David Gibson [Thu, 15 May 2008 06:46:39 +0000 (16:46 +1000)]
[POWERPC] Convert remaining dts-v0 files to v1
At the moment we have a mixture of left-over version 0 and new-format
version 1 files in arch/powerpc/boot/dts. This is potentially
confusing to people new to the dts format attempting to figure it out.
So, this patch converts all the as-yet unconverted dts v0 files and
converts them to v1. They're mechanically-converted, and not hand
tweaked so in some cases they're not 100% in keeping with usual v1
style, but the convertor program does have some heuristics so the
discrepancies aren't too bad.
I have checked that this patch produces no changes to the resulting
dtb binaries.
Previously, DCR support was configured at compile time to either use
MMIO or native dcr instructions. Although this works for most
platforms, it fails on FPGA platforms:
1) Systems may include more than one DCR bus.
2) Systems may be native DCR capable and still use memory mapped DCR interface.
This patch provides runtime support based on the device trees for the
case where CONFIG_PPC_DCR_MMIO and CONFIG_PPC_DCR_NATIVE are both
selected. Previously, this was a poorly defined configuration, which
happened to provide NATIVE support. The runtime selection is made
based on the dcr-controller having a 'dcr-access-method' attribute
in the device tree. If only one of the above options is selected,
then the code uses #defines to select only the used code in order to
avoid introducing overhead in existing usage.
Signed-off-by: Stephen Neuendorffer <stephen.neuendorffer@xilinx.com> Signed-off-by: Josh Boyer <jwboyer@linux.vnet.ibm.com>
Ilpo Järvinen [Thu, 29 May 2008 10:25:23 +0000 (03:25 -0700)]
tcp: Reorganize tcp_sock to fill 64-bit holes & improve locality
I tried to group recovery related fields nearby (non-CA_Open related
variables, to be more accurate) so that one to three cachelines would
not be necessary in CA_Open. These are now contiguously deployed:
...and they're then followed by URG slowpath & keepalive related
variables.
Head of the out_of_order_queue always needed for empty checks, if
that's empty (and TCP is in CA_Open), following ~200 bytes (in 64-bit)
shouldn't be necessary for anything. If only OFO queue exists but TCP
is in CA_Open, selective_acks (and possibly duplicate_sack) are
necessary besides the out_of_order_queue but the rest of the block
again shouldn't be (ie., the other direction had losses).
As the cacheline boundaries depend on many factors in the preceeding
stuff, trying to align considering them doesn't make too much sense.
Commented one ordering hazard.
There are number of low utilized u8/16s that could be combined get 2
bytes less in total so that the hole could be made to vanish (includes
at least ecn_flags, urg_data, urg_mode, frto_counter, nonagle).
Signed-off-by: Ilpo Järvinen <ilpo.jarvinen@helsinki.fi> Acked-by: Eric Dumazet <dada1@cosmosbay.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Alexey Dobriyan [Thu, 29 May 2008 10:19:37 +0000 (03:19 -0700)]
netfilter: nf_conntrack_expect: fix error path unwind in nf_conntrack_expect_init()
Signed-off-by: Alexey Dobriyan <adobriyan@parallels.com> Signed-off-by: Patrick McHardy <kaber@trash.net> Signed-off-by: David S. Miller <davem@davemloft.net>
Ingo Molnar [Thu, 29 May 2008 09:28:57 +0000 (11:28 +0200)]
revert ("sched: fair-group: SMP-nice for group scheduling")
Yanmin Zhang reported:
Comparing with 2.6.25, volanoMark has big regression with kernel 2.6.26-rc1.
It's about 50% on my 8-core stoakley, 16-core tigerton, and Itanium Montecito.
Matt Carlson [Mon, 26 May 2008 06:51:01 +0000 (23:51 -0700)]
tg3: Update version to 3.93
This patch increments the version to 3.93.
Signed-off-by: Matt Carlson <mcarlson@broadcom.com> Signed-off-by: Michael Chan <mchan@broadcom.com> Signed-off-by: Benjamin Li <benli@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Matt Carlson [Mon, 26 May 2008 06:49:44 +0000 (23:49 -0700)]
tg3: Add shmem options.
This patch adds some options obtained through shared memory.
Signed-off-by: Matt Carlson <mcarlson@broadcom.com> Signed-off-by: Michael Chan <mchan@broadcom.com> Signed-off-by: Benjamin Li <benli@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Matt Carlson [Mon, 26 May 2008 06:48:31 +0000 (23:48 -0700)]
tg3: Add 5785 ASIC revision
This patch added the 5785 device ID and ASIC revision to the code.
