Our goal is to access no more than one cache line access per skb in
a write or receive queue when doing the various walks.
After recent TCP_SKB_CB() reorganizations, it is almost done.
Last part is tcp_skb_pcount() which currently uses
skb_shinfo(skb)->gso_segs, which is a terrible choice, because it needs
3 cache lines in current kernel (skb->head, skb->end, and
shinfo->gso_segs are all in 3 different cache lines, far from skb->cb)
This very simple patch reuses space currently taken by tcp_tw_isn
only in input path, as tcp_skb_pcount is only needed for skb stored in
write queue.
This considerably speeds up tcp_ack(), granted we avoid shinfo->tx_flags
to get SKBTX_ACK_TSTAMP, which seems possible.
This also speeds up all sack processing in general.
This speeds up tcp_sendmsg() because it no longer has to access/dirty
shinfo.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Eric Dumazet says:
====================
tcp: better TCP_SKB_CB layout
TCP had the assumption that IPCB and IP6CB are first members of skb->cb[]
This is fine, except that IPCB/IP6CB are used in TCP for a very short time
in input path.
What really matters for TCP stack is to get skb->next,
TCP_SKB_CB(skb)->seq, and TCP_SKB_CB(skb)->end_seq in the same cache line.
skb that are immediately consumed do not care because whole skb->cb[] is
hot in cpu cache, while skb that sit in wocket write queue or receive queues
do not need TCP_SKB_CB(skb)->header at all.
This patch set implements the prereq for IPv4, IPv6, and TCP to make this
possible. This makes TCP more efficient.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
TCP maintains lists of skb in write queue, and in receive queues
(in order and out of order queues)
Scanning these lists both in input and output path usually requires
access to skb->next, TCP_SKB_CB(skb)->seq, and TCP_SKB_CB(skb)->end_seq
These fields are currently in two different cache lines, meaning we
waste lot of memory bandwidth when these queues are big and flows
have either packet drops or packet reorders.
We can move TCP_SKB_CB(skb)->header at the end of TCP_SKB_CB, because
this header is not used in fast path. This allows TCP to search much faster
in the skb lists.
Even with regular flows, we save one cache line miss in fast path.
Thanks to Christoph Paasch for noticing we need to cleanup
skb->cb[] (IPCB/IP6CB) before entering IP stack in tx path,
and that I forgot IPCB use in tcp_v4_hnd_req() and tcp_v4_save_options().
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
ipv6_opt_accepted() assumes IP6CB(skb) holds the struct inet6_skb_parm
that it needs. Lets not assume this, as TCP stack might use a different
place.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
ip_options_echo() assumes struct ip_options is provided in &IPCB(skb)->opt
Lets break this assumption, but provide a helper to not change all call points.
ip_send_unicast_reply() gets a new struct ip_options pointer.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Steffen Klassert says:
====================
ipv6: Return an error when adding an already existing tunnel
The ipv6 tunnel locate functions should not return an existing
tunnel if create is true. Otherwise it is possible to add the
same tunnel multiple times without getting an error.
All our ipv6 tunnels have this bug from the very beginning.
Only the sit tunnel was fixed some years ago with:
commit 8db99e5717 ("sit: Fail to create tunnel, if it already exists").
This patchset fixes the remaining ipv6 tunnels.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
ip6gre_tunnel_locate() should not return an existing tunnel if
create is true. Otherwise it is possible to add the same
tunnel multiple times without getting an error.
So return NULL if the tunnel that should be created already
exists.
Signed-off-by: Steffen Klassert <steffen.klassert@secunet.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
vti6_locate() should not return an existing tunnel if
create is true. Otherwise it is possible to add the same
tunnel multiple times without getting an error.
So return NULL if the tunnel that should be created already
exists.
Signed-off-by: Steffen Klassert <steffen.klassert@secunet.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
ip6_tnl_locate() should not return an existing tunnel if
create is true. Otherwise it is possible to add the same
tunnel multiple times without getting an error.
So return NULL if the tunnel that should be created already
exists.
Signed-off-by: Steffen Klassert <steffen.klassert@secunet.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
The current open/lock state recovery unfortunately does not handle errors
such as NFS4ERR_CONN_NOT_BOUND_TO_SESSION correctly. Instead of looping,
just proceeds as if the state manager is finished recovering.
This patch ensures that we loop back, handle higher priority errors
and complete the open/lock state recovery.
Cc: stable@vger.kernel.org
Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
If a NFSv4.x server returns NFS4ERR_STALE_CLIENTID in response to a
CREATE_SESSION or SETCLIENTID_CONFIRM in order to tell us that it rebooted
a second time, then the client will currently take this to mean that it must
declare all locks to be stale, and hence ineligible for reboot recovery.
RFC3530 and RFC5661 both suggest that the client should instead rely on the
server to respond to inelegible open share, lock and delegation reclaim
requests with NFS4ERR_NO_GRACE in this situation.
Cc: stable@vger.kernel.org
Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
In e34b731faa ("dma: imx-sdma: Remove spurious __init annotation on
sdma_probe()"), Mark found an extraneous __init label and fixed it.
However, he missed another one, because now we get this other warning:
WARNING: drivers/dma/imx-sdma.o(.text+0x3bb4): Section mismatch in reference from the function sdma_probe() to the function .init.text:sdma_get_firmware()
The function sdma_probe() references
the function __init sdma_get_firmware().
Same reasoning as the last time, the function may get called at
runtime, so it can't be __init.
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: Vinod Koul <vinod.koul@intel.com>
Sharing Last Message (SLM) mode is currently checked in two
functions: ahci_host_activate() and ahci_init_interrupts().
This update consolidates SLM mode check with activation of
multiple MSIs mode.
Signed-off-by: Alexander Gordeev <agordeev@redhat.com>
Signed-off-by: Tejun Heo <tj@kernel.org>
Cc: linux-ide@vger.kernel.org
All byt-max98090 audio connections are known and described in DAPM routing
table. Set the fully_routed flag in order to disable unused codec pins.
Signed-off-by: Jarkko Nikula <jarkko.nikula@linux.intel.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
According to MAX98090 specification "Digital microphone clock (DMC) is
enabled once both data channels are enabled.". Therefore both digital
microphone data channels must be enabled also when using mono microphone
configuration.
Fix this by moving "DMICL_ENA" and "DMICR_ENA" supply widgets from "DMICL"
and "DMICR" inputs to "DMIC Mux" in order to enable both data channels
whenever there is active mono or stereo digital microphone input path.
Use of "DMICL" and "DMICR" inputs are retained for informative source and in
case the driver would find use for exact digital microphone configuration in
the future.
Signed-off-by: Jarkko Nikula <jarkko.nikula@linux.intel.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
Since we cannot make sure the 'val_count' will always be none zero
here, and then if it equals to zero, the kmemdup() will return
ZERO_SIZE_PTR, which equals to ((void *)16).
So this patch fix this with just doing the zero check before calling
kmemdup().
Signed-off-by: Xiubo Li <Li.Xiubo@freescale.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
Cc: stable@vger.kernel.org
If 'map->dev' is NULL and there will lead dev_name() to be NULL pointer
dereference. So before dev_name(), we need to have check of the map->dev
pionter.
We also should make sure that the 'name' pointer shouldn't be NULL for
debugfs_create_dir(). So here using one default "dummy" debugfs name when
the 'name' pointer and 'map->dev' are both NULL.
Signed-off-by: Xiubo Li <Li.Xiubo@freescale.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
Cc: stable@vger.kernel.org
Since we cannot make sure the 'params->num_regs' will always be none
zero here, and then if it equals to zero, the kmemdup() will return
ZERO_SIZE_PTR, which equals to ((void *)16).
So this patch fix this with just doing the zero check before calling
kmemdup().
Signed-off-by: Xiubo Li <Li.Xiubo@freescale.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
Cc: stable@vger.kernel.org
rename st-pwm to pwm-regulator. And support getting voltage & duty table from
device tree, other platforms can also use this driver without any modify.
Signed-off-by: Chris Zhong <zyw@rock-chips.com>
Reviewed-by: Doug Anderson <dianders@chromium.org>
Tested-by: Doug Anderson <dianders@chromium.org>
Signed-off-by: Mark Brown <broonie@kernel.org>
If CONFIG_TTY=n:
drivers/built-in.o: In function `rs_flush_buffer':
amiserial.c:(.text+0x1f626): undefined reference to `tty_wakeup'
drivers/built-in.o: In function `transmit_chars':
amiserial.c:(.text+0x1f6c8): undefined reference to `tty_wakeup'
drivers/built-in.o: In function `change_speed':
amiserial.c:(.text+0x1f80a): undefined reference to `tty_termios_baud_rate'
amiserial.c:(.text+0x1f88c): undefined reference to `tty_termios_baud_rate'
drivers/built-in.o: In function `check_modem_status':
amiserial.c:(.text+0x1faa6): undefined reference to `tty_hangup'
amiserial.c:(.text+0x1faec): undefined reference to `tty_wakeup'
drivers/built-in.o: In function `get_serial_info':
amiserial.c:(.text+0x1fb88): undefined reference to `tty_lock'
amiserial.c:(.text+0x1fbc0): undefined reference to `tty_unlock'
drivers/built-in.o: In function `rs_open':
amiserial.c:(.text+0x1fdc6): undefined reference to `tty_port_block_til_ready'
drivers/built-in.o: In function `set_serial_info':
amiserial.c:(.text+0x1fe0c): undefined reference to `tty_lock'
amiserial.c:(.text+0x1fea0): undefined reference to `tty_unlock'
amiserial.c:(.text+0x1fed0): undefined reference to `tty_unlock'
amiserial.c:(.text+0x1ffaa): undefined reference to `tty_unlock'
drivers/built-in.o: In function `receive_chars':
amiserial.c:(.text+0x2005c): undefined reference to `do_SAK'
amiserial.c:(.text+0x200e0): undefined reference to `tty_insert_flip_string_flags'
amiserial.c:(.text+0x2013c): undefined reference to `tty_insert_flip_string_flags'
amiserial.c:(.text+0x20148): undefined reference to `tty_flip_buffer_push'
drivers/built-in.o: In function `rs_close':
amiserial.c:(.text+0x20744): undefined reference to `tty_port_close_start'
amiserial.c:(.text+0x2078a): undefined reference to `tty_ldisc_flush'
amiserial.c:(.text+0x20798): undefined reference to `tty_port_close_end'
drivers/built-in.o: In function `amiga_serial_probe':
amiserial.c:(.init.text+0x96a): undefined reference to `__tty_alloc_driver'
amiserial.c:(.init.text+0x9c0): undefined reference to `tty_std_termios'
amiserial.c:(.init.text+0x9e6): undefined reference to `tty_set_operations'
amiserial.c:(.init.text+0xa3e): undefined reference to `tty_port_init'
amiserial.c:(.init.text+0xa5c): undefined reference to `tty_port_link_device'
amiserial.c:(.init.text+0xa6c): undefined reference to `tty_register_driver'
amiserial.c:(.init.text+0xb4a): undefined reference to `tty_unregister_driver'
amiserial.c:(.init.text+0xb58): undefined reference to `tty_port_destroy'
amiserial.c:(.init.text+0xb64): undefined reference to `put_tty_driver'
Signed-off-by: Geert Uytterhoeven <geert@linux-m68k.org>
drivers/net/ethernet/cirrus/cs89x0.c: In function ‘cs89x0_ioport_probe’:
drivers/net/ethernet/cirrus/cs89x0.c:1629: error: implicit declaration of function ‘ioport_unmap’
Add the missing ioport_unmap() implementation, and convert ioport_map()
from a macro to a static inline function while we're at it (both copied
from asm-generic).
Signed-off-by: Geert Uytterhoeven <geert@linux-m68k.org>
This fix ensures that we never meet an integer overflow while adding
255 while parsing a variable length encoding. It works differently from
commit 206a81c ("lzo: properly check for overruns") because instead of
ensuring that we don't overrun the input, which is tricky to guarantee
due to many assumptions in the code, it simply checks that the cumulated
number of 255 read cannot overflow by bounding this number.
The MAX_255_COUNT is the maximum number of times we can add 255 to a base
count without overflowing an integer. The multiply will overflow when
multiplying 255 by more than MAXINT/255. The sum will overflow earlier
depending on the base count. Since the base count is taken from a u8
and a few bits, it is safe to assume that it will always be lower than
or equal to 2*255, thus we can always prevent any overflow by accepting
two less 255 steps.
This patch also reduces the CPU overhead and actually increases performance
by 1.1% compared to the initial code, while the previous fix costs 3.1%
(measured on x86_64).
The fix needs to be backported to all currently supported stable kernels.
Reported-by: Willem Pinckaers <willem@lekkertech.net>
Cc: "Don A. Bailey" <donb@securitymouse.com>
Cc: stable <stable@vger.kernel.org>
Signed-off-by: Willy Tarreau <w@1wt.eu>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
This reverts commit 206a81c ("lzo: properly check for overruns").
As analysed by Willem Pinckaers, this fix is still incomplete on
certain rare corner cases, and it is easier to restart from the
original code.
Reported-by: Willem Pinckaers <willem@lekkertech.net>
Cc: "Don A. Bailey" <donb@securitymouse.com>
Cc: stable <stable@vger.kernel.org>
Signed-off-by: Willy Tarreau <w@1wt.eu>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Add a complete description of the LZO format as processed by the
decompressor. I have not found a public specification of this format
hence this analysis, which will be used to better understand the code.
Cc: Willem Pinckaers <willem@lekkertech.net>
Cc: "Don A. Bailey" <donb@securitymouse.com>
Cc: stable <stable@vger.kernel.org>
Signed-off-by: Willy Tarreau <w@1wt.eu>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Pull vfs fixes from Al Viro:
"Assorted fixes + unifying __d_move() and __d_materialise_dentry() +
minimal regression fix for d_path() of victims of overwriting rename()
ported on top of that"
* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs:
vfs: Don't exchange "short" filenames unconditionally.
fold swapping ->d_name.hash into switch_names()
fold unlocking the children into dentry_unlock_parents_for_move()
kill __d_materialise_dentry()
__d_materialise_dentry(): flip the order of arguments
__d_move(): fold manipulations with ->d_child/->d_subdirs
don't open-code d_rehash() in d_materialise_unique()
pull rehashing and unlocking the target dentry into __d_materialise_dentry()
ufs: deal with nfsd/iget races
fuse: honour max_read and max_write in direct_io mode
shmem: fix nlink for rename overwrite directory
Pull cgroup fixes from Tejun Heo:
"This is quite late but these need to be backported anyway.
This is the fix for a long-standing cpuset bug which existed from
2009. cpuset makes use of PF_SPREAD_{PAGE|SLAB} flags to modify the
task's memory allocation behavior according to the settings of the
cpuset it belongs to; unfortunately, when those flags have to be
changed, cpuset did so directly even whlie the target task is running,
which is obviously racy as task->flags may be modified by the task
itself at any time. This obscure bug manifested as corrupt
PF_USED_MATH flag leading to a weird crash.
The bug is fixed by moving the flag to task->atomic_flags. The first
two are prepatory ones to help defining atomic_flags accessors and the
third one is the actual fix"
* 'for-3.17-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/tj/cgroup:
cpuset: PF_SPREAD_PAGE and PF_SPREAD_SLAB should be atomic flags
sched: add macros to define bitops for task atomic flags
sched: fix confusing PFA_NO_NEW_PRIVS constant
Parentheses are missing under an if () statement in
acpi_ac_get_present() which makes the check work differently
from what was intended (at least according to the comment right
above it). Add the missing parens.
The problem was found by sparse.
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Cc: All applicable <stable@vger.kernel.org>
Setting 'flags' to zero will be certainly a misleading way to avoid
warning of 'flags' may be used uninitialized. uninitialized_var is
a correct way because the warning is a false possitive.
Signed-off-by: Xiubo Li <Li.Xiubo@freescale.com>
Signed-off-by: Mike Turquette <mturquette@linaro.org>
The kernel used to contain two functions for length-delimited,
case-insensitive string comparison, strnicmp with correct semantics
and a slightly buggy strncasecmp. The latter is the POSIX name, so
strnicmp was renamed to strncasecmp, and strnicmp made into a wrapper
for the new strncasecmp to avoid breaking existing users.
To allow the compat wrapper strnicmp to be removed at some point in
the future, and to avoid the extra indirection cost, do
s/strnicmp/strncasecmp/g.
Cc: Jens Axboe <axboe@kernel.dk>
Signed-off-by: Rasmus Villemoes <linux@rasmusvillemoes.dk>
Signed-off-by: Jens Axboe <axboe@fb.com>
Pull ARM SoC fixes from Olof Johansson:
"Here's our last set of fixes for 3.17. Most of these are for TI
platforms, fixing some noisy Kconfig issues, runtime clock and power
issues on several platforms and NAND timings on DRA7.
There are also a couple of bug fixes for i.MX, one for QCOM and a
small fix to avoid section mismatch noise on PXA.
Diffstat looks large, partially due to some tables being updated and
thus touching many lines. The qcom gsbi change also restructures
clock management a bit and thus touches a bunch of lines.
All in all, a bit more changes than we'd like at this point, but
nothing stands out as risky either so it seems like the right thing to
send it up now instead of holding it to the merge window"
* tag 'fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/arm/arm-soc:
drivers/soc: qcom: do not disable the iface clock in probe
ARM: imx: fix .is_enabled() of shared gate clock
ARM: OMAP3: Fix I/O chain clock line assertion timed out error
ARM: keystone: dts: fix bindings for pcie and usb clock nodes
bus: omap_l3_noc: Fix connID for OMAP4
ARM: DT: imx53: fix lvds channel 1 port
ARM: dts: cm-t54: fix serial console power supply.
ARM: dts: dra7-evm: Fix NAND GPMC timings
ARM: pxa: fix section mismatch warning for pxa_timer_nodt_init
ARM: OMAP: Fix Kconfig warning for omap1
Pull MIPS fixes from Ralf Baechle:
"The final round of fixes. One corner case in the math emulator and
another one in the mcount function for ftrace"
* 'upstream' of git://git.linux-mips.org/pub/scm/ralf/upstream-linus:
MIPS: mcount: Adjust stack pointer for static trace in MIPS32
MIPS: Fix MFC1 & MFHC1 emulation for 64-bit MIPS systems
Pull x86 fixes from Ingo Molnar:
"This has:
- EFI revert to fix a boot regression
- early_ioremap() fix for boot failure
- KASLR fix for possible boot failures
- EFI fix for corrupted string printing
- remove a misleading EFI bootup 'failed!' error message
Unfortunately it's all rather close to the merge window"
* 'x86-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
x86/efi: Truncate 64-bit values when calling 32-bit OutputString()
x86/efi: Delete misleading efi_printk() error message
Revert "efi/x86: efistub: Move shared dependencies to <asm/efi.h>"
x86/kaslr: Avoid the setup_data area when picking location
x86 early_ioremap: Increase FIX_BTMAPS_SLOTS to 8
Only exchange source and destination filenames
if flags contain RENAME_EXCHANGE.
In case if executable file was running and replaced by
other file /proc/PID/exe should still show correct file name,
not the old name of the file by which it was replaced.
The scenario when this bug manifests itself was like this:
* ALT Linux uses rpm and start-stop-daemon;
* during a package upgrade rpm creates a temporary file
for an executable to rename it upon successful unpacking;
* start-stop-daemon is run subsequently and it obtains
the (nonexistant) temporary filename via /proc/PID/exe
thus failing to identify the running process.
Note that "long" filenames (> DNAiME_INLINE_LEN) are still
exchanged without RENAME_EXCHANGE and this behaviour exists
long enough (should be fixed too apparently).
So this patch is just an interim workaround that restores
behavior for "short" names as it was before changes
introduced by commit da1ce0670c ("vfs: add cross-rename").
See https://lkml.org/lkml/2014/9/7/6 for details.
AV: the comments about being more careful with ->d_name.hash
than with ->d_name.name are from back in 2.3.40s; they
became obsolete by 2.3.60s, when we started to unhash the
target instead of swapping hash chain positions followed
by d_delete() as we used to do when dcache was first
introduced.
Acked-by: Miklos Szeredi <mszeredi@suse.cz>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Alexander Viro <viro@zeniv.linux.org.uk>
Cc: linux-fsdevel@vger.kernel.org
Cc: stable@vger.kernel.org
Fixes: da1ce0670c "vfs: add cross-rename"
Signed-off-by: Mikhail Efremov <sem@altlinux.org>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Allwinner Clocks Additions for 3.18
The most important part of this serie is the addition of the phase API to
handle the MMC clocks in the Allwinner SoCs.
Apart from that, the A23 gained a new mbus driver, and there's a fix for a
incorrect divider table on the APB0 clock.
