02 Jul, 2006
40 commits
-
Add __start_rodata and __end_rodata to sections.h to avoid extern
declarations. Needed by s390 code (see following patch).[akpm@osdl.org: update architectures]
Cc: Arjan van de Ven
Cc: Martin Schwidefsky
Signed-off-by: Heiko Carstens
Cc: Andi Kleen
Acked-by: Kyle McMartin
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
Always use do {} while (0). Failing to do so can cause subtle compile
failures or bugs.Cc: Chandra Seetharaman
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
This is a renamed and tested version of the previous S3C24XX RTC class
driver.The driver has been renamed from the original s3c2410-rtc, which is now too
narrow for the range of devices supported.The rtc-s3c has been chosen as there is the distinct possibility of this
driver being carried forward into newer Samsung SoC silicon.Signed-off-by: Ben Dooks
Cc: Alessandro Zummo
Cc: Russell King
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
o Raise the maximum error number in IS_ERR_VALUE to 4095.
o Make that number available as a new constant MAX_ERRNO.Signed-off-by: Ralf Baechle
Cc: "H. Peter Anvin"
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
This patch fixes buggy behaviour of UFS
in such kind of scenario:
open(, O_TRUNC...)
ftruncate(, 1024)
ftruncate(, 0)Such a scenario causes ufs_panic and remount read-only. This happen
because of according to specification UFS should always allocate block for
last byte, and many parts of our implementation rely on this, but
`ufs_truncate' doesn't care about this.To make possible return error code and to know about old size, this patch
removes `truncate' from ufs inode_operations and uses `setattr' method to
call ufs_truncate.Signed-off-by: Evgeniy Dushistov
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
Rename actually_do_remove() to remove_files_and_dir(), make it call
closedir(), make it ignore ENOENT (I see it frequently enough).ENOENT is probably due to multiple threads calling the exitcall functions
together*, but fixing that is non-trivial; and ignoring it is perfectly ok
in any case.* it can surely happen: last_ditch_exit() is installed as SIGTERM handler
at boot, and it's not removed on thread creation. So killall vmlinux
(which I do) surely causes that. I've seen also a crash which seems to
do the same.Installing the handler on only the main thread would make UML do no cleanup
when another thread exits, and we're not sure we want that. And mutual
exclusion in that context is tricky - we can't use spinlock in code not on
a kernel stack (spinlock debugging uses "current" a lot).Signed-off-by: Paolo 'Blaisorblade' Giarrusso
Cc: Jeff Dike
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
The bug occurred to me when a UML left an empty ~/.uml/Sarge-norm folder -
when trying to reuse not_dead_yet() failed one of its check. The comment
says that's ok and means that we can take the directory, but while normally
not_dead_yet() removes it and returns 0 (i.e. go on, use this), on failure
it returns 0 but forgets to remove it. The fix is to remove it anytime
we're going to return 0.But since "not_dead_yet" didn't make the interface so clear, causing this
bug, and I couldn't find a convenient name for the mix of things it did, I
split it into two parts:is_umdir_used() - returns a boolean, contains all checks of not_dead_yet()
umdir_take_if_dead - tries to remove the dir unless it's used - returns
whether it removed it, that is we now own it.With this changes the control flow is IMHO a bit clearer and needs less
comment for control flow.Signed-off-by: Paolo 'Blaisorblade' Giarrusso
Cc: Jeff Dike
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
Make __copy_*_user_inatomic really atomic to avoid "Sleeping function called in
atomic context" warnings, especially from futex code.This is made by adding another kmap_atomic slot and making copy_*_user_skas
use kmap_atomic; also copy_*_user() becomes atomic, but that's true and is not
a problem for i386 (and we can always add might_sleep there as done
elsewhere). For TT mode kmap is not used, so there's no need for this.I've had to use another slot since both KM_USER0 and KM_USER1 are used
elsewhere and could cause conflicts. Till now we reused the kmap_atomic slot
list from the subarch, but that's not needed as that list must contain the
common ones (used by generic code) + the ones used in architecture specific
code (and Uml till now used none); so I've taken the i386 one after comparing
it with ones from other archs, and added KM_UML_USERCOPY.Signed-off-by: Paolo 'Blaisorblade' Giarrusso
Cc: Jeff Dike
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
Hide the magic in alternative.h and provide some dummy inline functions
for the UP case (gcc should manage to optimize away these calls). No
changes in module.c.Cc: Dave Jones
Cc: Andi Kleen
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
Remove ips namespace from infinipath drivers. This renames ips_common.h to
ipath_common.h. Definitions, data structures, etc. that were not used by
kernel modules have moved to user-only headers. All names including ips have
