've
released 7.1. Make sure the code in HEAD doesn't refer to a flag we
don't need anyway.
snp(4) in the MPSAFE TTY P4 branch already works, but still needs some
polishing before it can be integrated to SVN.
fix a another driver bug related with tsleep() during detaching that
this also can be happened if we pull the USN stick out forcibly.
Currently the ZyDAS driver uses tsleep() when it try to query a read
command to the device and it'd make a timeout if the device doesn't
response within about 1 sec.
In a case of that the USB stick is gone by hand and the driver's
scanning with changing the channel numbers, the thread which is sleeping
until a command requested is responded can be waked up after all
detaching routines finished that means the zyd softc already freed.
Tring to touch the softc freed by the wakeup thread makes a panic.
So make sure that all sleeping threads should be waken up before the
detach is completed and any other new requests to the device should be
prevented.
When attaching framebuffer to "/chosen/stdout" node fails, try attaching
to "screen" node directly. The problem is that by default OF on some (all?)
Macs either doesn't provide "/chosen/stdout" or redirects it somewhere,
unless you boot in manual mode via CMD-ALT-O-F. It's nice to see normal
FreeBSD boot output instead of blank gray screen.
marius [Thu, 18 Sep 2008 19:45:22 +0000 (19:45 +0000)]
Clear any possibly pending PCI error bits left by the firmware. These
could trigger an error interrupt that we can't actually to do anything
against as soon as enabling the error handlers.
While at it don't bother about writing only to the write-one-to-clear
bits when clearing error bits.
- Make bsdlabel use libgeom to determine provider name, device path, the media
size and the sector size.
- Fix a bug where bsdlabel would try to read a regular file using the geom_bsd
class.
marius [Thu, 18 Sep 2008 13:56:30 +0000 (13:56 +0000)]
- Newer firmware versions no longer provide SUNW,stop-self so just
disable interrupts and loop forever with these.
- Hide all MP-related bits in <machine/smp.h> underneath #ifdef SMP.
- Inline ipi_all_but_self(9) and ipi_selected(9). We don't expose any
additional bits but save a few cycles by doing so.
- Remove ipi_all(9), which actually only called panic(9). It can't be
implemented natively anyway and having it removed at least causes
MI users to fail already fail when linking.
Some people have very strange notions of how large KVA_PAGES should be. The
core of this change generalizes the initial page directory setup so that
the kernel can be given arbitrarily large or small.
Remove the suser(9) interface from the kernel. It has been replaced from
years by the priv_check(9) interface and just very few places are left.
Note that compatibility stub with older FreeBSD version
(all above the 8 limit though) are left in order to reduce diffs against
old versions. It is responsibility of the maintainers for any module, if
they think it is the case, to axe out such cases.
This patch breaks KPI so __FreeBSD_version will be bumped into a later
commit.
This patch needs to be credited 50-50 with rwatson@ as he found time to
explain me how the priv_check() works in detail and to review patches.
Tested by: Giovanni Trematerra <giovanni dot trematerra at gmail dot com>
Reviewed by: rwatson
Decontext-alize the nfsserver module.
Now, only some few places still require thread passing (mostly the ones which
access to VOP_* functions) and will be fixed once the primitive also will be.
Tested by: Giovanni Trematerra <giovanni dot trematerra at gmail dot com>
Implement zero-copy bpf(4) buffer or "zbuf" support for libpcap. A slightly
different version has been committed upstream in the libpcap vendor branch.
This will allow people to experiment with zero-copy bpf(4) without requiring
external patches.
Note to enable this functionality:
sysctl net.bpf.zerocopy_enable=1
By default, libpcap will use the legacy buffering method unless this sysctl
variable is set to 1.
For the details about zero-copy bpf(4) implementation see svn change r177548.
Requested by: many
Discussed with: sam
In collaboration with: rwatson
Add 'device acpi' to GENERIC on i386. Most newer i386 machines require it,
so the benefit of having acpi.ko as a standalone module is outweighed by
the complications of drivers compiled into the kernel not including ACPI
attachments by default.
Left only parts surely required for basic troubleshooting and configuration
purposes. There is still very long output, but further shrinking makes it
less informative.
Original debugging can be enabled with hw.snd.verbose=4.
Allow child devices of vgapci(4) to query VPD strings and use MSI/MSI-X
interrupts. For the MSI/MSI-X case, we only allow 1 child device to use
MSI or MSI-X at a time.
o When not making a translation cache-inhibit and guarded (PTE_I|PTE_G)
make it memory-coherency enforced (PTE_M). This is required for SMP
to work.
o Serialize tlbie operations and implement the tlbie operation in a
function called tlbie(). Hardware can end up in a live-lock if
between the tlbsync and subsequent sync on one processor another
processor executes a tlbie or tlbsync.
o Eliminate the following defines:
TLBIE, TLBSYNC, SYNC and EIEIO
Use either inline assembly statements or inline functions defined
in <machine/cpufunc.h>
Retire the 'i_reclen' field from the in-memory i-node. Previously,
during a DELETE lookup operation, lookup would cache the length of the
directory entry to be deleted in 'i_reclen'. Later, the actual VOP to
remove the directory entry (ufs_remove, ufs_rename, etc.) would call
ufs_dirremove() which extended the length of the previous directory
entry to "remove" the deleted entry.
However, we always read the entire block containing the directory
entry when doing the removal, so we always have the directory entry to
be deleted in-memory when doing the update to the directory block.
Also, we already have to figure out where the directory entry that is
being removed is in the block so that we can pass the component name
to the dirhash code to update the dirhash. So, instead of passing
'i_reclen' from ufs_lookup() to the ufs_dirremove() routine, just read
the 'd_reclen' field directly out of the entry being removed when
updating the length of the previous entry in the block.
This avoids a cosmetic issue of writing to 'i_reclen' while holding a
shared vnode lock. It also slightly reduces the amount of side-band
data passed from ufs_lookup() to operations updating a directory via
the directory's i-node.
Bump __FreeBSD_version to reflect that kernel NFS mount code can accept
mount options in individual members of nmount() iovec array instead of
one big struct nfs_args.
Rewrite cpudep_ap_bootstrap(). We now enable L3, L2, L1D and L1I
caches if not yet enabed. This is required for coherency and
atomic operations to work, not to mention performance. We use the
L2 and L3 cache settings of the BSP to configure the APs caches.
Can't be bad.
Program NAP and not DOZE. DOZE is present only on earlier CPUs
and the bit is reserved on the MPC7441 & MPC7451. NAP will do
bus snooping to keep caches coherent.
Program the PIR with the cpuid. This may not be necessary...
o In decr_get_timecount() only read the low timebase register.
We're only returning a 32-bit counter.
o In decr_intr(), manually perform LICM, so that we don't test
a loop invariant condition inside a loop.
o Include <machine/smp.h>
Set pcpup->pc_curthread and pcpup->pc_curpcb before calling
pmap_activate. While pmap_activate doesn't need either, we
do need a valid curthread if we enable KTR_PMAP.
o Synchronize the APs timebase and decrementer values with the BSP.
o Don't set/get the PIR register. It's CPU dependent.
o Also initialize pcpup->pc_curpcb, in case it's dereferenced.
In powerpc_get_pcpup(), make the inline assembly statement
volatile so that the compiler won't perform CSE. For SMP,
this may result in us accessing the wrong PCPU and as such
results in a bogus curthread value.
Note that getting curthread is not quite MP-safe in the sense
that it requires two instructions that aren't performed
atomically. The first instruction gets the address of the PCPU
structure and the second instruction dereferences that pointer
to get curthread. If a thread is switched-out in between these
instructions and switched-in on a different CPU, we still get
the wrong curthread.
Fix a race with shared lookups on UFS. If the the dirhash code reached the
cap on memory usage, then shared LOOKUP operations could start free'ing
dirhash structures. Without these fixes, concurrent free's on the same
directory could result in one of the threads blocked on a lock in a dirhash
structure free'd by the other thread.
- Replace the lockmgr lock in the dirhash structure with an sx lock.
- Use a reference count managed with ufsdirhash_hold()/drop() to determine
when to free the dirhash structures. The directory i-node holds a
reference while the dirhash is attached to an i-node. Code that wishes
to lock the dirhash while holding a shared vnode lock must first
acquire a private reference to the dirhash while holding the vnode
interlock before acquiring the dirhash sx lock. After acquiring the sx
lock, it drops the private reference after checking to see if the
dirhash is still used by the directory i-node.
- Only set i_offset in the parent directory's i-node during a lookup for
non-LOOKUP operations.
- Relax a VOP assertion for a DELETE lookup. rename() uses WANTPARENT
instead of LOCKPARENT when looking up the source pathname. ufs_rename()
uses a relookup() to lock the parent directory when it decides to finally
remove the source path. Thus, it is ok for a DELETE with WANTPARENT set
instead of LOCKPARENT to use a shared vnode lock rather than an exclusive
vnode lock.
vdropl() drops the vnode interlock. Thus, the code in the QUOTA case that
upgrades the vnode lock if it is share locked was dropping the interlock
before actually checking VI_DOOMED. Fix this by do the vdropl() after the
check and relying on it to drop the vnode interlock.
ed [Tue, 16 Sep 2008 14:57:23 +0000 (14:57 +0000)]
Fix minor TTY API inconsistency.