Signed-off-by: Matt Carlson <mcarlson@broadcom.com> Signed-off-by: Michael Chan <mchan@broadcom.com> Signed-off-by: Benjamin Li <benli@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Matt Carlson [Mon, 26 May 2008 06:47:41 +0000 (23:47 -0700)]
tg3: Add libphy support.
This patch introduces the libphy support.
Signed-off-by: Matt Carlson <mcarlson@broadcom.com> Signed-off-by: Michael Chan <mchan@broadcom.com> Signed-off-by: Benjamin Li <benli@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Matt Carlson [Thu, 29 May 2008 08:37:54 +0000 (01:37 -0700)]
tg3: Add mdio bus registration
This patch introduces code to register and unregister the tg3 mdio bus
with the system.
Signed-off-by: Matt Carlson <mcarlson@broadcom.com> Signed-off-by: Michael Chan <mchan@broadcom.com> Signed-off-by: Benjamin Li <benli@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Matt Carlson [Mon, 26 May 2008 06:45:58 +0000 (23:45 -0700)]
tg3: Add TG3_FLG3_USE_PHYLIB
This patch introduces the TG3_FLG3_USE_PHYLIB flag and applies it to
some select places. This work makes later patches a little easier to
read.
Signed-off-by: Matt Carlson <mcarlson@broadcom.com> Signed-off-by: Michael Chan <mchan@broadcom.com> Signed-off-by: Benjamin Li <benli@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Matt Carlson [Mon, 26 May 2008 06:45:08 +0000 (23:45 -0700)]
tg3: Code cleanup.
This patch applies cleanups that would otherwise clutter later
patches.
Signed-off-by: Matt Carlson <mcarlson@broadcom.com> Signed-off-by: Michael Chan <mchan@broadcom.com> Signed-off-by: Benjamin Li <benli@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Matt Carlson [Mon, 26 May 2008 06:44:14 +0000 (23:44 -0700)]
tg3: Pure code movement.
This patch moves some functions towards the top of the file to avoid
unnecessary function prototypes.
Signed-off-by: Matt Carlson <mcarlson@broadcom.com> Signed-off-by: Michael Chan <mchan@broadcom.com> Signed-off-by: Benjamin Li <benli@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Arjan van de Ven [Thu, 29 May 2008 08:32:47 +0000 (01:32 -0700)]
bluetooth: fix locking bug in the rfcomm socket cleanup handling
in net/bluetooth/rfcomm/sock.c, rfcomm_sk_state_change() does the
following operation:
if (parent && sock_flag(sk, SOCK_ZAPPED)) {
/* We have to drop DLC lock here, otherwise
* rfcomm_sock_destruct() will dead lock. */
rfcomm_dlc_unlock(d);
rfcomm_sock_kill(sk);
rfcomm_dlc_lock(d);
}
}
which is fine, since rfcomm_sock_kill() will call sk_free() which will call
rfcomm_sock_destruct() which takes the rfcomm_dlc_lock()... so far so good.
HOWEVER, this assumes that the rfcomm_sk_state_change() function always gets
called with the rfcomm_dlc_lock() taken. This is the case for all but one
case, and in that case where we don't have the lock, we do a double unlock
followed by an attempt to take the lock, which due to underflow isn't
going anywhere fast.
This patch fixes this by moving the stragling case inside the lock, like
the other usages of the same call are doing in this code.
This was found with the help of the www.kerneloops.org project, where this
deadlock was observed 51 times at this point in time:
http://www.kerneloops.org/search.php?search=rfcomm_sock_destruct
Signed-off-by: Arjan van de Ven <arjan@linux.intel.com> Acked-by: Marcel Holtmann <marcel@holtmann.org> Signed-off-by: David S. Miller <davem@davemloft.net>
the problem is that drivers/xen/manage.c is built unconditionally if
CONFIG_XEN is enabled and makes use of xen_suspend(), but
drivers/char/hvc_xen.c, where the xen_suspend() method is implemented,
is only build if CONFIG_HVC_XEN=y as well.
i have solved this by providing a NOP implementation for xen_suspend()
in the !CONFIG_HVC_XEN case.
Philipp Zabel [Wed, 28 May 2008 16:58:06 +0000 (17:58 +0100)]
[ALSA] ASoC: Add UDA1380 driver
The UDA1380 codec is used by the HTC Magician and a number of Samsung
reference boards.
This driver has had a long out of tree history, having originally been
written by Giorgio Padrin and converted to ASoC by Richard Purdie.
Since conversion to ASoC most of the work on the driver has been done by
Philipp Zabel with some review and updates for new APIs by Liam Girdwood
and Mark Brown.
Signed-off-by: Richard Purdie <rpurdie@rpsys.net> Signed-off-by: Philipp Zabel <philipp.zabel@gmail.com> Signed-off-by: Liam Girdwood <lg@opensource.wolfsonmicro.com> Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com> Signed-off-by: Takashi Iwai <tiwai@suse.de>
Mark Brown [Wed, 28 May 2008 16:58:05 +0000 (17:58 +0100)]
[ALSA] ASoC: Add SOC_DOUBLE_S8_TLV control type
The SOC_DOUBLE_S8_TLV control type was originally implemented in the
UDA1380 driver by Philipp Zabel and was moved into the core by me.
Signed-off-by: Philipp Zabel <philipp.zabel@gmail.com> Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com> Signed-off-by: Takashi Iwai <tiwai@suse.de>
mac80211: fix alignment issue with compare_ether_addr()
This addresses an alignment issue with compare_ether_addr().
The addresses passed to compare_ether_addr should be two bytes aligned.
It may function properly in x86 platform. However may not work properly
on IA-64 or ARM processor.
This also fixes a typo in mlme.c where the sk_buff struct name is incorect.
Though sizeof() works for any incorrect structure pointer name as its just
a pointer length that we want, lets just fix it.
Signed-off-by: Senthil Balasubramanian <senthilkumar@atheros.com> Signed-off-by: Luis R. Rodriguez <lrodriguez@atheros.com> Signed-off-by: John W. Linville <linville@tuxdriver.com>
mac80211: Fix for NULL pointer dereference in sta_info_get()
This addresses a NULL pointer dereference in sta_info_get().
TID and sta_info are extracted in ADDBA Timer expiry function
through the timer handler's argument.
The problem is extracging the TID (which was stored in
timer_to_tid[] array of type "u8") through "int *" typecast which
may also yield unwanted bytes for the MSB of TID that results
in incorrect sta_info and ieee80211_local pointers.
ieee80211_local pointer is NULL as illustrated below, it crashes in
sta_info_get(). The problem started when extracting ieee80211_local
pointer out of sta_info iteself and eventually crashed in
stat_info_get().
The proper way to fix is to change the data type of TID to u8
instead of u16. However changing all the occurences requires
some prototype changes as well. We should fix this in upcoming
patches.
Signed-off-by: Senthil Balasubramanian <senthilkumar@atheros.com> Signed-off-by: Luis Rodriguez <lrodriguez@atheros.com> Signed-off-by: John W. Linville <linville@tuxdriver.com>
Guy Cohen [Tue, 27 May 2008 03:29:35 +0000 (11:29 +0800)]
iwlwifi: fix rate scale TLC column selection bug
This patch fixes a case that a wrong maximal rate is selected when
searching for better configurations.
Signed-off-by: Guy Cohen <guy.cohen@intel.com> Signed-off-by: Tomas Winkler <tomas.winkler@intel.com> Signed-off-by: John W. Linville <linville@tuxdriver.com>
Guy Cohen [Tue, 27 May 2008 03:29:34 +0000 (11:29 +0800)]
iwlwifi: fix exit from stay_in_table state
When exiting from stay in table state (e.g. timer expiration),
all the statistics are reset and the RS flow should not continue
but only after enough statistics are collected again.
Signed-off-by: Guy Cohen <guy.cohen@intel.com> Signed-off-by: Tomas Winkler <tomas.winkler@intel.com> Signed-off-by: John W. Linville <linville@tuxdriver.com>
Scott Ashcroft [Mon, 26 May 2008 21:06:15 +0000 (00:06 +0300)]
rndis_wlan: Make connections to TKIP PSK networks work
This patch allows the rndis_wlan driver to connect to TKIP PSK
networks. It uses the ASSOCIATION_INFORMATION RNDIS call to pull back
the IEs and sends them back to userspace using wireless events. Tested
on a few wireless networks I have access to. Based on the similar
code in ndiswrapper.
Signed-off-by: Scott Ashcroft <scott.ashcroft@talk21.com>
[edit: cleanups] Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Signed-off-by: John W. Linville <linville@tuxdriver.com>
Abhijeet Kolekar [Fri, 23 May 2008 17:15:26 +0000 (10:15 -0700)]
mac80211 : Fixes the status message for iwconfig
iwconfig was showing incorrect status messages when disassociated.
Patch fixes this by always checking for association status in
ioctl calls for getting ap address.
Signed-off-by: Abhijeet Kolekar <abhijeet.kolekar@intel.com> Acked-by: Dan Williams <dcbw@redhat.com> Signed-off-by: John W. Linville <linville@tuxdriver.com>
Ivo van Doorn [Fri, 23 May 2008 16:14:02 +0000 (18:14 +0200)]
rt2x00: Use atomic interface iteration in irq context
rt2x00lib_beacondone() is called from interrupt context,
this means we cannot use the mac80211 interface iterator
that uses the rtnl lock (since that uses a mutex which can sleep).
Instead we should use the atomic mac80211 interface iterator.
Signed-off-by: Ivo van Doorn <IvDoorn@gmail.com> Signed-off-by: John W. Linville <linville@tuxdriver.com>
Ivo van Doorn [Fri, 23 May 2008 16:13:56 +0000 (18:13 +0200)]
rt2x00: Reset antenna RSSI after switch
When the antenna configuration has changed we should reset
the antenna RSSI value. Otherwise the value will be influenced
by the previous configuration quality which in turn will affect
the antenna diversity.
Signed-off-by: Ivo van Doorn <IvDoorn@gmail.com> Signed-off-by: John W. Linville <linville@tuxdriver.com>
Ivo van Doorn [Fri, 23 May 2008 16:13:49 +0000 (18:13 +0200)]
rt2x00: Don't count retries as failure
Link quality estimation became quite low for all rt2x00 drivers
because the number of retries it took to send the frame were
counted as failure.
This does not correspond to the legacy driver link quality calculation,
by not counting it we will send somewhat more optimistic values to
mac80211.
Signed-off-by: Ivo van Doorn <IvDoorn@gmail.com> Signed-off-by: John W. Linville <linville@tuxdriver.com>
Tomas Winkler [Thu, 22 May 2008 22:36:36 +0000 (01:36 +0300)]
mac80211: reorder channel and freq reporting in wext scan report
This patch switch order of channel and freq (SIOCGIWFREQ) reports
in scan results in order to overcome wpa_supplicant inability
to handle channel numbers in 5.2Ghz band.
Wext reporting channel number is ambiguous as channels 7-12 (802.11j)
exist on both bands.
Signed-off-by: Tomas Winkler <tomas.winkler@intel.com> Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com> Acked-by: Dan Williams <dcbw@redhat.com> Signed-off-by: John W. Linville <linville@tuxdriver.com>
Andrea Merello [Sat, 10 May 2008 11:34:16 +0000 (13:34 +0200)]
rtl8180: fix wrong parameter in grf5101_rf_set_channel
The grf5101 RF code needs to invoke grf5101_write_phy_antenna every time the
channel is being switch.
This should be done passing the channel number to that function.
Incorrectly we were passing the same value that is written on the
channel RF register.
This may cause problems when operating on ch 14.
This patch fixes it.
Thanks to Alessandro Di Marco who found this issue!
Signed-off-by: Andrea Merello <andreamrl@tiscali.it> Signed-off-by: John W. Linville <linville@tuxdriver.com>
Andrea Merello [Sat, 10 May 2008 11:32:34 +0000 (13:32 +0200)]
rtl8180: fix wrong parameter in max2820_rf_set_channel
The max2820 RF code needs to invoke max2820_write_phy_antenna every time the
channel is being switch.
This should be done passing the channel number to that function.
Incorrectly we were passing the same value that is written on the
channel RF register.
This may cause problems when operating on ch 14.
This patch fixes it.
Thanks to Alessandro Di Marco who found this issue!
Signed-off-by: Andrea Merello <andreamrl@tiscali.it> Signed-off-by: John W. Linville <linville@tuxdriver.com>
Andrea Merello [Sat, 10 May 2008 11:30:12 +0000 (13:30 +0200)]
rtl8180: fix wrong parameter in sa2400_rf_set_channel
The sa2400 RF code needs to invoke sa2400_write_phy_antenna every time the
channel is being switch.
This should be done passing the channel number to that function.
Incorrectly we were passing the same value that is written on the
channel RF register.
This may cause problems when operating on ch 14.
This patch fixes it.
Thanks to Alessandro Di Marco who found this issue!
Signed-off-by: Andrea Merello <andreamrl@tiscali.it> Signed-off-by: John W. Linville <linville@tuxdriver.com>
John W. Linville [Tue, 27 May 2008 21:01:55 +0000 (17:01 -0400)]
rtl8180: avoid NULL dereference in max2820_rf_set_channel
The static function max2820_rf_set_channel is called with conf == NULL
within its compilation unit. Originally this defaulted to b/g channel
1, but "cfg80211 API for channels/bitrates, mac80211 and driver
conversion" (commit 8318d78a44d49ac1edf2bdec7299de3617c4232e) mistakenly
dropped this check. This patch minimally restores the expected
behavior.
Reported-by: Colin Lai <colin_sh@163.com> Signed-off-by: John W. Linville <linville@tuxdriver.com>
Paul Walmsley [Thu, 22 May 2008 17:33:59 +0000 (11:33 -0600)]
OMAP3 clockdomains: include D2D clockdomain on 34xx ES2+
The OMAP34xx ES2+ TRMs do not document the CLKTRCTRL_D2D bit, even
though the clockdomain logic remains on the chip, and must be
programmed appropriately for the CORE clockdomain to go inactive.
Thanks to Rajendra Najak <rnayak@ti.com> and Richard Woodruff
<r-woodruff2@ti.com> of TI for providing more detail on this issue;
and to Jouni Högander <jouni.hogander@nokia.com> for proposing this
fix.
This patch obsoletes the D2D clockdomain workaround in pm34xx.c in
Jouni's recent patch set.
Signed-off-by: Paul Walmsley <paul@pwsan.com> Signed-off-by: Tony Lindgren <tony@atomide.com>
Paul Walmsley [Thu, 22 May 2008 17:33:54 +0000 (11:33 -0600)]
OMAP3 PM: convert open-coded iterator to use pwrdm_for_each_clkdm()
Convert iterators of the list of clockdomains in a powerdomain to use
pwrdm_for_each_clkdm(), rather than an open-coded for-loop.
This also fixes a bug when the D2D clockdomain is added to the 3430ES2
chip. The open-coded for-loop tested for a NULL pointer in pwrdm_clkdms[]
to determine when to end the loop. This causes the iterator to fall
off the end of the array when the powerdomain contains the maximum number
of clockdomains per powerdomain (currently 3).
Signed-off-by: Paul Walmsley <paul@pwsan.com> Signed-off-by: Tony Lindgren <tony@atomide.com>
The recent powerdomain and clockdomain layer updates for OMAP24XX should
now control the DSP powerdomain and clockdomain automatically when
DSP Gateway calls clk_{enable,disable}(). Remove the manual
clockdomain/powerdomain register tweaks from DSP Gateway.
Signed-off-by: Paul Walmsley <paul@pwsan.com> Signed-off-by: Tony Lindgren <tony@atomide.com>
Paul Walmsley [Wed, 21 May 2008 00:41:52 +0000 (18:41 -0600)]
24xx PM: convert pm24xx.c code to use pwrdm/clkdm interfaces
Convert the existing OMAP24xx pm24xx.c code to use the recent
powerdomain/clockdomain interface code. This mostly consists of removing
direct register reads and writes and replacing it with code that can be
more generic.
Also update some of the comments to better reflect what is actually
happening on the chip.
Signed-off-by: Paul Walmsley <paul@pwsan.com> Signed-off-by: Tony Lindgren <tony@atomide.com>
Paul Walmsley [Wed, 21 May 2008 00:41:35 +0000 (18:41 -0600)]
clockdomain: add clkdm_get_pwrdm()
Add clkdm_get_pwrdm() to the clockdomain code. It will return a
pointer to the powerdomain struct that the clockdomain is contained
within. Used by the PM code.
Signed-off-by: Paul Walmsley <paul@pwsan.com> Signed-off-by: Tony Lindgren <tony@atomide.com>
Paul Walmsley [Wed, 21 May 2008 00:41:33 +0000 (18:41 -0600)]
powerdomain: add pwrdm_get_mem_bank_count()
Add a function pwrdm_get_mem_bank_count() to the powerdomain code to
return the number of memory banks contained inside an OMAP
powerdomain. This will be used by the PM code to configure the memory
bank retention state.
Signed-off-by: Paul Walmsley <paul@pwsan.com> Signed-off-by: Tony Lindgren <tony@atomide.com>
Driver for omap34xx temperature sensor. second version.
This patch adds a hwmon driver for the OMAP34xx internal temperature
sensor. I get unusually high values (80 degrees and more) from this
sensor though.
Signed-off-by: Peter 'p2' De Schrijver <peter.de-schrijver@nokia.com> Signed-off-by: Tony Lindgren <tony@atomide.com>
Antonio Ospite [Wed, 28 May 2008 18:35:52 +0000 (14:35 -0400)]
Input: pxa27x_keypad - miscellaneous fixes
1. Set input bits for direct keys codes
2. Set input bits for rotary encoder codes only if rotary
encoder is enabled
3. Enable EV_REL only if rotary encoder is enabled and rel_codes
are set up
Signed-off-by: Antonio Ospite <ao2@openezx.org> Signed-off-by: Dmitry Torokhov <dtor@mail.ru>
Jason Wessel [Wed, 28 May 2008 17:49:57 +0000 (12:49 -0500)]
kgdbts: Use HW breakpoints with CONFIG_DEBUG_RODATA
Whenever CONFIG_DEBUG_RODATA is set in the kernel config many kernel
text sections become read-only, and the use of software breakpoints in
the kgdb tests will cause the kernel to fail to complete the start up.
Until such time that there is an official API for modifying read-only
text sections hardware breakpoints must be used to run the do_fork or
sys_open tests or the tests get skipped.
Also fix the duplicated include reported by:
Huang Weiyi <weiyi.huang@gmail.com>
Signed-off-by: Jason Wessel <jason.wessel@windriver.com>
Harvey Harrison [Wed, 28 May 2008 17:49:56 +0000 (12:49 -0500)]
kgdb: use common ascii helpers and put_unaligned_be32 helper
Signed-off-by: Harvey Harrison <harvey.harrison@gmail.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Jason Wessel <jason.wessel@windriver.com>
Dave Jones [Wed, 28 May 2008 16:57:13 +0000 (12:57 -0400)]
x86: Fix up silly i1586 boot message.
Trying to boot a 64-bit kernel on a 32bit Pentium 4 gets
you an amusing message along the lines of.
"you need an x86-64, but you only have an i1586"
due to the P4 being family F. Munge it to be 686.
Signed-off-by: Dave Jones <davej@redhat.com> Signed-off-by: H. Peter Anvin <hpa@zytor.com>
Isaku Yamahata [Wed, 28 May 2008 16:41:58 +0000 (09:41 -0700)]
[IA64] pv_ops: move some functions in ivt.S to avoid lack of space.
move interrupt, page_fault, non_syscall, dispatch_unaligned_handler and
dispatch_to_fault_handler to avoid lack of instructin space.
The change set 4dcc29e1574d88f4465ba865ed82800032f76418 bloated
SAVE_MIN_WITH_COVER, SAVE_MIN_WITH_COVER_R19 so that it bloated the
functions which uses those macros.
In the native case, only dispatch_illegal_op_fault had to be moved.
When paravirtualized case the all functions which use the macros need
to be moved to avoid the lack of space.
Signed-off-by: Isaku Yamahata <yamahata@valinux.co.jp> Signed-off-by: Tony Luck <tony.luck@intel.com>
This doesn't work if SLAB is selected and slab debugging is enabled as
these are passed to the preprocessor, and the preprocessor doesn't
understand sizeof.
Linus Torvalds [Wed, 28 May 2008 15:00:51 +0000 (08:00 -0700)]
Merge branch 'for-linus' of git://git.kernel.dk/linux-2.6-block
* 'for-linus' of git://git.kernel.dk/linux-2.6-block:
cfq-iosched: fix RCU problem in cfq_cic_lookup()
block: make blktrace use per-cpu buffers for message notes
Added in elevator switch message to blktrace stream
Added in MESSAGE notes for blktraces
block: reorder cfq_queue to save space on 64bit builds
block: Move the second call to get_request to the end of the loop
splice: handle try_to_release_page() failure
splice: fix sendfile() issue with relay
David Howells [Wed, 28 May 2008 14:36:34 +0000 (15:36 +0100)]
FRV: Specify the minimum slab/kmalloc alignment
Specify the minimum slab/kmalloc alignment to be 8 bytes. This fixes a
crash when SLOB is selected as the memory allocator. The FRV arch needs
this so that it can use the load- and store-double instructions without
faulting. By default SLOB sets the minimum to be 4 bytes.
Signed-off-by: David Howells <dhowells@redhat.com> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Takashi Iwai [Tue, 27 May 2008 15:59:24 +0000 (17:59 +0200)]
[ALSA] Improve the slots option handling
Fix and improve the slots option handling. The sound core tries to
find the slot with the given module name first and assign if it's
still available. If all pre-given slots are unavailable, then try
to find another free slot.
Also, when a module name begins with '!', it means the negative match:
the slot will be given for any modules but that one.
Jens Axboe [Wed, 28 May 2008 12:45:33 +0000 (14:45 +0200)]
block: make blktrace use per-cpu buffers for message notes
Currently it uses a single static char array, but that risks
being corrupted when multiple users issue message notes at the
same time. Make the buffers dynamically allocated when the trace
is setup and make them per-cpu instead.
The default max message size of 1k is also very large, the
interface is mainly for small text notes. So shrink it to 128 bytes.
Zhang, Yanmin [Thu, 22 May 2008 13:13:29 +0000 (15:13 +0200)]
block: Move the second call to get_request to the end of the loop
In function get_request_wait, the second call to get_request could be
moved to the end of the while loop, because if the first call to
get_request fails, the second call will fail without sleep.
The thing is, the crash makes little sense at first sight. We crash on a
benign-looking printk. The code around it got changed in -tip but
checking those topic branches individually did not reproduce the bug.
Which is somewhat surprising, as on native hardware Xen client side
should have little to no side-effects.
After some head scratching, it turns out the following happened:
randconfig enabled the following Xen options:
CONFIG_XEN=y
CONFIG_XEN_MAX_DOMAIN_MEMORY=8
# CONFIG_XEN_BLKDEV_FRONTEND is not set
# CONFIG_XEN_NETDEV_FRONTEND is not set
CONFIG_HVC_XEN=y
# CONFIG_XEN_BALLOON is not set
which activated this piece of code in arch/x86/xen/mmu.c:
> @@ -69,6 +69,13 @@
> __attribute__((section(".data.page_aligned"))) =
> { [ 0 ... TOP_ENTRIES - 1] = &p2m_missing[0] };
>
> +/* Arrays of p2m arrays expressed in mfns used for save/restore */
> +static unsigned long p2m_top_mfn[TOP_ENTRIES]
> + __attribute__((section(".bss.page_aligned")));
> +
> +static unsigned long p2m_top_mfn_list[TOP_ENTRIES / P2M_ENTRIES_PER_PAGE]
> + __attribute__((section(".bss.page_aligned")));
The problem is, you must only put variables into .bss.page_aligned that
have a _size_ that is _exactly_ page aligned. In this case the size of
p2m_top_mfn_list is not page aligned:
So all subsequent variables get unaligned which, depending on luck,
breaks the kernel in various funny ways. In this case what killed the
kernel first was the misaligned bootmap pte page, resulting in that
creative crash above.
Anyway, this was a fun bug to track down :-)
I think the moral is that .bss.page_aligned is a dangerous construct in
its current form, and the symptoms of breakage are very non-trivial, so
i think we need build-time checks to make sure all symbols in
.bss.page_aligned are truly page aligned.
Linus Torvalds [Wed, 28 May 2008 01:47:59 +0000 (18:47 -0700)]
Merge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jbarnes/pci-2.6
* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jbarnes/pci-2.6:
pciehp: add message about pciehp_slot_with_bus option
pci hotplug core: add check of duplicate slot name
pciehp: move msleep after power off
pciehp: poll cmd completion if hotplug interrupt is disabled
pciehp: fix slow probing
pciehp: fix NULL dereference in interrupt handler
shpchp: add message about shpchp_slot_with_bus option
PCI: don't enable ASPM on devices with mixed PCIe/PCI functions
Kenji Kaneshige [Tue, 27 May 2008 10:07:33 +0000 (19:07 +0900)]
pciehp: add message about pciehp_slot_with_bus option
Some (broken?) platform assign the same slot name to multiple hotplug
slots. On such system, slot initialization would fail because of name
collision. The pciehp driver already have a "slot_with_bus" module
option which adds the bus number into the slot name. This patch adds
the message about this module option that will be displayed when slot
name collision is detected.
Kenji Kaneshige [Tue, 27 May 2008 10:07:01 +0000 (19:07 +0900)]
pci hotplug core: add check of duplicate slot name
Fix the following errors reported by Jan C. Nordholz in
http://bugzilla.kernel.org/show_bug.cgi?id=10751.
kobject_add_internal failed for 2 with -EEXIST, don't try to register things with the same name in the same directory.
Pid: 1, comm: swapper Tainted: G W 2.6.26-rc3 #1
[<c0266980>] kobject_add_internal+0x140/0x190
[<c0266afd>] kobject_init_and_add+0x2d/0x40
[<c027bc91>] pci_hp_register+0x81/0x2f0
[<c027fd07>] pciehp_probe+0x1a7/0x470
[<c01b3b84>] sysfs_add_one+0x44/0xa0
[<c01b3c1f>] sysfs_addrm_start+0x3f/0xb0
[<c01b497a>] sysfs_create_link+0x8a/0xf0
[<c0279570>] pcie_port_probe_service+0x50/0x80
[<c02e0545>] driver_sysfs_add+0x55/0x70
[<c02e0662>] driver_probe_device+0x82/0x180
[<c02e07cc>] __driver_attach+0x6c/0x70
[<c02dfe0a>] bus_for_each_dev+0x3a/0x60
[<c05db2d0>] pcied_init+0x0/0x80
[<c02e04e6>] driver_attach+0x16/0x20
[<c02e0760>] __driver_attach+0x0/0x70
[<c02e0341>] bus_add_driver+0x1a1/0x220
[<c05db2d0>] pcied_init+0x0/0x80
[<c02e09cd>] driver_register+0x4d/0x120
[<c05db050>] ibm_acpiphp_init+0x0/0x190
[<c0125aab>] printk+0x1b/0x20
[<c05db2d0>] pcied_init+0x0/0x80
[<c05db2de>] pcied_init+0xe/0x80
[<c05c751a>] kernel_init+0x10a/0x300
[<c0120138>] schedule_tail+0x18/0x50
[<c0103b9a>] ret_from_fork+0x6/0x1c
[<c05c7410>] kernel_init+0x0/0x300
[<c05c7410>] kernel_init+0x0/0x300
[<c010485b>] kernel_thread_helper+0x7/0x1c
=======================
pci_hotplug: Unable to register kobject '2'<3>pciehp: pci_hp_register failed with error -22
Slot with the same name can be registered multiple times if shpchp or
pciehp driver is loaded after acpiphp is loaded because ACPI based
hotplug driver and Native OS hotplug driver trying to handle the same
physical slot. In this case, current pci_hotplug core will call
kobject_init_and_add() muliple time with the same name. This is the
cause of this problem. To fix this problem, this patch adds the check
into pci_hp_register() to see if the slot with the same name.
Kenji Kaneshige [Tue, 27 May 2008 10:06:22 +0000 (19:06 +0900)]
pciehp: move msleep after power off
According to the PCI Express specification, we must wait for at least
1 second after turning power off before taking any action that relies
on power having been removed from the slot/adapter. For this, current
pciehp wait for 1 second after issuing the power off command in
hpc_power_off_slot() function. But waiting for 1 second in
hpc_power_off_slot() can make pciehp probing slow-down because pciehp
probe code calls hpc_power_off_slot() if the slot is not occupied just
in case. We don't need to wait for 1 second at the pciehp probe time
because there is no action on that empty slot. So move 1 second wait
from hpc_power_off_slot() to the caller of hpc_power_off_slot().
Kenji Kaneshige [Tue, 27 May 2008 10:05:26 +0000 (19:05 +0900)]
pciehp: poll cmd completion if hotplug interrupt is disabled
Fix improper long wait for command completion in pciehp probing.
As described in PCI Express specification, software notification is
not generated if the command that occurs as a result of a write to the
Slot Control register that disables software notification of command
completed events. Since pciehp driver doesn't take it into account,
such command is issued in pciehp probing, and it causes improper long
wait for command completion.
This patch changes the pciehp driver to take such command into
account.
Kenji Kaneshige [Tue, 27 May 2008 10:04:30 +0000 (19:04 +0900)]
pciehp: fix slow probing
Fix the "pciehp probing slow" problem reported from Jan C. Nordholz in
http://bugzilla.kernel.org/show_bug.cgi?id=10751.
The command completed bit in Slot Status register applies only to
commands issued to control the attention indicator, power indicator,
power controller, or electromechanical interlock. However, writes to
other parts of the Slot Control register would end up writing to the
control fields. Hence, any write to Slot Control register is
considered as a command. However, if the controller doesn't support
any of attention indicator, power indicator, power controller and
electromechanical interlock, command completed bit would not set in
writing to Slot Control register. In this case, we should not wait for
command completed bit set, otherwise all commands would be considered
not completed in timeout seconds (1 sec.).
The cause of the problem is pciehp driver didn't take this situation
into account. This patch changes pciehp to take it into account. This
patch also add the check for "No Command Completed Support" bit in
Slot Capability register. If it is set, we should not wait for command
completed bit set as well.
This problem seems to be revealed by the commit c27fb883dffe11aa4cb35ecea1fa1832ba45d4da that fixed the bug that
pciehp did not wait for command completed properly (pciehp just
ignored the command completion event).
The situation under which it occurs is hw and timing related: it appears
to happen on a system that has PCI hotplug hardware but with no active
hotplug cards, and another interrupt in the same (shared) IRQ line
arrives too early, before the hotplug-slot entry has been set up - as
triggered by CONFIG_DEBUG_SHIRQ=y:
This patch contains the following two fixes.
(1) Clear all events bits in Slot Status register to prevent the pciehp
driver from detecting the spurious events that would have been occur
before pciehp loading.
(2) Add check whether slot initialization had been already done.
This is short term fix. We need more structural fixes to install
interrupt handler after slot initialization is done.
Kenji Kaneshige [Tue, 27 May 2008 10:08:23 +0000 (19:08 +0900)]
shpchp: add message about shpchp_slot_with_bus option
Some (broken?) platform assign the same slot name to multiple hotplug
slots. On such system, slot initialization would fail because of name
collision. The shpchp driver already have a "slot_with_bus" module
option which adds the bus number into the slot name. This patch adds
the message about this module option that will be displayed when slot
name collision is detected.
projects / linux-2.6-omap-h63xx.git / log