| Age | Commit message (Collapse) | Author |
|---|
|
Pull networking fixes from David Miller:
1) Limit the number of can filters to avoid > MAX_ORDER allocations.
Fix from Marc Kleine-Budde.
2) Limit GSO max size in netvsc driver to avoid problems with NVGRE
configurations. From Stephen Hemminger.
3) Return proper error when memory allocation fails in
ser_gigaset_init(), from Dan Carpenter.
4) Missing linkage undo in error paths of ipvlan_link_new(), from Gao
Feng.
5) Missing necessayr SET_NETDEV_DEV in lantiq and cpmac drivers, from
Florian Fainelli.
6) Handle probe deferral properly in smsc911x driver.
* git://git.kernel.org/pub/scm/linux/kernel/git/davem/net:
net: mlx5: Fix Kconfig help text
net: smsc911x: back out silently on probe deferrals
ibmveth: set correct gso_size and gso_type
net: ethernet: cpmac: Call SET_NETDEV_DEV()
net: ethernet: lantiq_etop: Call SET_NETDEV_DEV()
vhost-vsock: fix orphan connection reset
cxgb4/cxgb4vf: Assign netdev->dev_port with port ID
driver: ipvlan: Unlink the upper dev when ipvlan_link_new failed
ser_gigaset: return -ENOMEM on error instead of success
NET: usb: cdc_mbim: add quirk for supporting Telit LE922A
can: peak: fix bad memory access and free sequence
phy: Don't increment MDIO bus refcount unless it's a different owner
netvsc: reduce maximum GSO size
drivers: net: cpsw-phy-sel: Clear RGMII_IDMODE on "rgmii" links
can: raw: raw_setsockopt: limit number of can_filter that can be set
|
|
NFS: NFSoRDMA Client Side Changes
New Features:
- Support for SG_GAP devices
Bugfixes and cleanups:
- Cap size of callback buffer resources
- Improve send queue and RPC metric accounting
- Fix coverity warning
- Avoid calls to ro_unmap_safe()
- Refactor FRMR invalidation
- Error message improvements
|
|
There are two problems with refcounting of auth_gss messages.
First, the reference on the pipe->pipe list (taken by a call
to rpc_queue_upcall()) is not counted. It seems to be
assumed that a message in pipe->pipe will always also be in
pipe->in_downcall, where it is correctly reference counted.
However there is no guaranty of this. I have a report of a
NULL dereferences in rpc_pipe_read() which suggests a msg
that has been freed is still on the pipe->pipe list.
One way I imagine this might happen is:
- message is queued for uid=U and auth->service=S1
- rpc.gssd reads this message and starts processing.
This removes the message from pipe->pipe
- message is queued for uid=U and auth->service=S2
- rpc.gssd replies to the first message. gss_pipe_downcall()
calls __gss_find_upcall(pipe, U, NULL) and it finds the
*second* message, as new messages are placed at the head
of ->in_downcall, and the service type is not checked.
- This second message is removed from ->in_downcall and freed
by gss_release_msg() (even though it is still on pipe->pipe)
- rpc.gssd tries to read another message, and dereferences a pointer
to this message that has just been freed.
I fix this by incrementing the reference count before calling
rpc_queue_upcall(), and decrementing it if that fails, or normally in
gss_pipe_destroy_msg().
It seems strange that the reply doesn't target the message more
precisely, but I don't know all the details. In any case, I think the
reference counting irregularity became a measureable bug when the
extra arg was added to __gss_find_upcall(), hence the Fixes: line
below.
The second problem is that if rpc_queue_upcall() fails, the new
message is not freed. gss_alloc_msg() set the ->count to 1,
gss_add_msg() increments this to 2, gss_unhash_msg() decrements to 1,
then the pointer is discarded so the memory never gets freed.
Fixes: 9130b8dbc6ac ("SUNRPC: allow for upcalls for same uid but different gss service")
Cc: stable@vger.kernel.org
Link: https://bugzilla.opensuse.org/show_bug.cgi?id=1011250
Signed-off-by: NeilBrown <neilb@suse.com>
Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
|
|
We should set the error code if kzalloc() fails.
Fixes: 67cf5b09a46f ("ext4: add the basic function for inline data support")
Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
Cc: stable@vger.kernel.org
|
|
Don't load an inode with a negative size; this causes integer overflow
problems in the VFS.
[ Added EXT4_ERROR_INODE() to mark file system as corrupted. -TYT]
Fixes: a48380f769df (ext4: rename i_dir_acl to i_size_high)
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
Cc: stable@kernel.org
|
|
gcc warns about the newly added driver when phys_addr_t is wider than
a pointer:
drivers/uio/uio_hv_generic.c: In function 'hv_uio_mmap':
drivers/uio/uio_hv_generic.c:71:17: error: cast to pointer from integer of different size [-Werror=int-to-pointer-cast]
virt_to_phys((void *)info->mem[mi].addr) >> PAGE_SHIFT,
drivers/uio/uio_hv_generic.c: In function 'hv_uio_probe':
drivers/uio/uio_hv_generic.c:140:5: error: cast from pointer to integer of different size [-Werror=pointer-to-int-cast]
= (phys_addr_t)dev->channel->ringbuffer_pages;
drivers/uio/uio_hv_generic.c:147:3: error: cast from pointer to integer of different size [-Werror=pointer-to-int-cast]
(phys_addr_t)vmbus_connection.int_page;
drivers/uio/uio_hv_generic.c:153:3: error: cast from pointer to integer of different size [-Werror=pointer-to-int-cast]
(phys_addr_t)vmbus_connection.monitor_pages[1];
I can't see why we store a virtual address in a phys_addr_t here,
as the only user of that variable converts it into a physical
address anyway, so this moves the conversion to where it logically
fits according to the types.
Fixes: 95096f2fbd10 ("uio-hv-generic: new userspace i/o driver for VMBus")
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Acked-by: Stephen Hemminger <sthemmin@microsoft.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
|
|
Fix overflows seen when writing large values into various temperature limit
attributes.
The input value passed to DIV_ROUND_CLOSEST() needs to be clamped to avoid
such overflows.
Reviewed-by: Jean Delvare <jdelvare@suse.de>
Signed-off-by: Guenter Roeck <linux@roeck-us.net>
|
|
Fix overflows seen when writing large values into temperature limit,
voltage limit, and pwm hysteresis attributes.
The input parameter to DIV_ROUND_CLOSEST() needs to be clamped to avoid
such overflows.
Reviewed-by: Jean Delvare <jdelvare@suse.de>
Signed-off-by: Guenter Roeck <linux@roeck-us.net>
|
|
Fix overflows seen when writing large values into voltage limit,
temperature limit, temperature offset, and DAC attributes.
Overflows are seen due to unbound multiplications and additions.
While at it, change the low temperature limit to -128 degrees C,
since this is the minimum temperature accepted by the chip.
Reviewed-by: Jean Delvare <jdelvare@suse.de>
Signed-off-by: Guenter Roeck <linux@roeck-us.net>
|
|
Writes into voltage limit attributes can overflow due to an unbound
multiplication.
Reviewed-by: Jean Delvare <jdelvare@suse.de>
Signed-off-by: Guenter Roeck <linux@roeck-us.net>
|
|
Install the callbacks via the state machine and let the core invoke the
callbacks on the already online CPUs. When the hotplug state is
unregistered the cleanup function is called for each cpu. So both cpu loops
in init() and exit() are not longer required.
Cc: Jean Delvare <jdelvare@suse.com>
Cc: Guenter Roeck <linux@roeck-us.net>
Cc: linux-hwmon@vger.kernel.org
Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Signed-off-by: Guenter Roeck <linux@roeck-us.net>
|
|
Simple hwmon binding documentation.
Signed-off-by: John Muir <john@jmuir.com>
Signed-off-by: Guenter Roeck <linux@roeck-us.net>
|
|
Add support for the TI TMP108 temperature sensor with some device
configuration parameters.
Signed-off-by: John Muir <john@jmuir.com>
[groeck: Initialize of_match_table]
Signed-off-by: Guenter Roeck <linux@roeck-us.net>
|
|
Allocating the sysfs attribute name only if needed and only with the
required minimum length looks optimal, but does not take the additional
overhead for both devm_ data structures and the allocation header itself
into account. This also results in unnecessary memory fragmentation.