been renamed to ipath. Some names have had an ipath prefix added.Signed-off-by: Christian Bell
Signed-off-by: Bryan O'Sullivan
Cc: "Michael S. Tsirkin"
Cc: Roland Dreier
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
The receive work queue size should be ignored if the QP is created to use a
shared receive queue according to the IB spec.Signed-off-by: Ralph Campbell
Signed-off-by: Bryan O'Sullivan
Cc: "Michael S. Tsirkin"
Cc: Roland Dreier
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
Signed-off-by: Dave Olson
Signed-off-by: Bryan O'Sullivan
Cc: "Michael S. Tsirkin"
Cc: Roland Dreier
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
We can't tell for sure if any packets are in the infinipath receive buffer
when we shut down a chip port. Normally this is taken care of by orderly
shutdown, but when processes are terminated, or sending process has a bug, we
can continue to receive packets. So rather than writing zero to the address
registers for the closing port, we point it at a dummy memory.Signed-off-by: Dave Olson
Signed-off-by: Bryan O'Sullivan
Cc: "Michael S. Tsirkin"
Cc: Roland Dreier
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
We must increment uaddr by size we are reading or writing, since it's passed
as a char *, not a pointer to the appropriate size.Signed-off-by: Dave Olson
Signed-off-by: Bryan O'Sullivan
Cc: "Michael S. Tsirkin"
Cc: Roland Dreier
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
We do a few more explicit checks for specific models, and now also support the
old PathScale serial number style, or new QLogic style.This is backwards compatible with previous versions of software and hardware.
That is, older software will see a plausible serial number and correct GUID
when used with a new board, while newer software will correctly handle an
older board.Signed-off-by: Mike Albaugh
Signed-off-by: Dave Olson
Signed-off-by: Bryan O'Sullivan
Cc: "Michael S. Tsirkin"
Cc: Roland Dreier
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
This attribute group made it into the original driver, but should not have.
Signed-off-by: Bryan O'Sullivan
Cc: "Michael S. Tsirkin"
Cc: Roland Dreier
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
The two arrays only had space for 4 units.
Also changed from ipath_set_sps_lid() to ipath_set_lid(); the sps was
leftover.Signed-off-by: Dave Olson
Signed-off-by: Bryan O'Sullivan
Cc: "Michael S. Tsirkin"
Cc: Roland Dreier
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
This patch separates QP state used for sending and receiving RC packets so the
processing in the receive interrupt handler can be done mostly without locks
being held. ACK packets are now sent without requiring synchronization with
the send tasklet.Signed-off-by: Ralph Campbell
Signed-off-by: Bryan O'Sullivan
Cc: "Michael S. Tsirkin"
Cc: Roland Dreier
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
This patch fixes some problems uncovered during IB compliance testing to
return the right values for error counters returned by the Performance Get
Counters packet.Signed-off-by: Ralph Campbell
Signed-off-by: Bryan O'Sullivan
Cc: "Michael S. Tsirkin"
Cc: Roland Dreier
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
Signed-off-by: Ralph Campbell
Signed-off-by: Bryan O'Sullivan
Cc: "Michael S. Tsirkin"
Cc: Roland Dreier
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
The tail register read became redundant as the result of earlier receive
interrupt bug fixes.Drop another unneeded register read.
And another line that got duplicated.
Signed-off-by: Dave Olson
Signed-off-by: Bryan O'Sullivan
Cc: "Michael S. Tsirkin"
Cc: Roland Dreier
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
Signed-off-by: Ralph Campbell
Signed-off-by: Bryan O'Sullivan
Cc: "Michael S. Tsirkin"
Cc: Roland Dreier
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
Signed-off-by: Ralph Campbell
Signed-off-by: Bryan O'Sullivan
Cc: "Michael S. Tsirkin"
Cc: Roland Dreier
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
Do an extra check to see if in-memory tail changed while processing packets,
and if so, going back through the loop again (but only once per call to
ipath_kreceive()). In practice, this seems to be enough to guarantee that if
we crossed the clearing of an interrupt at start of ipath_intr with a
scheduled tail register update, that we'll process the "extra" packet that
lost the interrupt because we cleared it just as it was about to arrive.Signed-off-by: Dave Olson
Signed-off-by: Bryan O'Sullivan
Cc: "Michael S. Tsirkin"
Cc: Roland Dreier
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
The problem was that I was updating the head register multiple times in the
rcvhdrq processing loop, and setting the counter on each update. Since that
meant that the tail register was ahead of head for all but the last update, we
would get extra interrupts. The fix was to not write the counter value except
on the last update.I also changed to update rcvhdrhead and rcvegrindexhead at most every 16
packets, if there were lots of packets in the queue (and of course, on the
last packet, regardless).I also made some small cleanups while debugging this.
With these changes, xeon/monty typically sees two openib packets per interrupt
on sdp and ipoib, opteron/monty is about 1.25 pkts/intr.I'm seeing about 3800 Mbit/s monty/xeon, and 5000-5100 opteron/monty with
netperf sdp. Netpipe doesn't show as good as that, peaking at about 4400 on
opteron/monty sdp. Plain ipoib xeon is about 2100+ netperf, opteron 2900+, at
128KBSigned-off-by: olson@eng-12.pathscale.com
Signed-off-by: Bryan O'Sullivan
Cc: "Michael S. Tsirkin"
Cc: Roland Dreier
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
Also count the number of interrupts where that works (fastrcvint). On any
interrupt where the port0 head and tail registers are not equal, just call the
ipath_kreceive code without reading the interrupt status, thus saving the
approximately 0.25usec processor stall waiting for the read to return. If any
other interrupt bits are set, or head==tail, take the normal path, but that
has been reordered to handle read ahead of pioavail. Also no longer call
ipath_kreceive() from ipath_qcheck(), because that just seems to make things
worse, and isn't really buying us anything, these days.Also no longer loop in ipath_kreceive(); better to not hold things off too
long (I saw many cases where we would loop 4-8 times, and handle thousands (up
to 3500) in a single call).Signed-off-by: Dave Olson
Signed-off-by: Bryan O'Sullivan
Cc: "Michael S. Tsirkin"
Cc: Roland Dreier
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
Made in-memory rcvhdrq tail update be in dma_alloc'ed memory, not random user
or special kernel (needed for ppc, also "just the right thing to do").Some cleanups to make unexpected link transitions less likely to produce
complaints about packet errors, and also to not leave SMA packets stuck and
unable to go out.A few other random debug and comment cleanups.
Always init rcvhdrq head/tail registers to 0, to avoid race conditions (should
have been that way some time ago).Signed-off-by: Dave Olson
Signed-off-by: Bryan O'Sullivan
Cc: "Michael S. Tsirkin"
Cc: Roland Dreier
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
This is not a DMA target, so no need to use dma_alloc_coherent on it.
Signed-off-by: Bryan O'Sullivan
Cc: "Michael S. Tsirkin"
Cc: Roland Dreier
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
This helps us to survive better when memory is fragmented.
Signed-off-by: Dave Olson
Signed-off-by: Bryan O'Sullivan
Cc: "Michael S. Tsirkin"
Cc: Roland Dreier
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
Signed-off-by: Dave Olson
Signed-off-by: Bryan O'Sullivan
Cc: "Michael S. Tsirkin"
Cc: Roland Dreier
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
Signed-off-by: Dave Olson
Signed-off-by: Bryan O'Sullivan
Cc: "Michael S. Tsirkin"
Cc: Roland Dreier
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
Signed-off-by: Dave Olson
Signed-off-by: Bryan O'Sullivan
Cc: "Michael S. Tsirkin"
Cc: Roland Dreier
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
These limits are somewhat artificial in that we don't actually have any
device limits. However, the verbs layer expects that such limits exist
and are enforced, so we make up arbitrary (but sensible) limits.Signed-off-by: Robert Walsh
Signed-off-by: Bryan O'Sullivan
Cc: "Michael S. Tsirkin"
Cc: Roland Dreier
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
Signed-off-by: Robert Walsh
Signed-off-by: Bryan O'Sullivan
Cc: "Michael S. Tsirkin"
Cc: Roland Dreier
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
Signed-off-by: Robert Walsh
Signed-off-by: Bryan O'Sullivan
Cc: "Michael S. Tsirkin"
Cc: Roland Dreier
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
Signed-off-by: Robert Walsh
Signed-off-by: Bryan O'Sullivan
Cc: "Michael S. Tsirkin"
Cc: Roland Dreier
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
Signed-off-by: Robert Walsh
Signed-off-by: Bryan O'Sullivan
Cc: "Michael S. Tsirkin"
Cc: Roland Dreier
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
Signed-off-by: Robert Walsh
Signed-off-by: Bryan O'Sullivan
Cc: "Michael S. Tsirkin"
Cc: Roland Dreier
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
Signed-off-by: Robert Walsh
Signed-off-by: Bryan O'Sullivan
Cc: "Michael S. Tsirkin"
Cc: Roland Dreier
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
There is no longer a /dev/ipath_diag file; instead, there's
/dev/ipath_diag0, 1, etc.It's still not possible to have diags run on more than one unit at a time,
but that's easy to fix at some point.Signed-off-by: Bryan O'Sullivan
Cc: "Michael S. Tsirkin"
Cc: Roland Dreier
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds
