13 Feb, 2007
40 commits
-
This is an attempt of providing an interface for memory scrubbing control in
EDAC.This patch modifies the EDAC Core to provide the Interface for memory
controller modules to implment.The following things are still outstanding:
- K8 is the first implemenation,
The patch provide a method of configuring the K8 hardware memory scrubber
via the 'mcX' sysfs directory. There should be some fallback to a generic
scrubber implemented in software if the hardware does not support
scrubbing.Or .. the scrubbing sysfs entry should not be visible at all.
- Only works with SDRAM, not cache,
The K8 can scrub cache and l2cache also - but I think this is not so
useful as the cache is busy all the time (one hopes).One would also expect that cache scrubbing requires hardware support.
- Error Handling,
I would like that errors are returned to the user in "terms of file
system".- Presentation,
I chose Bandwidth in Bytes/Second as a representation of the scrubbing
rate for the following reasons:I like that the sysfs entries are sort-of textual, related to something
that makes sense instead of magical values that must be looked up."My People" wants "% main memory scrubbed per hour" others prefer "%
memory bandwidth used" as representation, "bandwith used" makes it easy to
calculate both versions in one-liner scripts.If one later wants to scrub cache, the scaling becomes wierd for K8
changing from "blocks of 64 byte memory" to "blocks of 64 cache lines" to
"blocks of 64 bit". Using "bandwidth used" makes sense in all three cases,
(I.M.O. anyway ;-).- Discovery,
There is no way to discover the possible settings and what they do
without reading the code and the documentation.*I* do not know how to make that work in a practical way.
- Bugs(??),
other tools can set invalid values in the memory scrub control register,
those will read back as '-1', requiring the user to reset the scrub rate.
This is how *I* think it should be.- Afflicting other areas of code,
I made changes to edac_mc.c and edac_mc.h which will show up globally -
this is not nice, it would be better that the memory scrubbing fuctionality
and interface could be entirely contained within the memory controller it
applies to.Frithiof Jensen
edac_mc.c and its .h file is a CORE helper module for EDAC
driver modules. This provides the abstraction for device specific
drivers. It is fine to modify this CORE to provide help for
new features of the the driversdoug thompson
Signed-off-by: Frithiof Jensen
Signed-off-by: doug thompson
Acked-by: Alan Cox
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
This fix/change returns the offset into the page for the ce/ue error, instead
of just 0. The e752x dram controller reads 34:6 of the linear address with
the error.Signed-off-by: Mike Chan
Signed-off-by: doug thompson
Acked-by: Alan Cox
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
The reading of the DRA registers should be a byte at a time (one register at a
time) instead of 4 bytes at a time (four registers). Reading a dword at a
time retrieves erroneous information from all but the first register. A
change was made to read in each register in a loop prior to using the data in
those registers.Signed-off-by: Brian Pomerantz
Signed-off-by: Dave Jiang
Signed-off-by: Doug Thompson
Cc: Alan Cox
Cc: Andi Kleen
Cc: "Eric W. Biederman"
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
The fatal vs. non-fatal mask for the sysbus FERR status is incorrect
according to the E7520 datasheet. This patch corrects the mask to correctly
handle fatal and non-fatal errors.Signed-off-by: Brian Pomerantz
Signed-off-by: Dave Jiang
Signed-off-by: Doug Thompson
Cc: Alan Cox
Cc: Andi Kleen
Cc: "Eric W. Biederman"
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
Now that I have changed all of the in-tree users remove the old version of
these functions. This should make it clear to any out of tree users that they
should be using kill_pgrp kill_pgrp_info or __kill_pgrp_info instead.Signed-off-by: Eric W. Biederman
Cc: Alan Cox
Cc: Oleg Nesterov
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
Now that I have changed all of the users remove the old version of these
functions. This should be a clear hint to any out of tree users that they
should use do_each_pid_task and while_each_pid_task for new code.Signed-off-by: Eric W. Biederman
Cc: Alan Cox
Cc: Oleg Nesterov
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
There isn't any real advantage to this change except that it allows the old
functions to be removed. Which is easier on maintenance and puts the code in