Move the sysfs name string into struct hwmon_device_attribute and give it
a sufficient length to reduce this overhead.
Signed-off-by: Guenter Roeck <linux@roeck-us.net>
|
|
The 'groups' parameter of hwmon_device_register_with_info() and
devm_hwmon_device_register_with_info() is only necessary if extra
non-standard attributes need to be provided. Rename the parameter
to extra_groups and clarify the documentation.
Signed-off-by: Guenter Roeck <linux@roeck-us.net>
|
|
A list of sysfs attribute groups is NULL-terminated, so we always need
to allocate data for at least two groups (the dynamically generated group
plus the NULL pointer). Add a comment to explain the situation.
Signed-off-by: Guenter Roeck <linux@roeck-us.net>
|
|
The is_visible callback provides the sysfs attribute mode and is thus
truly mandatory as documented. Check it once at registration and remove
other checks for its existence.
Signed-off-by: Guenter Roeck <linux@roeck-us.net>
|
|
Inform the user that hwmon_device_register() is deprecated,
and suggest conversion to the newest API. Also remove
hwmon_device_register() from the kernel API documentation.
Note that hwmon_device_register() is not marked as __deprecated()
since doing so might result in build errors.
Signed-off-by: Guenter Roeck <linux@roeck-us.net>
|
|
Describing chip attributes as "attributes which apply to the entire chip"
is confusing. Rephrase to "attributes which are not bound to a specific
input or output".
Also rename hwmon_chip_attr_templates[] to hwmon_chip_attrs[] to indicate
that the respective strings strings are not templates but actual attribute
names.
Signed-off-by: Guenter Roeck <linux@roeck-us.net>
|
|
The new API is so far only suited for data attributes and does not work
well for string attributes, specifically for the 'label' attributes.
Provide a separate callback function for those.
Signed-off-by: Guenter Roeck <linux@roeck-us.net>
|
|
The callback descrption in hwmon.h was misleading and stated that read and
write callbacks would be optional. More accurate is is that the callbacks
are mandatory if readable / writeable attributes are present.
Signed-off-by: Guenter Roeck <linux@roeck-us.net>
|
|
The lm90 driver also supports the Texas Instruments TMP451 sensor chip.
Since the Kconfig description for the driver includes a list of all
compatible chips, mention the TI TMP451 there as well.
Signed-off-by: Tobias Klauser <tklauser@distanz.ch>
Signed-off-by: Guenter Roeck <linux@roeck-us.net>
|
|
Keeping track of the per package platform devices requires an extra object,
which is held in a linked list.
The maximum number of packages is known at init() time. So the extra object
and linked list management can be replaced by an array of platform device
pointers in which the per package devices pointers can be stored. Lookup
becomes a simple array lookup instead of a list walk.
The mutex protecting the list can be removed as well because the array is
only accessed from cpu hotplug callbacks which are already serialized.
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Guenter Roeck <linux@roeck-us.net>
|
|
The cpu online callback returns success unconditionally even when the
device has no support, micro code mismatches or device allocation fails.
Only if CPU_HOTPLUG is disabled, the init function checks whether the
device list is empty and removes the driver.
This does not make sense. If CPU HOTPLUG is enabled then there is no point
to keep the driver around when it failed to initialize on the already
online cpus. The chance that not yet online CPUs will provide a functional
interface later is very close to zero.
Add proper error return codes, so the setup of the cpu hotplug states fails
when the device cannot be initialized and remove all the magic cruft.
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Guenter Roeck <linux@roeck-us.net>
|
|
Install the callbacks via the state machine. Setup and teardown are handled
by the hotplug core.
Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Cc: linux-hwmon@vger.kernel.org
Cc: Fenghua Yu <fenghua.yu@intel.com>
Cc: Jean Delvare <jdelvare@suse.com>
Cc: rt@linuxtronix.de
Cc: Guenter Roeck <linux@roeck-us.net>
Link: http://lkml.kernel.org/r/20161117183541.8588-5-bigeasy@linutronix.de
Signed-off-by: Guenter Roeck <linux@roeck-us.net>
|
|
No point in looking up the same thing over and over.