Unlike tty_rel_gone() and tty_rel_sess(), the tty_rel_pgrp() routine
does not unlock the TTY. I once had the idea to make the code call
tty_rel_pgrp() and tty_rel_sess(), picking up the TTY lock once. This
turned out a little harder than I expected, so this is how it works now.
It's a lot easier if we just let tty_rel_pgrp() unlock the TTY, because
the other routines do this anyway.
When attempt is made to suspend a filesystem that is already syspended,
wait until the current suspension is lifted instead of silently returning
success immediately. The consequences of calling vfs_write() resume when
not owning the suspension are not well-defined at best.
Add the vfs_susp_clean() mount method to be called from
vfs_write_resume(). Set it to process_deferred_inactive() for ffs, and
stop calling it manually.
Add the thread flag TDP_IGNSUSP that allows to bypass the suspension
point in the vn_start_write. It is intended for use by VFS in the
situations where the suspender want to do some i/o requiring calls to
vn_start_write(), and this i/o cannot be done later.
Add the ffs structures introspection functions for ddb.
Show the b_dep value for the buffer in the show buffer command.
Add a comand to dump the dirty/clean buffer list for vnode.
Reviewed by: tegge
Tested and used by: pho
MFC after: 1 month
When downgrading the read-write mount to read-only, do_unmount() sets
MNT_RDONLY flag before the VFS_MOUNT() is called. In ufs_inactive()
and ufs_itimes_locked(), UFS verifies whether the fs is read-only by
checking MNT_RDONLY, but this may cause loss of the IN_MODIFIED flag
for inode on the fs being remounted rw->ro.
Introduce UFS_RDONLY() struct ufsmount' method that reports the value
of the fs_ronly. The later is set to 1 only after the remount is
finished.
The struct inode *ip supplied to softdep_freefile is not neccessary the
inode having number ino. In r170991, the ip was marked IN_MODIFIED, that
is not quite correct.
Mark only the right inode modified by checking inode number.
Allow multiple locks to be acquired by detecting corresponding
bit flag, otherwise if a thread acquired a lock, another thread
or the current thread itself can no longer acquire another lock
because thread_mask_set() return whole flag word, this results
bit leaking in the word and misbehavior in later locking and
unlocking.
Remove the tracing from the AP startup. The AP is known
to start and the tracing can interfere with AP startup.
Instead, use the available space in the reset vector
for the initial stack.
sam [Mon, 15 Sep 2008 22:45:14 +0000 (22:45 +0000)]
Make ddb command registration dynamic so modules can extend
the command set (only so long as the module is present):
o add db_command_register and db_command_unregister to add and remove
commands, respectively
o replace linker sets with SYSINIT's (and SYSUINIT's) that register
commands
o expose 3 list heads: db_cmd_table, db_show_table, and db_show_all_table
for registering top-level commands, show operands, and show all operands,
respectively
While here also:
o sort command lists
o add DB_ALIAS, DB_SHOW_ALIAS, and DB_SHOW_ALL_ALIAS to add aliases
for existing commands
o add "show all trace" as an alias for "show alltrace"
o add "show all locks" as an alias for "show alllocks"
Rework the handling of interrupt handlers for children of ppc and ppbus:
- Retire IVARs for passing IRQs around. Instead, ppbus and ppc now allow
child devices to access the interrupt by via a rid 0 IRQ resource
using bus_alloc_resource_any().
- ppc creates its own interrupt event to manage the interrupt handlers of
child devices. ppc does not allow child devices to use filters. It
could allow this if needed, but none of the current drivers use them
and it adds a good bit of complication. It uses
intr_event_execute_handlers() to fire the child device interrupt handlers
from its threaded interrupt handler.
- Remove the ppbus_dummy_intr() hack. Now the ppc device always has an
interrupt handler registered and we no longer bounce all the way up to
nexus to manage adding/removing ppbus child interrupt handlers. Instead,
the child handlers are added and removed to the private interrupt event
in the ppc device.
Expose a new public routine intr_event_execute_handlers() which executes
all the non-filter handlers attached to an interrupt event. This can be
used by device drivers which multiplex their interrupt onto the interrupt
handlers for child devices.