sched: clean up fastcall uses of sched_fork()/sched_exit()
sched_fork()/sched_exit() does not need to specify fastcall anymore,
as the x86 kernel defaults to regparm3, and no assembly code calls
these functions.
track TSC-unstable events and propagate it to the scheduler code.
Also allow sched_clock() to be used when the TSC is unstable,
the rq_clock() wrapper creates a reliable clock out of it.
this change switches over the scheduler core to CFS's modular
design and makes use of kernel/sched_fair/rt/idletask.c to implement
Linux's scheduling policies.
thanks to Andrew Morton and Thomas Gleixner for lots of detailed review
feedback and for fixlets.
remove the sleep_type heuristics from the core scheduler - scheduling
policy is implemented in the scheduling-policy modules. (and CFS does
not use this type of sleep-type heuristics)
add rq_clock()/__rq_clock(), a robust wrapper around sched_clock(),
used by CFS. It protects against common type of sched_clock() problems
(caused by hardware): time warps forwards and backwards.
add kernel/sched_idletask.c - which implements the idle thread
scheduling class. This further simplifies sched.c (under CFS),
for example a number of 'if (p == rq->idle)' type of special-cases
can be removed from sched.c, and schedule() gets simpler too.
add the init_idle_bootup_task() callback to the bootup thread,
unused at the moment. (CFS will use it to switch the scheduling
class of the boot thread to the idle class)
sched: zap the migration init / cache-hot balancing code
the SMP load-balancer uses the boot-time migration-cost estimation
code to attempt to improve the quality of balancing. The reason for
this code is that the discrete priority queues do not preserve
the order of scheduling accurately, so the load-balancer skips
tasks that were running on a CPU 'recently'.
this code is fundamental fragile: the boot-time migration cost detector
doesnt really work on systems that had large L3 caches, it caused boot
delays on large systems and the whole cache-hot concept made the
balancing code pretty undeterministic as well.
(and hey, i wrote most of it, so i can say it out loud that it sucks ;-)
under CFS the same purpose of cache affinity can be achieved without
any special cache-hot special-case: tasks are sorted in the 'timeline'
tree and the SMP balancer picks tasks from the left side of the
tree, thus the most cache-cold task is balanced automatically.
enum idle_type (used by the load-balancer) clashes with the
SCHED_IDLE name that we want to introduce. 'CPU_IDLE' instead
of 'SCHED_IDLE' is more descriptive as well.
HID: handle cases of volume knobs generating relative values
There are some devices (for example Dell Multimedia Keyboard SK-8135) that have
a volume control knob which generates relative events instead of absolute.
hid-input maps them to ABS_VOLUME. HUT pages don't restrict volume to absolute
values.
Adding REL_VOLUME doesn't seem feasible, nothing knows how to handle it. This
patch translates relative ABS_VOLUME events into appropriate number of series
of VOLUME_UP or VOLUME_DOWN events respectively, so that userspace sees the
correct values in the end.
Ryo Dairiki [Mon, 25 Jun 2007 08:31:12 +0000 (10:31 +0200)]
HID: support for logitech cordless desktop LX500 special mapping
This keyboard has wireless mouse which has left, middle, right buttons and
2-dimensional scrolling wheel. Unfornetuly, this wheel reports side scrolling
events and 11 or 12 button events at the same time.
I've wrote a patch to fix this mapping. I'm not sure if this mapping is proper
for buttons, because , for example, there is no entry for "burn cd" in input.h.
The patch also supress 11 and 12 button events from mouse when you scroll the
wheel left and right. With this patch, only side scrolling events are
reported. (This mouse has only 4 buttons and 2D wheel. There is no such
buttons like 11 and 12.)
Diogo Kastrup [Tue, 19 Jun 2007 12:16:15 +0000 (14:16 +0200)]
HID: fix autocentering of PID devices
When setting the autocentering of PID devices, PID_DIRECTION_ENABLE is not
being explicitely set to 1. This results in autocentering working only on the
vertical axis when this field is preset to 0.
Fix that by setting it explicitely to 1 when preparing the set_effect report
for autocentering spring effect.
Jiri Kosina [Tue, 19 Jun 2007 12:09:14 +0000 (14:09 +0200)]
HID: separate quirks for report descriptor fixup
Lately there have been quite a lot of bug reports against broken devices
which require us to fix their report descriptor in the runtime, before it
is passed to the HID parser. Those devices have eaten quite an amount of
our quirks space, which isn't particularly necessary - the quirks are not
needed after the report descriptor is parsed, and they just consume bits.
Therefore this patch separates the quirks for report descriptor fixup, and
moves their handling into separate code. The quirks are then forgotten as
soon as the report descriptor has been parsed.
Module parameter 'rdesc_quirks' is introduced to be able to modify these
quirks in runtime in a similar way to 'quirks' parameter for ordinary HID
quirks.
Jiri Kosina [Mon, 18 Jun 2007 15:41:14 +0000 (17:41 +0200)]
HID: Add NOGET quirk for all NCR devices
Devices manufactured by NCR have userspace hiddev-based drivers,
which do all the necessary device querying by themselves. The devices
must not be queried directly by the generic HID driver, as reported by
NCR engineers.
Jiri Kosina [Mon, 18 Jun 2007 11:30:11 +0000 (13:30 +0200)]
HID: support for Petalynx Maxter remote control
Petalynx Maxter remote control [1] 0x18b1/0x0037 emits 0xfa and 0xfc from
consumer page (reserved in HUT 1.12) for back and more keys. It also emits
a few usages from LOGIVENDOR page, which need adding.
Also, this device has broken report descriptor - the reported maximum is too
low - it doesn't contain the range for 'back' and 'more' keys, so we need to
bump it up before the report descriptor is being parsed.
Hans de Goede [Tue, 12 Jun 2007 20:04:32 +0000 (22:04 +0200)]
HID: fix mismatch between hid-input HUT find/search mapping and the HUT
When comparing usb vs ps2 / testing the keycodes generated for the easy access
keys on my trust (microsoft compatible) keyboard. I noticed the search key
generated the keycode for find when connected through USB. This lead me to
check the consumer page mappings in hid-input.c . And it turns out the the
mapping for ID 0x221 deviates from the HUT standard document:
http://www.usb.org/developers/devclass_docs/Hut1_12.pdf
Currently it is incorrectly mapped to find, whereas it should be mapped to
search. I also added missing bindings for ID 0x21f, the real find and for
0x222, goto.
Jan Engelhardt [Mon, 30 Apr 2007 11:27:48 +0000 (13:27 +0200)]
HID: Use menuconfig objects
Make a "menuconfig" out of the Kconfig objects "menu, ..., endmenu",
so that the user can disable all the options in that menu at once
instead of having to disable each option separately.
Jiri Kosina [Sat, 19 May 2007 14:28:04 +0000 (16:28 +0200)]
HID: force hid-input for Microsoft SideWinder GameVoice device
Microsoft SideWinder GameVoice driver is a trivial device with a few buttons
(0x09 HID usage) and an audio connector, which just forwards the audio input
into oridinary sound card present in the computer.
Despite this fact, the only interface of this device reports itself as a
Telephony/Headset type of HID device. This is apparently incorrect - the device
itself doesn't provide any audio/telephony functionality. This is achieved in
userland application which only needs to receive the button events from the HID
driver.
This patch establishes a new quirk which forces hid-input to claim a device it
will otherwise leave untouched.
Jiri Kosina [Mon, 14 May 2007 08:20:29 +0000 (10:20 +0200)]
HID: input mapping for Chicony KU-0418 tactical pad
Chicony KU-0418 (aka Saitek PZ08AU gaming keyboard) has a separate
"tactical pad" with 11 non-functional buttons - they generate usage
codes from 0xff00 (MSVENDOR) usage page. Special case handling for
this keyboard added, so no later clash with MSVENDOR mappings is going
to occur.
Jiri Kosina [Wed, 30 May 2007 13:07:13 +0000 (15:07 +0200)]
HID: make debugging output runtime-configurable
There have been many reports recently about broken HID devices, the
diagnosis of which required users to recompile their kernels in order
to be able to provide debugging output needed for coding a quirk for
a particular device.
This patch makes CONFIG_HID_DEBUG default y if !EMBEDDED and makes it
possible to control debugging output produced by HID code by supplying
'debug=1' module parameter.
Changes to last version:
- use netdev_alloc_skb
- make init_rx_bufs just fail and not panic, if skb alloc fails
- don/t free_irq, if request_irq failed
Thomas.
Extracted chip specific code out of lasi_82596.c and placed into a lib82596.c
to make it usable for other 82596 drivers
Francois Romieu [Wed, 31 Jan 2007 22:47:43 +0000 (23:47 +0100)]
r8169: mac address change support
Merged from Realtek's r8169-6.001 driver.
I have added some locking to protect against the arp monitoring
timer in the bonding driver. Accessing the configuration registers
is otherwise performed under RTNL locking.
Francois Romieu [Wed, 21 Feb 2007 21:40:46 +0000 (22:40 +0100)]
r8169: cleanup
No functionnal change:
- trim the old history log
- whitespace/indent/case police
- unsigned int where signedness does not matter
- removal of obsolete assert
- needless cast from void * (dev_instance)
- remove dead code once related to power management
- use netdev_alloc_skb.
Francois Romieu [Tue, 20 Feb 2007 21:20:51 +0000 (22:20 +0100)]
r8169: confusion between hardware and IP header alignment
The rx copybreak part is straightforward.
The align field in struct rtl_cfg_info is related to the alignment
requirements of the DMA operation. Its value is set at 2 to limit the
scale of possible regression but my old v1.21 8169 datasheet claims a
8 bytes requirements (which never appeared in the driver, of course)
and the 8101/8168 go with a plain 8 bytes alignment.
Francois Romieu [Mon, 19 Feb 2007 23:00:26 +0000 (00:00 +0100)]
r8169: merge with version 8.001.00 of Realtek's r8168 driver
This one includes:
- more tweaks to rtl_hw_start_8168
- a work around for a Rx FiFO overflow issue on the 8168Bb
- rtl8169_{intr_mask/napi_event} are replaced with per-device fields,
namely tp->{intr/napi}_event
- rtl_cfg_info is converted to C99 for readability but the values are
not changed for the 8169/8110 and the 8101
Includes ChipCmd fix from Bernhard Walle <[email protected]> (2007/02/24).
Francois Romieu [Tue, 13 Feb 2007 22:38:05 +0000 (23:38 +0100)]
r8169: merge with version 6.001.00 of Realtek's r8169 driver
- new identifier for the 8110SCe
- the PCI latency timer is set unconditionally. This part is identical
in Realtek's r8168 (8.001.00) and r8101 (1.001.00)
- initialization of the cache line size register is for the 8169s only
- more magic in rtl_hw_start_8169
- it is not possible to factor out the setting of the the irq event mask
with the 8168 and the 8101 any more. Pushed it into the hw_start handler.
- rtl_set_rx_tx_config_registers() and write to the ChipCmd register are
issued identically for the whole 8169/8110 family: the 8110SCd/8110SCe
are handled the same way
- work around for AMD platform.
Some registers definitions in Realtek's driver are let aside for later.
Francois Romieu [Fri, 23 Feb 2007 22:36:39 +0000 (23:36 +0100)]
r8169: add hooks for per-device hw_start handler
Rationale: rtl8169_hw_start will not help maintaining an unified
driver for different chipsets but people at Realtek are probably
too polite to say it distinctly.
Let's add the hook and keep hw_start handler unchanged.
As can be seen from the content of rtl8169_pci_tbl, the RTL_CFG_1
entry in rtl_cfg_info was unused. I recycled it for the 0x8168.
- pci_dma_sync_single_for_cpu is not needed for a single large packet
- remove the function pointer to help gcc optimizing the inline
pci_dma functions.
Francois Romieu [Sat, 16 Jun 2007 21:28:45 +0000 (23:28 +0200)]
r8169: de-obfuscate modulo arithmetic
The former style suggests a modulo arithmetic misuse but
the expression should never be < 0. Even if it does, the
driver will simply loop longer than expected (not that
the remaining parts of the system will necessarily
appreciate it...).
Let's warn the user when something goes wrong and try
to go over it.
Enable support for Yukon EX chipset (88e8071).
Most of changes are related to new commands to chip for transmit,
and change in status and checksumming.
The General Purpose I/O register is yet another hardware workaround
catchall. Enable workaround that vendor driver does to stay
but for bug compatiable.
On Yukon EX reading some of the undocumented places in the
memory space will cause a hang. Since they don't provide useful
information, just skip the reserved areas.
Linas Vepstas [Mon, 11 Jun 2007 19:09:16 +0000 (14:09 -0500)]
spidernet: fix misnamed flag
The transmit frame tail bit is stranglely misnamed as
"no checksum". Fix the name to what it should be:
"transmit frame tail". No functional change,
just a name change.
projects / linux.git / log