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Guenter Roeck <linux@roeck-us.net>
|
|
The coretemp driver provides a sysfs interface per physical core. If
hyperthreading is enabled and one of the siblings goes offline the sysfs
interface is removed and then immeditately created again for the
sibling. The only difference of them is the target cpu for the
rdmsr_on_cpu() in the sysfs show functions.
It's way simpler to keep a cpumask of cpus which are active in a package
and only remove the interface when the last sibling goes offline. Otherwise
just move the target cpu for the sysfs show functions to the still online
sibling.
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Guenter Roeck <linux@roeck-us.net>
|
|
When a CPU is offlined nothing checks whether it is the target CPU for the
package temperature sysfs interface.
As a consequence all future readouts of the package temperature return
crap:
90000
which is Tjmax of that package.
Check whether the outgoing CPU is the target for the package and assign it
to some other still online CPU in the package. Protect the change against
the rdmsr_on_cpu() in show_crit_alarm().
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Guenter Roeck <linux@roeck-us.net>
|
|
Module test reports overflows when writing into temperature and voltage
limit attributes
temp1_min: Suspected overflow: [127000 vs. 0]
temp1_max: Suspected overflow: [127000 vs. 0]
temp1_offset: Suspected overflow: [127000 vs. 0]
temp2_min: Suspected overflow: [127000 vs. 0]
temp2_max: Suspected overflow: [127000 vs. 0]
temp2_offset: Suspected overflow: [127000 vs. 0]
temp3_min: Suspected overflow: [127000 vs. 0]
temp3_max: Suspected overflow: [127000 vs. 0]
temp3_offset: Suspected overflow: [127000 vs. 0]
in0_min: Suspected overflow: [3320 vs. 0]
in0_max: Suspected overflow: [3320 vs. 0]
in4_min: Suspected overflow: [15938 vs. 0]
in4_max: Suspected overflow: [15938 vs. 0]
in6_min: Suspected overflow: [1992 vs. 0]
in6_max: Suspected overflow: [1992 vs. 0]
in7_min: Suspected overflow: [2391 vs. 0]
in7_max: Suspected overflow: [2391 vs. 0]
The problem is caused by conversions from unsigned long to long and
from long to int.
Reviewed-by: Jean Delvare <jdelvare@suse.de>
Signed-off-by: Guenter Roeck <linux@roeck-us.net>
|
|
Signed-off-by: Madalin Bucur <madalin.bucur@nxp.com>
Signed-off-by: Scott Wood <oss@buserror.net>
|
|
The alias is used by the boot loader to perform a device tree
fixup.
Signed-off-by: Madalin Bucur <madalin.bucur@nxp.com>
Signed-off-by: Scott Wood <oss@buserror.net>
|
|
Signed-off-by: Madalin Bucur <madalin.bucur@nxp.com>
Signed-off-by: Scott Wood <oss@buserror.net>
|
|
Signed-off-by: Madalin Bucur <madalin.bucur@nxp.com>
Signed-off-by: Scott Wood <oss@buserror.net>
|
|
Signed-off-by: Fabian Frederick <fabf@skynet.be>
Signed-off-by: Scott Wood <oss@buserror.net>
|
|
Signed-off-by: Fabian Frederick <fabf@skynet.be>
Signed-off-by: Scott Wood <oss@buserror.net>
|
|
8xx uses a two level page table with two different linux page size
support (4k and 16k). 8xx also support two different hugepage sizes
512k and 8M. In order to support them on linux we define two different
page table layout.
The size of pages is in the PGD entry, using PS field (bits 28-29):
00 : Small pages (4k or 16k)
01 : 512k pages
10 : reserved
11 : 8M pages
For 512K hugepage size a pgd entry have the below format
[<hugepte address >0101] . The hugepte table allocated will contain 8
entries pointing to 512K huge pte in 4k pages mode and 64 entries in
16k pages mode.
For 8M in 16k mode, a pgd entry have the below format
[<hugepte address >1101] . The hugepte table allocated will contain 8
entries pointing to 8M huge pte.
For 8M in 4k mode, multiple pgd entries point to the same hugepte
address and pgd entry will have the below format
[<hugepte address>1101]. The hugepte table allocated will only have one
entry.
For the time being, we do not support CPU15 ERRATA when HUGETLB is
selected
Signed-off-by: Christophe Leroy <christophe.leroy@c-s.fr>
Reviewed-by: Aneesh Kumar K.V <aneesh.kumar@linux.vnet.ibm.com> (v3, for the generic bits)
Signed-off-by: Scott Wood <oss@buserror.net>
|
|
Today there are two implementations of hugetlbpages which are managed
by exclusive #ifdefs:
* FSL_BOOKE: several directory entries points to the same single hugepage
* BOOK3S: one upper level directory entry points to a table of hugepages
In preparation of implementation of hugepage support on the 8xx, we
need a mix of the two above solutions, because the 8xx needs both cases
depending on the size of pages:
* In 4k page size mode, each PGD entry covers a 4M bytes area. It means
that 2 PGD entries will be necessary to cover an 8M hugepage while a
single PGD entry will cover 8x 512k hugepages.
* In 16 page size mode, each PGD entry covers a 64M bytes area. It means
that 8x 8M hugepages will be covered by one PGD entry and 64x 512k
hugepages will be covers by one PGD entry.
This patch:
* removes #ifdefs in favor of if/else based on the range sizes
* merges the two huge_pte_alloc() functions as they are pretty similar
* merges the two hugetlbpage_init() functions as they are pretty similar
Signed-off-by: Christophe Leroy <christophe.leroy@c-s.fr>
Reviewed-by: Aneesh Kumar K.V <aneesh.kumar@linux.vnet.ibm.com> (v3)
Signed-off-by: Scott Wood <oss@buserror.net>
|
|
Today powerpc64 uses a set of pgtable_caches while powerpc32 uses
standard pages when using 4k pages and a single pgtable_cache
if using other size pages.
In preparation of implementing huge pages on the 8xx, this patch
replaces the specific powerpc32 handling by the 64 bits approach.
This is done by:
* moving 64 bits pgtable_cache_add() and pgtable_cache_init()
in a new file called init-common.c
* modifying pgtable_cache_init() to also handle the case
without PMD
* removing the 32 bits version of pgtable_cache_add() and
pgtable_cache_init()
* copying related header contents from 64 bits into both the
book3s/32 and nohash/32 header files
On the 8xx, the following cache sizes will be used:
* 4k pages mode:
- PGT_CACHE(10) for PGD
- PGT_CACHE(3) for 512k hugepage tables
* 16k pages mode:
- PGT_CACHE(6) for PGD
- PGT_CACHE(7) for 512k hugepage tables
- PGT_CACHE(3) for 8M hugepage tables
Signed-off-by: Christophe Leroy <christophe.leroy@c-s.fr>
Reviewed-by: Aneesh Kumar K.V <aneesh.kumar@linux.vnet.ibm.com>
Signed-off-by: Scott Wood <oss@buserror.net>
|
|
Clients can set the umask attribute when creating files to cause the
server to apply it always except when inheriting permissions from the
parent directory. That way, the new files will end up with the same
permissions as files created locally.
See https://tools.ietf.org/html/draft-ietf-nfsv4-umask-02 for more details.
Signed-off-by: Andreas Gruenbacher <agruenba@redhat.com>
Signed-off-by: J. Bruce Fields <bfields@redhat.com>
Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
|
|
Since the following commit, Infiniband and Ethernet have not been
mutually exclusive.
Fixes: 4aa17b28 mlx5: Enable mutual support for IB and Ethernet
Signed-off-by: Christopher Covington <cov@codeaurora.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
It seems attackers can also send UDP packets with no payload at all.
skb_condense() can still be a win in this case.
It will be possible to replace the custom code in tcp_add_backlog()
to get full benefit from skb_condense()
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
When trying to get a regulator we may get deferred and we see
this noise:
smsc911x 1b800000.ethernet-ebi2 (unnamed net_device) (uninitialized):
couldn't get regulators -517
Then the driver continues anyway. Which means that the regulator
may not be properly retrieved and reference counted, and may be
switched off in case noone else is using it.
Fix this by returning silently on deferred probe and let the
system work it out.
Cc: Jeremy Linton <jeremy.linton@arm.com>
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
git://git.kernel.org/pub/scm/linux/kernel/git/jberg/mac80211-next
Johannes Berg says:
====================
Three fixes:
* fix a logic bug introduced by a previous cleanup
* fix nl80211 attribute confusing (trying to use
a single attribute for two purposes)
* fix a long-standing BSS leak that happens when an
association attempt is abandoned
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
This patch is based on an earlier one submitted
by Jon Maxwell with the following commit message:
"We recently encountered a bug where a few customers using ibmveth on the
same LPAR hit an issue where a TCP session hung when large receive was
enabled. Closer analysis revealed that the session was stuck because the
one side was advertising a zero window repeatedly.
We narrowed this down to the fact the ibmveth driver did not set gso_size
which is translated by TCP into the MSS later up the stack. The MSS is
used to calculate the TCP window size and as that was abnormally large,
it was calculating a zero window, even although the sockets receive buffer
was completely empty."
We rely on the Virtual I/O Server partition in a pseries
environment to provide the MSS through the TCP header checksum
field. The stipulation is that users should not disable checksum
offloading if rx packet aggregation is enabled through VIOS.
Some firmware offerings provide the MSS in the RX buffer.
This is signalled by a bit in the RX queue descriptor.
Reviewed-by: Brian King <brking@linux.vnet.ibm.com>
Reviewed-by: Pradeep Satyanarayana <pradeeps@linux.vnet.ibm.com>
Reviewed-by: Marcelo Ricardo Leitner <marcelo.leitner@gmail.com>
Reviewed-by: Jonathan Maxwell <jmaxwell37@gmail.com>
Reviewed-by: David Dai <zdai@us.ibm.com>
Signed-off-by: Thomas Falcon <tlfalcon@linux.vnet.ibm.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
Eric Dumazet says:
====================
udp: receive path optimizations
This patch series provides about 100 % performance increase under flood.
v2: added Paolo feedback on udp_rmem_release() for tiny sk_rcvbuf
added the last patch touching sk_rmem_alloc later
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
In flood situations, keeping sk_rmem_alloc at a high value
prevents producers from touching the socket.
It makes sense to lower sk_rmem_alloc only at the end
of udp_rmem_release() after the thread draining receive
queue in udp_recvmsg() finished the writes to sk_forward_alloc.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
If udp_recvmsg() constantly releases sk_rmem_alloc
for every read packet, it gives opportunity for
producers to immediately grab spinlocks and desperatly
try adding another packet, causing false sharing.
We can add a simple heuristic to give the signal
by batches of ~25 % of the queue capacity.
This patch considerably increases performance under
flood by about 50 %, since the thread draining the queue
is no longer slowed by false sharing.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
In UDP RX handler, we currently clear skb->dev before skb
is added to receive queue, because device pointer is no longer
available once we exit from RCU section.
Since this first cache line is always hot, lets reuse this space
to store skb->truesize and thus avoid a cache line miss at
udp_recvmsg()/udp_skb_destructor time while receive queue
spinlock is held.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
Idea of busylocks is to let producers grab an extra spinlock
to relieve pressure on the receive_queue spinlock shared by consumer.
This behavior is requested only once socket receive queue is above
half occupancy.
Under flood, this means that only one producer can be in line
trying to acquire the receive_queue spinlock.
These busylock can be allocated on a per cpu manner, instead of a
per socket one (that would consume a cache line per socket)
This patch considerably improves UDP behavior under stress,
depending on number of NIC RX queues and/or RPS spread.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
Timur Tabi says:
====================
net: qcom/emac: simplify support for different SOCs
On SOCs that have the Qualcomm EMAC network controller, the internal
PHY block is always different. Sometimes the differences are small,
sometimes it might be a completely different IP. Either way, using version
numbers to differentiate them and putting all of the init code in one
file does not scale.
This patchset does two things: The first breaks up the current code into
different files, and the second patch adds support for a third SOC, the
Qualcomm Technologies QDF2400 ARM Server SOC.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
|
