| Age | Commit message (Collapse) | Author |
|---|
|
BKL was not really needed, just came from earlier push downs.
The only part that's a bit dodgy is the lseek function. Would
need another lock or atomic access to fpos on 32bit?
Better to have a libfs lseek
Signed-off-by: Andi Kleen <ak@linux.intel.com>
Cc: Arnd Bergmann <arnd@arndb.de>
Cc: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
|
|
implicit slab.h inclusion from percpu.h
percpu.h is included by sched.h and module.h and thus ends up being
included when building most .c files. percpu.h includes slab.h which
in turn includes gfp.h making everything defined by the two files
universally available and complicating inclusion dependencies.
percpu.h -> slab.h dependency is about to be removed. Prepare for
this change by updating users of gfp and slab facilities include those
headers directly instead of assuming availability. As this conversion
needs to touch large number of source files, the following script is
used as the basis of conversion.
http://userweb.kernel.org/~tj/misc/slabh-sweep.py
The script does the followings.
* Scan files for gfp and slab usages and update includes such that
only the necessary includes are there. ie. if only gfp is used,
gfp.h, if slab is used, slab.h.
* When the script inserts a new include, it looks at the include
blocks and try to put the new include such that its order conforms
to its surrounding. It's put in the include block which contains
core kernel includes, in the same order that the rest are ordered -
alphabetical, Christmas tree, rev-Xmas-tree or at the end if there
doesn't seem to be any matching order.
* If the script can't find a place to put a new include (mostly
because the file doesn't have fitting include block), it prints out
an error message indicating which .h file needs to be added to the
file.
The conversion was done in the following steps.
1. The initial automatic conversion of all .c files updated slightly
over 4000 files, deleting around 700 includes and adding ~480 gfp.h
and ~3000 slab.h inclusions. The script emitted errors for ~400
files.
2. Each error was manually checked. Some didn't need the inclusion,
some needed manual addition while adding it to implementation .h or
embedding .c file was more appropriate for others. This step added
inclusions to around 150 files.
3. The script was run again and the output was compared to the edits
from #2 to make sure no file was left behind.
4. Several build tests were done and a couple of problems were fixed.
e.g. lib/decompress_*.c used malloc/free() wrappers around slab
APIs requiring slab.h to be added manually.
5. The script was run on all .h files but without automatically
editing them as sprinkling gfp.h and slab.h inclusions around .h
files could easily lead to inclusion dependency hell. Most gfp.h
inclusion directives were ignored as stuff from gfp.h was usually
wildly available and often used in preprocessor macros. Each
slab.h inclusion directive was examined and added manually as
necessary.
6. percpu.h was updated not to include slab.h.
7. Build test were done on the following configurations and failures
were fixed. CONFIG_GCOV_KERNEL was turned off for all tests (as my
distributed build env didn't work with gcov compiles) and a few
more options had to be turned off depending on archs to make things
build (like ipr on powerpc/64 which failed due to missing writeq).
* x86 and x86_64 UP and SMP allmodconfig and a custom test config.
* powerpc and powerpc64 SMP allmodconfig
* sparc and sparc64 SMP allmodconfig
* ia64 SMP allmodconfig
* s390 SMP allmodconfig
* alpha SMP allmodconfig
* um on x86_64 SMP allmodconfig
8. percpu.h modifications were reverted so that it could be applied as
a separate patch and serve as bisection point.
Given the fact that I had only a couple of failures from tests on step
6, I'm fairly confident about the coverage of this conversion patch.
If there is a breakage, it's likely to be something in one of the arch
headers which should be easily discoverable easily on most builds of
the specific arch.
Signed-off-by: Tejun Heo <tj@kernel.org>
Guess-its-ok-by: Christoph Lameter <cl@linux-foundation.org>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Lee Schermerhorn <Lee.Schermerhorn@hp.com>
|
|
The code in uhci-q.c doesn't have to use pseudo-negative values. I did
it that way because it was easy and because it would give the expected
output during debugging. But it doesn't have to work that way. Here's
another approach.
Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Cc: Roel Kluin <roel.kluin@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
|
|
This patch (as970) adds a new urb->unlinked field, which is used to
store the status of unlinked URBs since we can't use urb->status for
that purpose any more. To help simplify the HCDs, usbcore will check
urb->unlinked before calling the completion handler; if the value is
set it will automatically override the status reported by the HCD.
Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
CC: David Brownell <david-b@pacbell.net>
CC: Olav Kongas <ok@artecdesign.ee>
CC: Yoshihiro Shimoda <shimoda.yoshihiro@renesas.com>
CC: Tony Olech <tony.olech@elandigitalsystems.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
|
|
This patch (as871) fixes a problem introduced by an earlier change.
It turns out that some systems really do need to have a terminating
skeleton QH present whenever FSBR is on. I don't know any way to tell
which systems do need it and which don't; the easiest answer is to
have it there always.
This fixes the NumLock-hang bug reported by Jiri Slaby.
Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
|
|
This patch (as856) attempts to improve the performance of uhci-hcd by
removing the asynchronous skeleton Queue Headers. They don't contain
any useful information but the controller has to read through them at
least once every millisecond, incurring a non-zero DMA overhead.
Now all the asynchronous queues are combined, along with the period-1
interrupt queue, into a single list with a single skeleton QH. The
start of the low-speed control, full-speed control, and bulk sublists
is determined by linear search. Since there should rarely be more
than a couple of QHs in the list, the searches should incur a much
smaller total load than keeping the skeleton QHs.
Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
|
|
This patch (as855) adds some convenience macros to uhci-hcd, to help
simplify the code for computing hardware DMA pointers.
Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
|
|
This patch (as840) fixes the bandwidth allocation mechanism in
uhci-hcd. It has never worked correctly.
Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
|
|
Isochronous queues don't need a dummy TD because the Queue Header
isn't managed by the hardware. This patch (as836) removes the
unnecessary dummy TDs.
The patch also fixes a long-standing typo in a comment (a "don't" was
missing -- potentially very confusing!).
Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
|
|
This patch (as768) improves the debugging checks for the uhci-hcd
frame list. The number of entries displayed is limited to 10, and the
driver now checks for the correct Skeleton QH link value at the end of
each chain of Isochronous TDs. The code to compute these link values
is now used in two spots, so it is moved into its own separate
subroutine.
Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
|
|
Making structs const prevents accidental bugs and with the proper debug
options they're protected against corruption.
Signed-off-by: Luiz Fernando N. Capitulino <lcapitulino@mandriva.com.br>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
|
|
The following patches reduce the size of the VFS inode structure by 28 bytes
on a UP x86. (It would be more on an x86_64 system). This is a 10% reduction
in the inode size on a UP kernel that is configured in a production mode
(i.e., with no spinlock or other debugging functions enabled; if you want to
save memory taken up by in-core inodes, the first thing you should do is
disable the debugging options; they are responsible for a huge amount of bloat
in the VFS inode structure).
This patch:
The filesystem or device-specific pointer in the inode is inside a union,
which is pretty pointless given that all 30+ users of this field have been
using the void pointer. Get rid of the union and rename it to i_private, with
a comment to explain who is allowed to use the void pointer. This is just a
cleanup, but it allows us to reuse the union 'u' for something something where
the union will actually be used.
[judith@osdl.org: powerpc build fix]
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
Signed-off-by: Judith Lebzelter <judith@osdl.org>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
|
|
Signed-off-by: Jörn Engel <joern@wohnheim.fh-wedel.de>
Signed-off-by: Adrian Bunk <bunk@stusta.de>
|
|
This patch (as690) does the same thing for ISO TDs as as680 did for
non-ISO TDs: free them as they are used rather than all at once when an
URB is complete. At the same time it fixes a minor buglet (I'm not
aware of it ever affecting anyone): An ISO TD should be retired when its
frame is over, regardless of whether or not the hardware has marked it
inactive.
Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
|
|
This patch (as689) stores the period for periodic transfers (interrupt
and ISO) in the queue header. This is necessary for proper bandwidth
tracking (not yet implemented). It also makes the scheduling of ISO
transfers a bit more rigorous, with checks for out-of-bounds frame
numbers.
Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
|
|
This patch (as687) changes uhci-hcd to keep track of frame numbers as
full-sized integers rather than 11-bit values. This makes them a lot
easier to handle and makes it possible to schedule beyond a 2-second
window, should anyone ever want to do so.
Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
|
|
This patch (as683) re-implements Full-Speed Bandwidth Reclamation (FSBR)
properly. It keeps track of which endpoint queues have advanced, and
when none have advanced for a sufficiently long time, FSBR is turned
off. The next TD on each of the non-moving queues is modified to
generate an interrupt on completion, so that FSBR can be re-enabled as
soon as the hardware starts to make some progress.
Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
|
|
This patch (as680) frees non-isochronous TDs as they are used, rather
than all at once when an URB is complete. Although not a terribly
important change in itself, it opens the door to a later enhancement
that will reduce storage requirements by allocating only a limited
number of TDs at any time for each endpoint queue.
Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
|
|
This patch (as675) simplifies uhci-hcd slightly by storing each endpoint's
type in the corresponding Queue Header structure.
Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
|
|
This patch (as626) makes some improvements to the debugging code in
uhci-hcd. The main change is that now the code won't get compiled if
CONFIG_USB_DEBUG isn't set. But there are other changes too, like
adding a missing .owner field and printing a debugging dump if the
controller dies.
Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
|
|
As part of reorienting uhci-hcd away from URBs and toward endpoint
queues, this patch (as625) eliminates the driver's main list of URBs.
The list wsa used mainly in checking for URB completions; now the driver
goes through the list of active endpoints and checks the members of the
queues.
As a side effect, I had to remove the code that looks for FSBR timeouts.
For now, FSBR will remain on so long as any URBs on a full-speed control
or bulk queue request it, even if the queue isn't advancing. A later
patch can add more intelligent handling. This isn't a huge drawback;
it's pretty rare for an URB to get stuck for more than a fraction of a
second. (And it will help the people trying to use those insane HP USB
devices.)
Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
|
|
This patch (as624) fixes a hardware race in uhci-hcd by adding a dummy
TD to the end of each endpoint's queue. Without the dummy the host
controller will effectively turn off the queue when it reaches the end,
which happens asynchronously. This leads to a potential problem when
new transfer descriptors are added to the end of the queue; they may
never get used.
With a dummy TD present the controller never turns off the queue;
instead it just stops at the dummy and leaves the queue on but inactive.
When new TDs are added to the end of the queue, the first new one gets
written over the dummy. Thus there's never any question about whether
the queue is running or needs to be restarted.
Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
|
|
This patch (as623) changes the uhci-hcd driver to make it use one QH per
device endpoint, instead of a QH per URB as it does now. Numerous areas
of the code are affected by this. For example, the distinction between
"queued" URBs and non-"queued" URBs no longer exists; all URBs belong to
a queue and some just happen to be at the queue's head.
Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
|
|
This patch (as615b) edits a large number of comments in the uhci-hcd code,
mainly removing excess apostrophes.
Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
|
|
patch below marks various USB tables and variables as const so that they
end up in .rodata section and don't cacheline share with things that get
written to. For the non-array variables it also allows gcc to optimize
more.
Signed-off-by: Arjan van de Ven <arjan@infradead.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
|
|
This patch (as563) splits the physical and logical framelist arrays in
uhci-hcd into two separate pieces. This will allow slightly better memory
utilization, since each piece is no larger than a single page whereas
before the whole thing was a little bigger than two pages. It also allows
the logical array to be allocated in non-DMA-coherent memory.
Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
|
|
This patch (as562) removes from the uhci-hcd driver a few unused fields
and some unnecessary tests against NULL and assignments to NULL. In fact
it wasn't until fairly recently that the tests became unnecessary.
Before last winter it was possible that the driver's stop() routine would
get called even if the start() routine returned an error, but now that
can't happen. Hence there's no longer any need to check for partial
initialization.
Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
|
|
This patch starts making some serious changes to the UHCI driver.
There's a set of private states for the root hub, and the internal
routines for suspending and resuming work completely differently, with
transitions based on the new states. Now the driver distinguishes
between a privately auto-stopped state and a publicly suspended state,
and it will properly suspend controllers with broken resume-detect
interrupts instead of resetting them.
Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
|
|
Initial git repository build. I'm not bothering with the full history,
even though we have it. We can create a separate "historical" git
archive of that later if we want to, and in the meantime it's about
3.2GB when imported into git - space that would just make the early
git days unnecessarily complicated, when we don't have a lot of good
infrastructure for it.
Let it rip!
|
