fix 2.6.28 EC interrupt storm regression
Signed-off-by: NAlexey Starikovskiy <astarikovskiy@suse.de>
Signed-off-by: NLen Brown <len.brown@intel.com>

Referencies: http://bugzilla.kernel.org/show_bug.cgi?id=12004Signed-off-by: NAlexey Starikovskiy <astarikovskiy@suse.de>
Signed-off-by: NLen Brown <len.brown@intel.com>

http://bugzilla.kernel.org/show_bug.cgi?id=11892Signed-off-by: NAlexey Starikovskiy <astarikovskiy@suse.de>
Signed-off-by: NLen Brown <len.brown@intel.com>

Disabling gpe might interfere with gpe detection/handling,
thus producing "interrupt not handled" errors.
Ironically, disabling of GPE from interrupt context is already
under spinlock, so only userspace needs to start using it.
Signed-off-by: NAlexey Starikovskiy <astarikovskiy@suse.de>
Signed-off-by: NLen Brown <len.brown@intel.com>

Restart current transaction if we recieved unexpected GPEs instead
of needed ones.

http://bugzilla.kernel.org/show_bug.cgi?id=11896Signed-off-by: NAlexey Starikovskiy <astarikovskiy@suse.de>
Signed-off-by: NLen Brown <len.brown@intel.com>

There is a possibility that EC might break if next command is
issued within 1 us after write or burst-disable command.
Suggestd-by: NZhao Yakui <yakui.zhao@intel.com>
Signed-off-by: NAlexey Starikovskiy <astarikovskiy@suse.de>
Signed-off-by: NLen Brown <len.brown@intel.com>

Make sure we can tell if the GPE storm workaround gets activated,
and avoid flooding the logs afterwards.

http://bugzilla.kernel.org/show_bug.cgi?id=11841
"plenty of line "ACPI: EC: non-query interrupt received,
 switching to interrupt mode" in dmesg"
Signed-off-by: NAlan Jenkins <alan-jenkins@tuffmail.co.uk>
Acked-by: NAlexey Starikovskiy <astarikovskiy@suse.de>
Signed-off-by: NLen Brown <len.brown@intel.com>

With the better solution for EC interrupt storm issue,
there is no need to use msleep over udelay.

References:
	http://bugzilla.kernel.org/show_bug.cgi?id=11810
	http://bugzilla.kernel.org/show_bug.cgi?id=10724Signed-off-by: NAlexey Starikovskiy <astarikovskiy@suse.de>
Signed-off-by: NLen Brown <len.brown@intel.com>

breakage introduced by following patch
commit 27663c58
Author: Matthew Wilcox <willy@linux.intel.com>
Date:   Fri Oct 10 02:22:59 2008 -0400

acpi_evaluate_integer() does not clear passed variable if
there is an error at evaluation.
So if we ignore error, we must supply initialized variable.

http://bugzilla.kernel.org/show_bug.cgi?id=11917Signed-off-by: NAlexey Starikovskiy <astarikovskiy@suse.de>
Tested-by: NAlan Jenkins <alan-jenkins@tuffmail.co.uk>
Signed-off-by: NLen Brown <len.brown@intel.com>

On some broken BIOS the ACPI object in EC _REG method can't be found in
ACPI namespace, which causes that the status code of AE_NOT_FOUND is returned by
the EC _REG object. In such case the EC device can't be initialized correctly,
which causes that battery/AC adapter can't work normally. As the EC address
space handler is not removed and the memory pointed by its input argument is
already free, sometimes the kernel will also be panic when EC internal register
is still accessed. But the windows can work well on such broken BIOS.

Maybe it will be reasonable that OS ignores the AE_NOT_FOUND error
returned by the EC _REG object and continues to initialize EC device
on some broken BIOS.
For example: the ACPI object in EC _REG method can't be found and status error
code is AE_NOT_FOUND.

http://bugzilla.kernel.org/show_bug.cgi?id=8953
http://bugzilla.kernel.org/show_bug.cgi?id=10237

lenb: we may find a more general solution to this in the future.
Signed-off-by: NZhao Yakui <yakui.zhao@intel.com>
Signed-off-by: NLin Ming <ming.m.lin@intel.com>
Signed-off-by: NLen Brown <len.brown@intel.com>

Signed-off-by: NAlexey Starikovskiy <astarikovskiy@suse.de>
Tested-by: NAlan Jenkins <alan-jenkins@tuffmail.co.uk>
Signed-off-by: NLen Brown <len.brown@intel.com>

As of version 2.0, ACPI can return 64-bit integers.  The current
acpi_evaluate_integer only supports 64-bit integers on 64-bit platforms.
Change the argument to take a pointer to an acpi_integer so we support
64-bit integers on all platforms.

lenb: replaced use of "acpi_integer" with "unsigned long long"
lenb: fixed bug in acpi_thermal_trips_update()
Signed-off-by: NMatthew Wilcox <willy@linux.intel.com>
Signed-off-by: NLen Brown <len.brown@intel.com>

Catch attempts to use of acpi_driver_data on pointers of wrong type.

akpm: rewritten to use proper C typechecking and remove the
"function"-used-as-lvalue thing.
Signed-off-by: NPavel Machek <pavel@suse.cz>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLen Brown <len.brown@intel.com>

No functional changes.
Signed-off-by: NAlexey Starikovskiy <astarikovskiy@suse.de>
Acked-by: NRafael J. Wysocki <rjw@suse.com>
Signed-off-by: NLen Brown <len.brown@intel.com>

It is easier and faster to do transaction directly from interrupt context
rather than waking control thread.
Also, cleaner GPE storm avoidance is implemented.
References: 	http://bugzilla.kernel.org/show_bug.cgi?id=9998
		http://bugzilla.kernel.org/show_bug.cgi?id=10724
		http://bugzilla.kernel.org/show_bug.cgi?id=10919
		http://bugzilla.kernel.org/show_bug.cgi?id=11309
		http://bugzilla.kernel.org/show_bug.cgi?id=11549Signed-off-by: NAlexey Starikovskiy <astarikovskiy@suse.de>
Tested-by: NSitsofe Wheeler <sitsofe@yahoo.com>
Signed-off-by: NLen Brown <len.brown@intel.com>

When EC is in Polling mode, OS will check the EC status continually by using
the following source code:
       clear_bit(EC_FLAGS_WAIT_GPE, &ec->flags);
       while (time_before(jiffies, delay)) {
               if (acpi_ec_check_status(ec, event))
       	            return 0;
               msleep(1);
       }
But msleep is realized by the function of schedule_timeout. At the same time
although one process is already waken up by some events, it won't be scheduled
immediately. So maybe there exists the following phenomena:
     a. The current jiffies is already after the predefined jiffies.
	But before timeout happens, OS has no chance to check the EC
	status again.
     b. If preemptible schedule is enabled, maybe preempt schedule will happen
	before checking loop. When the process is resumed again, maybe
	timeout already happens, which means that OS has no chance to check
	the EC status.

In such case maybe EC status is already what OS expects when timeout happens.
But OS has no chance to check the EC status and regards it as AE_TIME.

So it will be more appropriate that OS will try to check the EC status again
when timeout happens. If the EC status is what we expect, it won't be regarded
as timeout. Only when the EC status is not what we expect, it will be regarded
as timeout, which means that EC controller can't give a response in time.

http://bugzilla.kernel.org/show_bug.cgi?id=9823
http://bugzilla.kernel.org/show_bug.cgi?id=11141Signed-off-by: NZhao Yakui <yakui.zhao@intel.com>
Signed-off-by: NZhang Rui  <rui.zhang@intel.com>
Signed-off-by: NAndi Kleen <ak@linux.intel.com>

On some ASUS laptops the ECDT gives the incorrect command/status & Data I/O
register address.

AK: it seems like the command/data addresses are exchanged.

In such case it will cause that EC device can't be
initialized correctly.
To add the EC dmi table is to fix this issue. If the laptop falls into the
EC dmi table, the EC command/data I/O address will be fixed.

AK: Add comments describing this better

http://bugzilla.kernel.org/show_bug.cgi?id=9399Signed-off-by: NZhao Yakui <yakui.zhao@intel.com>
tested-by    : Jan Kasprzak  <kas@fi.muni.cz>
Signed-off-by: NAndi Kleen <ak@linux.intel.com>

http://bugzilla.kernel.org/show_bug.cgi?id=10724Signed-off-by: NAlexey Starikovskiy <astarikovskiy@suse.de>
Signed-off-by: NLen Brown <len.brown@intel.com>

Use proc_create()/proc_create_data() to make sure that ->proc_fops and ->data
be setup before gluing PDE to main tree.

Add correct ->owner to proc_fops to fix reading/module unloading race.
Signed-off-by: NDenis V. Lunev <den@openvz.org>
Cc: Len Brown <lenb@kernel.org>
Cc: Alexey Dobriyan <adobriyan@gmail.com>
Cc: "Eric W. Biederman" <ebiederm@xmission.com>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

Signed-off-by: NAlexey Starikovskiy <astarikovskiy@suse.de>
Signed-off-by: NLen Brown <len.brown@intel.com>

Signed-off-by: NAlexey Starikovskiy <astarikovskiy@suse.de>
Signed-off-by: NLen Brown <len.brown@intel.com>

Problem seems to be that hw fails to clear GPE after we service it and write 1
into corresponding bit. Thus, as soon as we get interrupts enabled again, we
receive a new one. Google gives too many results for "acer interrupt storm" for
this being one-broken-machine case.

Reference: http://bugzilla.kernel.org/show_bug.cgi?id=9998Signed-off-by: NAlexey Starikovskiy <astarikovskiy@suse.de>
Signed-off-by: NLen Brown <len.brown@intel.com>

Signed-off-by: NAlexey Starikovskiy <astarikovskiy@suse.de>
Signed-off-by: NLen Brown <len.brown@intel.com>

Signed-off-by: NAlexey Starikovskiy <astarikovskiy@suse.de>
Signed-off-by: NLen Brown <len.brown@intel.com>

Signed-off-by: NAlexey Starikovskiy <astarikovskiy@suse.de>
Signed-off-by: NLen Brown <len.brown@intel.com>

If we can not use interrupt mode of EC for some reason, start polling
EC for events periodically.
Signed-off-by: NAlexey Starikovskiy <astarikovskiy@suse.de>
Signed-off-by: NLen Brown <len.brown@intel.com>

This fixes keyboard event handling on some systems.

Note that this delay was thought unnecessary, and removed
from linux-2.6.20 with 50c1e113
'ACPI: ec: Drop udelay() from poll mode. Loop by reading status field instead.'
Signed-off-by: NAlexey Starikovskiy <astarikovskiy@suse.de>
Signed-off-by: NLen Brown <len.brown@intel.com>

This reverts commit 2c81ce4c.

It caused several new troubles (eg suspend slowdown bisected down to
this patch by Pavel Machek), so just revert it for now.
Signed-off-by: NAlexey Starikovskiy <astarikovskiy@suse.de>
Cc: Pavel Machek <pavel@suse.cz>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

On some Acer systems, the HW fails to clear the GPE source,
causing an interrupt storm.

So in EC interrupt mode, we count how many interrupts we
receive when waiting.  If we get more than 5, we give
up on interrupt mode and revert to polling mode.

Also, for polling mode to work on Acers, we need
to insert a delay.
Signed-off-by: NAlexey Starikovskiy <astarikovskiy@suse.de>
Signed-off-by: NLen Brown <len.brown@intel.com>

This reverts commit 208c70a4.

http://bugzilla.kernel.org/show_bug.cgi?id=10100Signed-off-by: NLen Brown <len.brown@intel.com>

Fall back to ACPI_ROOT_HANDLE only in case of error.

ACPI: EC: EC description table is found, configuring boot EC
ACPI Error (evregion-0316): No handler for Region [ECOR] (ffff81007a651620) [EmbeddedControl] [20070126]
ACPI Error (exfldio-0289): Region EmbeddedControl(3) has no handler [20070126]

http://bugzilla.kernel.org/show_bug.cgi?id=9916Signed-off-by: NAlexey Starikovskiy <astarikovskiy@suse.de>
Signed-off-by: NLen Brown <len.brown@intel.com>

The "DEBUG" symbol needs to be defined before #including <linux/kernel.h> to
get the pr_debug() working.
Signed-off-by: NMárton Németh <nm127@freemail.hu>
Signed-off-by: NLen Brown <len.brown@intel.com>

Add leading zeros to pr_debug() calls. For example if x=0x0a, the format
"0x%2x" will result the string "0x a", the format "0x%2.2x" will result "0x0a".
Signed-off-by: NMárton Németh <nm127@freemail.hu>
Signed-off-by: NLen Brown <len.brown@intel.com>

Return OBF_1 optimization workaround

http://bugzilla.kernel.org/show_bug.cgi?id=8459Signed-off-by: NAlexey Starikovskiy <astarikovskiy@suse.de>
Signed-off-by: NLen Brown <len.brown@intel.com>

Burst mode temporary (50 ms) locks EC to do only transactions with
driver, without it some hardware returns abstract garbage.

Reference: http://bugzilla.kernel.org/show_bug.cgi?id=9341Signed-off-by: NAlexey Starikovskiy <astarikovskiy@suse.de>
Signed-off-by: NLen Brown <len.brown@intel.com>

Specification allows only byte access for EC region, so
make it separate from bug-compatible multi-byte access.
Also do not allow return of garbage in supplied *value.

Reference: http://bugzilla.kernel.org/show_bug.cgi?id=9341Signed-off-by: NAlexey Starikovskiy <astarikovskiy@suse.de>
Signed-off-by: NLen Brown <len.brown@intel.com>

Some _STA methods called during bus_scan() might require EC region handler,
which might be enabled later in the scan.
Enable it explicitly before scan to avoid errors.

Reference: http://bugzilla.kernel.org/show_bug.cgi?id=9627Signed-off-by: NAlexey Starikovskiy <astarikovskiy@suse.de>
Signed-off-by: NLen Brown <len.brown@intel.com>

Some controllers fail to send confirmation GPE after address or data write.
Detect this and don't expect such confirmation in future.
This is a generalization of previous workaround
(66c5f4e7), which did only read address.

Reference: http://bugzilla.kernel.org/show_bug.cgi?id=9327Signed-off-by: NAlexey Starikovskiy <astarikovskiy@suse.de>
Tested-by: NRomano Giannetti <romano.giannetti@gmail.com>
Tested-by: Mats Johannesson
Signed-off-by: NLen Brown <len.brown@intel.com>

Sometimes it is usefull to see raw protocol dump.
Uncomment '#define DEBUG' at the beginning of file to make EC
really verbose.
Signed-off-by: NMárton Németh <nm127@freemail.hu>
Signed-off-by: NAlexey Starikovskiy <astarikovskiy@suse.de>
Signed-off-by: NLen Brown <len.brown@intel.com>

This reverts commit f2d68935.