a more uniform style.Signed-off-by: Eric W. Biederman
Cc: Alan Cox
Cc: Oleg Nesterov
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
Of kernel subsystems that work with pids the tty layer is probably the largest
consumer. But it has the nice virtue that the assiation with a session only
lasts until the session leader exits. Which means that no reference counting
is required. So using struct pid winds up being a simple optimization to
avoid hash table lookups.In the long term the use of pid_nr also ensures that when we have multiple pid
spaces mixed everything will work correctly.Signed-off-by: Eric W. Biederman
Cc: Alan Cox
Cc: Oleg Nesterov
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
Every call to is_orphaned_pgrp passed in process_group(current) which is racy
with respect to another thread changing our process group. It didn't bite us
because we were dealing with integers and the worse we would get would be a
stale answer.In switching the checks to use struct pid to be a little more efficient and
prepare the way for pid namespaces this race became apparent.So I simplified the calls to the more specialized is_current_pgrp_orphaned so
I didn't have to worry about making logic changes to avoid the race.Signed-off-by: Eric W. Biederman
Cc: Alan Cox
Cc: Oleg Nesterov
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
Modify has_stopped_jobs and will_become_orphan_pgrp to use struct pid based
process groups. This reduces the number of hash tables looks ups and paves
the way for multiple pid spaces.Signed-off-by: Eric W. Biederman
Cc: Alan Cox
Cc: Oleg Nesterov
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
To properly implement a pid namespace I need to deal exclusively in terms of
struct pid, because pid_t values become ambiguous.To this end session_of_pgrp is transformed to take and return a struct pid
pointer. To avoid the need to worry about reference counting I now require my
caller to hold the appropriate locks. Leaving callers repsonsible for
increasing the reference count if they need access to the result outside of
the locks.Since session_of_pgrp currently only has one caller and that caller simply
uses only test the result for equality with another process group, the locking
change means I don't actually have to acquire the tasklist_lock at all.tiocspgrp is also modified to take and release the lock. The logic there is a
little more complicated but nothing I won't need when I convert pgrp of a tty
to a struct pid pointer.Signed-off-by: Eric W. Biederman
Cc: Alan Cox
Cc: Oleg Nesterov
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
The goal is to remove users of the old signal helper functions so they can be
removed.Signed-off-by: Eric W. Biederman
Cc: Alan Cox
Cc: Oleg Nesterov
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
I am slowly moving to a model where all process killing is struct pid based
instead of pid_t based. The sunos compatibility code is one of the last users
of the old pid_t based kill_pg in the kernel. By being complete I allow for
the future removal of kill_pg from the kernel, which will ensure I don't miss
something.Signed-off-by: Eric W. Biederman
Cc: Alan Cox
Cc: Oleg Nesterov
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
commit 24ec839c431eb79bb8f6abc00c4e1eb3b8c4d517 while fixing the locking for
signal->tty got the locking wrong for signal->session. This places our
accesses of signal->session back under the tasklist_lock where they belong.Signed-off-by: Eric W. Biederman
Cc: Alan Cox
Cc: Oleg Nesterov
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
The code to look at tty_old_pgrp and send SIGHUP and SIGCONT when it is
present only executes when disassociate_ctty is called from do_exit. Make
this clear by adding an explict on_exit check, and explicitly setting
tty_old_pgrp to 0.In addition fix the locking by reading tty_old_pgrp under the siglock.
Signed-off-by: Eric W. Biederman
Cc: Alan Cox
Cc: Oleg Nesterov
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
The aim of this patch set is to start wrapping up the struct pid conversions.
As such this patchset culminates with the removal of kill_pg, kill_pg_info,
__kill_pg_info, do_each_task_pid, and while_each_task_pid.kill_proc, daemonize, and kernel_thread are still in my sights but there is
still work to get to them.The first three are basic cleanups around disassociate_ctty, while working on
converting it I found several issues. tty_old_pgrp can be a tricky concept to
wrap your head around.1 tty: Make __proc_set_tty static.
2 tty: Clarify disassociate_ctty
3 tty: Fix the locking for signal->session in disassociate_cttyThese just stop using the old helper functions.
4 signal: Use kill_pgrp not kill_pg in the sunos compatibility code.
5 signal: Rewrite kill_something_info so it uses newer helpers.Then the grind to convert the tty layer and all of it's helper functions to
struct pid.6 pid: Make session_of_pgrp use struct pid instead of pid_t.
7 pid: Use struct pid for talking about process groups in exit.c
8 pid: Replace is_orphaned_pgrp with is_current_pgrp_orphaned
9 tty: Update the tty layer to work with struct pid.A final helper function update.
10 pid: Replace do/while_each_task_pid with do/while_each_pid_task
And the removal of the functions that are now unused.
11 pid: Remove now unused do_each_task_pid and while_each_task_pid
12 pid: Remove the now unused kill_pg kill_pg_info and __kill_pg_infoAll of these should be fairly simple and to the point.
This patch:
Currently all users of __proc_set_tty are in tty_io.c so make the function
static.Signed-off-by: Eric W. Biederman
Cc: Alan Cox
Cc: Oleg Nesterov
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
This morning I needed to read a Minix V3 filesystem, but unfortunately my
2.6.19 did not support that, and neither did the downloaded 2.6.20rc4.Fortunately, google told me that Daniel Aragones had already done the work,
patch found at http://www.terra.es/personal2/danarag/Unfortunaly, looking at the patch was painful to my eyes, so I polished it
a bit before applying. The resulting kernel boots, and reads the
filesystem it needed to read.Signed-off-by: Daniel Aragones
Signed-off-by: Andries Brouwer
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
This is adds a simple SPI EEPROM driver, providing access to the EEPROM
through sysfs much like the I2C "eeprom" driver ... except this driver
supports write access, and multiple EEPROM sizes.From: "Tuppa, Walter"
Since I have EEPROMs on SPI with different address sizing, I made some
changes to your at25.c to support them. Works perfectly. (Also includes a
small bugfix for the "what size address" test.)Signed-off-by: David Brownell
Signed-off-by: Walter Tuppa
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
This adds a new "wakealarm" sysfs attribute to RTC class devices which support
alarm operations and are wakeup-capable:- It reads as either empty, or the scheduled alarm time as seconds
since the POSIX epoch. (That time may already have passed, since
nothing currently enforces one-shot alarm semantics.)- It can be written with an alarm time in the future, again seconds
since the POSIX epoch, which enables the alarm.- It can be written with an alarm time not in the future (such as 0,
the start of the POSIX epoch) to disable the alarm.Usage examples (some need GNU date) after "cd /sys/class/rtc/rtcN":
alarm after 10 minutes:
# echo $(( $(cat since_epoch) + 10 * 60 )) > wakealarm
alarm tuesday evening 10pm:
# date -d '10pm tuesday' "+%s" > wakealarm
disable alarm:
# echo 0 > wakealarmThis resembles the /proc/acpi/alarm file in that nothing happens when the
alarm triggers ... except possibly waking the system from sleep. It's also
like that in a nasty way: not much can be done to prevent one task from
clobbering another task's alarm settings.It differs from that file in that there's no in-kernel date parser.
Note that a few RTCs ignore rtc_wkalrm.enabled when setting alarms, or aren't
set up correctly, so they won't yet behave with this attribute.Signed-off-by: David Brownell
Acked-by: Pavel Machek
Cc: Alessandro Zummo
Cc: Greg KH
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
This clarifies some aspects of the SPI programming interface, based on
feedback from Hans-Peter Nilsson. The in-memory representation of words is
right-aligned, so for example a twelve bit word is stored using sixteen bits
with four undefined bits in the MSB. And controller drivers must reject
protocol tweaking modes they do not support.Signed-off-by: David Brownell
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
I'd like to assign NULL to kfree()d members of a structure. I can't do
that without ugly casting (see the PXA patch) when the structure pointed to
is const-qualified. I don't really see a reason why the cleanup method
isn't allowed to alter the object it should clean up. :-)No, I didn't test the PXA patch, but I verified that the NULL-assignment
doesn't stop me from doing rmmod/insmodding my own spi_bitbang-based
driver.Signed-off-by: Hans-Peter Nilsson
Signed-off-by: David Brownell
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
A small bug-fix for spi_bitbang: it must always call the setup_transfer
function via the overridable pointer, not assume that its
spi_bitbang_setup_transfer is sufficient. Otherwise, if all options in the
transfers are default (0), the overrided function will never be called.Granted, the function replacing it must call spi_bitbang_setup_transfer,
but it might also have other important things to do, even if the second
argument (the spi_transfer) is NULL. Tested together with the other
patches on the spi_crisv32_sser and spi_crisv32_gpio drivers (not yet in
the kernel, will IIUC be submitted as part of the usual
arch-maintainer-pushes).Signed-off-by: Hans-Peter Nilsson
Signed-off-by: David Brownell
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
Make the spi_unregister_driver() code fit in with the rest of the header
file, and only do the action if the driver passed is non-NULL.This also makes the code a line smaller.
Signed-off-by: Ben Dooks
Signed-off-by: David Brownell
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
The spi_register_driver() sets the bus_type field of the spi_driver being
registered, so there is no need to have it set in the driver itself.Signed-off-by: Ben Dooks
Signed-off-by: David Brownell
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
Add wrappers for getting and setting the driver data using spi_device
instead of using dev_{get|set}_drvdata with &spi->dev, to mirror the
platform_{get|set}_drvdata.Signed-off-by: Ben Dooks
Signed-off-by: David Brownell
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
Add the SPI controller driver for Freescale i.MX(S/L/1).
Main features summary:> Per chip setup via board specific code and/or protocol driver.
> Per transfer setup.
> PIO transfers.
> DMA transfers.
> Managing of NULL tx / rx buffer for rd only / wr only transfers.This patch replace patch-2.6.20-rc4-spi_imx with the following changes:
> Few cosmetic changes.
> Function map_dma_buffers now return 0 for success and -1 for failure.
> Solved a bug inside spi_imx_probe function (wrong error path).
> Solved a bug inside setup function (bad undo setup for max_speed_hz).
> For read-only transfers, always write zero bytes.This is almost the same as the 'BIS' version sent by Andrea, except for
updating the 'DUMMY' byte so that read-only transfers shift out zeroes.
That part of the API changed recently, since some half duplex peripheral
chips require that semantic.Signed-off-by: Andrea Paterniani
Signed-off-by: David Brownell
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
This adds a SPI driver for the Microwire controller on OMAP1 chips. This
driver has been used in the Linux-OMAP tree for some time now, including
with some of those displays using standardized 9-bit commands followed by
data with 8-bit words.Microwire only supports half duplex transfers, but that's all that most SPI
protocols need. When full duplex, or higher speeds, are needed there are
several other controllers that can be used on OMAP.[akpm@osdl.org: cleanups]
Signed-off-by: David Brownell
Signed-off-by: Imre Deak
Signed-off-by: Juha Yrjola
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
Minor Kconfig cleanup ... put the SPI_S3C24XX entry in the
correct location (alphabetical order).Signed-off-by: David Brownell
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
drivers/char/ds1302.c::get_rtc_time() contains local_irq_disable() call
after local_irq_save(). This looks redundant.drivers/char/ds1302.c::rtc_ioctl() contains local_irq_disable() call after
local_irq_save(). This looks redundant.Signed-off-by: Jiri Kosina
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
This patch adds the line discipline based driver for the Gigaset M101
wireless RS232 adapter. It also improves the documentation a bit.Signed-off-by: Tilman Schmidt
Signed-off-by: Hansjoerg Lipp
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
Convert:
BOOL -> bool
FALSE -> false
TRUE -> trueChange a variable ('mContinue') to boolean from char, since it is used
as boolean.Signed-off-by: Richard Knutsson
Acked-by: Randy Dunlap
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
- Use timer macros to set function and data members and to modify
expiration time.
- Use DEFINE_TIMER for global timers and do not init them at run-time in
these cases.
- del_timer_sync is common in most cases -- we want to wait for timer
function if it's still running.Signed-off-by: Jiri Slaby
Cc: Dave Airlie
Cc: David Woodhouse
Cc: Dominik Brodowski
Cc: Alessandro Zummo
Cc: Paul Fulghum
Cc: Kylene Jo Hall
Cc: Wim Van Sebroeck
Acked-by: Dmitry Torokhov (Input bits)
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
specialix, isr have 2 params
pt_regs are no longer the third parameter of isr, call sx_interrupt without
it.Signed-off-by: Jiri Slaby
Cc:
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
Change __init to __devinit in rtc drivers' probe functions.
Resolves MODPOST warnings:
WARNING: drivers/rtc/rtc-ds1553.o - Section mismatch: reference to
.init.text:ds1553_rtc_probe from .data.rel between 'ds1553_rtc_driver' (at
offset 0x0) and 'ds1553_nvram_attr'
WARNING: drivers/rtc/rtc-ds1742.o - Section mismatch: reference to
.init.text:ds1742_rtc_probe from .data.rel between 'ds1742_rtc_driver' (at
offset 0x0) and 'ds1742_nvram_attr'Signed-off-by: Prarit Bhargava
Cc: Alessandro Zummo
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
Was ufs_fs.h purposefully not exported to userspace or did it just slip
through the cracks ? assuming the latter scenario, the attached patch touches
up the relationship between ufs_fs.h and its sub headers (like ufs_fs_sb.h) so
that we can export it ... the silo bootloader takes advantage of this header
for example.Signed-off-by: Mike Frysinger
Cc: Evgeniy Dushistov
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
close_files() can sometimes take long enough to trigger the soft lockup
detector.Cc: Eric Dumazet
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
Reduce the number of kernel messages the Gigaset drivers produce in case of an
excessively long device response, from one per character exceeding the limit
to one per overlong message.Signed-off-by: Tilman Schmidt
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
Do not lock the cardstate structure mutex earlier than necessary.
Signed-off-by: Tilman Schmidt
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
oprofile hunting showed a stall in rw_verify_area(), because of triple
indirection and potential cache misses.
(file->f_path.dentry->d_inode->i_flock)By moving initialization of 'struct inode' pointer before the pos/count
sanity tests, we allow the compiler and processor to perform two loads by
anticipation, reducing stall, without prefetch() hints. Even x86 arch has
enough registers to not use temporary variables and not increase text size.I validated this patch running a bench and studied oprofile changes, and
absolute perf of the test program.Results of my epoll_pipe_bench (source available on request) on a Pentium-M
1.6 GHz machineBefore :
# ./epoll_pipe_bench -l 30 -t 20
Avg: 436089 evts/sec read_count=8843037 write_count=8843040 21.218390 samples
per call
(best value out of 10 runs)After :
# ./epoll_pipe_bench -l 30 -t 20
Avg: 470980 evts/sec read_count=9549871 write_count=9549894 21.216694 samples
per call
(best value out of 10 runs)oprofile CPU_CLK_UNHALTED events gave a reduction from 5.3401 % to 2.5851 %
for the rw_verify_area() function.Signed-off-by: Eric Dumazet
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
Need to export com20020 symbols for com20020_cs also.
WARNING: "com20020_found" [drivers/net/pcmcia/com20020_cs.ko] undefined!
WARNING: "com20020_check" [drivers/net/pcmcia/com20020_cs.ko] undefined!Signed-off-by: Randy Dunlap
Cc: Esben Nielsen
Cc: Jeff Garzik
Cc: Dominik Brodowski
Cc: "David S. Miller"
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds
