| Age | Commit message (Collapse) | Author | Files | Lines |
|---|
|
This reverts commit 01ab1167cd2d861d20195eda08505652c536df97, it is
incorrect, a future patch will fix this up properly.
Cc: Maxime Ripard <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
|
|
There have been reports of EFI crashes since -rc1. The following two
commits fix known issues.
* Fix boot failure on 32-bit EFI due to the recent EFI memmap changes
merged during the merge window - Borislav Petkov
* Avoid a crash during efi_bgrt_init() by detecting invalid BGRT
headers based on the 'status' field.
Signed-off-by: H. Peter Anvin <[email protected]>
|
|
Pull KVM fixes from Paolo Bonzini:
"A small error handling problem and a compile breakage for ARM64"
* tag 'for-linus' of git://git.kernel.org/pub/scm/virt/kvm/kvm:
arm64: KVM: Add VGIC device control for arm64
KVM: return an error code in kvm_vm_ioctl_register_coalesced_mmio()
|
|
git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip
Pull x86 fixes from Peter Anvin:
"A collection of small fixes:
- There still seem to be problems with asm goto which requires the
empty asm hack.
- If SMAP is disabled at compile time, don't enable it nor try to
interpret a page fault as an SMAP violation.
- Fix a case of unbounded recursion while tracing"
* 'x86-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
x86, smap: smap_violation() is bogus if CONFIG_X86_SMAP is off
x86, smap: Don't enable SMAP if CONFIG_X86_SMAP is disabled
compiler/gcc4: Make quirk for asm_volatile_goto() unconditional
x86: Use preempt_disable_notrace() in cycles_2_ns()
|
|
git://git.kernel.org/pub/scm/linux/kernel/git/rafael/linux-pm
Pull ACPI and power management fixes from Rafael Wysocki:
"These include a fix for a recent intel_pstate regression, a fix for a
regression in the ACPI-based PCI hotplug (ACPIPHP) code introduced
during the 3.12 cycle, fixes for two bugs in the ACPI core introduced
recently and a MAINTAINERS update related to cpufreq.
Specifics:
- Fix for a recent regression in the intel_pstate driver that
introduced a race condition causing systems to crash during
initialization in some situations. This removes the affected code
altogether. From Dirk Brandewie.
- ACPIPHP fix for a regression introduced during the 3.12 cycle
causing devices to be dropped as a result of bus check
notifications after system resume on some systems due to the way
ACPIPHP interprets _STA return values (arguably incorrectly). From
Mika Westerberg.
- ACPI dock driver fix for a problem causing docking to fail due to a
check that always fails after recent ACPI core changes (found by
code inspection).
- ACPI container driver fix to prevent memory from being leaked in an
error code path after device_register() failures.
- Update of the arm_big_little cpufreq driver maintainer's e-mail
address"
* tag 'pm+acpi-3.14-rc3' of git://git.kernel.org/pub/scm/linux/kernel/git/rafael/linux-pm:
MAINTAINERS / cpufreq: update Sudeep's email address
intel_pstate: Remove energy reporting from pstate_sample tracepoint
ACPI / container: Fix error code path in container_device_attach()
ACPI / hotplug / PCI: Relax the checking of _STA return values
ACPI / dock: Use acpi_device_enumerated() to check if dock is present
|
|
Pull EDAC fixes from Borislav Petkov:
"Fix polling timeout setting through sysfs.
You're surely wondering why the patches are not based on an rc. Well,
Andrew sent you 79040cad3f82 ("drivers/edac/edac_mc_sysfs.c: poll
timeout cannot be zero sent you") already (it got in in -rc2) but it
is not enough as a fix because for one, setting too low polling
intervals (< 1sec) don't make any sense and cause unnecessary polling
load on the system.
Then, even if we set some interval, we explode with
[ 4143.094342] WARNING: CPU: 1 PID: 0 at kernel/workqueue.c:1393 __queue_work+0x1d7/0x340()
because the workqueue setup path is used also for the timeout period
resetting and we're doing INIT_DELAYED_WORK() on an already active
workqueue. Which is total bollocks. So this is taken care of by the
second patch.
I've CCed stable for those two"
* tag 'edac_for_3.14' of git://git.kernel.org/pub/scm/linux/kernel/git/bp/bp:
EDAC: Correct workqueue setup path
EDAC: Poll timeout cannot be zero, p2
|
|
git://git.kernel.org/pub/scm/linux/kernel/git/tomba/linux
Pull fbdev fixes from Tomi Valkeinen:
"Minor fbdev fixes for 3.14"
* tag 'fbdev-fixes-3.14' of git://git.kernel.org/pub/scm/linux/kernel/git/tomba/linux:
video: Kconfig: Allow more broad selection of the imxfb framebuffer driver.
video: exynos: Fix S6E8AX0 LCD driver build error
OMAPDSS: fix fck field types
OMAPDSS: DISPC: decimation rounding fix
|
|
Pull block IO fixes from Jens Axboe:
"Second round of updates and fixes for 3.14-rc2. Most of this stuff
has been queued up for a while. The notable exception is the blk-mq
changes, which are naturally a bit more in flux still.
The pull request contains:
- Two bug fixes for the new immutable vecs, causing crashes with raid
or swap. From Kent.
- Various blk-mq tweaks and fixes from Christoph. A fix for
integrity bio's from Nic.
- A few bcache fixes from Kent and Darrick Wong.
- xen-blk{front,back} fixes from David Vrabel, Matt Rushton, Nicolas
Swenson, and Roger Pau Monne.
- Fix for a vec miscount with integrity vectors from Martin.
- Minor annotations or fixes from Masanari Iida and Rashika Kheria.
- Tweak to null_blk to do more normal FIFO processing of requests
from Shlomo Pongratz.
- Elevator switching bypass fix from Tejun.
- Softlockup in blkdev_issue_discard() fix when !CONFIG_PREEMPT from
me"
* 'for-linus' of git://git.kernel.dk/linux-block: (31 commits)
block: add cond_resched() to potentially long running ioctl discard loop
xen-blkback: init persistent_purge_work work_struct
blk-mq: pair blk_mq_start_request / blk_mq_requeue_request
blk-mq: dont assume rq->errors is set when returning an error from ->queue_rq
block: Fix cloning of discard/write same bios
block: Fix type mismatch in ssize_t_blk_mq_tag_sysfs_show
blk-mq: rework flush sequencing logic
null_blk: use blk_complete_request and blk_mq_complete_request
virtio_blk: use blk_mq_complete_request
blk-mq: rework I/O completions
fs: Add prototype declaration to appropriate header file include/linux/bio.h
fs: Mark function as static in fs/bio-integrity.c
block/null_blk: Fix completion processing from LIFO to FIFO
block: Explicitly handle discard/write same segments
block: Fix nr_vecs for inline integrity vectors
blk-mq: Add bio_integrity setup to blk_mq_make_request
blk-mq: initialize sg_reserved_size
blk-mq: handle dma_drain_size
blk-mq: divert __blk_put_request for MQ ops
blk-mq: support at_head inserations for blk_execute_rq
...
|
|
git://git.kernel.org/pub/scm/linux/kernel/git/tiwai/sound
Pull sound fixes from Takashi Iwai:
"Again only fixes for HD-audio:
- regression fixes due to the modularization
- a few fixups for Dell, Sony and HP laptops
- a revert of the previous fix as it leads to another regression"
* tag 'sound-3.14-rc3' of git://git.kernel.org/pub/scm/linux/kernel/git/tiwai/sound:
ALSA: Revert "ALSA: hda/realtek - Avoid invalid COEFs for ALC271X"
ALSA: hda - Fix undefined symbol due to builtin/module mixup
ALSA: hda - Fix mic capture on Sony VAIO Pro 11
ALSA: hda - Add a headset quirk for Dell XPS 13
ALSA: hda - Fix inconsistent Mic mute LED
ALSA: hda - Fix leftover ifdef checks after modularization
|
|
git://git.kernel.org/pub/scm/linux/kernel/git/roland/infiniband
Pull RDMA/InfiniBand fixes from Roland Dreier:
- Fix some rough edges from the "IP addressing for IBoE" merge
- Other misc fixes, mostly to hardware drivers
* tag 'rdma-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/roland/infiniband: (21 commits)
RDMA/ocrdma: Fix load time panic during GID table init
RDMA/ocrdma: Fix traffic class shift
IB/iser: Fix use after free in iser_snd_completion()
IB/iser: Avoid dereferencing iscsi_iser conn object when not bound to iser connection
IB/usnic: Fix smatch endianness error
IB/mlx5: Remove dependency on X86
mlx5: Add include of <linux/slab.h> because of kzalloc()/kfree() use
IB/qib: Add missing serdes init sequence
RDMA/cxgb4: Add missing neigh_release in LE-Workaround path
IB: Report using RoCE IP based gids in port caps
IB/mlx4: Build the port IBoE GID table properly under bonding
IB/mlx4: Do IBoE GID table resets per-port
IB/mlx4: Do IBoE locking earlier when initializing the GID table
IB/mlx4: Move rtnl locking to the right place
IB/mlx4: Make sure GID index 0 is always occupied
IB/mlx4: Don't allocate range of steerable UD QPs for Ethernet-only device
RDMA/amso1100: Fix error return code
RDMA/nes: Fix error return code
IB/mlx5: Don't set "block multicast loopback" capability
IB/mlx5: Fix binary compatibility with libmlx5
...
|
|
git://git.kernel.org/pub/scm/linux/kernel/git/groeck/linux-staging
Pull hwmon fix from Guenter Roeck:
"Fix arithmetic overflow in ntc_thermistor driver"
* tag 'hwmon-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/groeck/linux-staging:
hwmon: (ntc_thermistor) Avoid math overflow
|
|
Pull battery fixes from Dmitry Eremin-Solenikov:
- Fix NULL pointer dereference in max17040 driver
- Add bq2415x dts bindings documentation
- Fix misleading comment in ds2782 driver
- Remove useless check in isp1704 charger driver.
* tag 'for-v3.14-fixes' of git://git.infradead.org/battery-2.6:
power: max17040: Fix NULL pointer dereference when there is no platform_data
dt: binding documentation for bq2415x charger
isp1704_charger: remove useless check in isp1704_charger_probe()
power: ds2782_battery: Typo in comment
|
|
'ocrdma', 'qib' and 'usnic' into for-next
|
|
We should use rdma_vlan_dev_real_dev() instead of using vlan_dev_real_dev()
when building the GID table for a vlan interface.
Signed-off-by: Devesh Sharma <[email protected]>
Signed-off-by: Roland Dreier <[email protected]>
|
|
Use correct value for obtaining traffic class from device
response for Query QP request.
Signed-off-by: Devesh Sharma <[email protected]>
Reported-by: Dan Carpenter <[email protected]>
Signed-off-by: Roland Dreier <[email protected]>
|
|
We use "tx_desc" again after we free it.
Signed-off-by: Dan Carpenter <[email protected]>
Acked-by: Or Gerlitz <[email protected]>
Signed-off-by: Roland Dreier <[email protected]>
|
|
connection
Fix a possible NULL pointer dereference in disconnection flow. This
can happen if the target disconnected/rejected the connection request,
e.g before the binding stage between iscsi connection to the transport
connection.
Signed-off-by: Alex Tabachnik <[email protected]>
Signed-off-by: Roi Dayan <[email protected]>
Signed-off-by: Or Gerlitz <[email protected]>
Signed-off-by: Roland Dreier <[email protected]>
|
|
Error reported at http://marc.info/?l=linux-rdma&m=138995755801039&w=2
Fix short to int cast for big endian systems.
Signed-off-by: Upinder Malhi <[email protected]>
Signed-off-by: Roland Dreier <[email protected]>
|
|
This fixes the build breakage introduced by
c07a0191ef2de1f9510f12d1f88e3b0b5cd8d66f and adds support for the device
control API and save/restore of the VGIC state for ARMv8.
The defines were simply missing from the arm64 header files and
uaccess.h must be implicitly imported from somewhere else on arm.
Signed-off-by: Christoffer Dall <[email protected]>
Signed-off-by: Paolo Bonzini <[email protected]>
|
|
Madper reported seeing the following crash,
BUG: unable to handle kernel paging request at ffffffffff340003
IP: [<ffffffff81d85ba4>] efi_bgrt_init+0x9d/0x133
Call Trace:
[<ffffffff81d8525d>] efi_late_init+0x9/0xb
[<ffffffff81d68f59>] start_kernel+0x436/0x450
[<ffffffff81d6892c>] ? repair_env_string+0x5c/0x5c
[<ffffffff81d68120>] ? early_idt_handlers+0x120/0x120
[<ffffffff81d685de>] x86_64_start_reservations+0x2a/0x2c
[<ffffffff81d6871e>] x86_64_start_kernel+0x13e/0x14d
This is caused because the layout of the ACPI BGRT header on this system
doesn't match the definition from the ACPI spec, and so we get a bogus
physical address when dereferencing ->image_address in efi_bgrt_init().
Luckily the status field in the BGRT header clearly marks it as invalid,
so we can check that field and skip BGRT initialisation.
Reported-by: Madper Xie <[email protected]>
Suggested-by: Toshi Kani <[email protected]>
Cc: "Rafael J. Wysocki" <[email protected]>
Cc: Matthew Garrett <[email protected]>
Cc: Josh Triplett <[email protected]>
Signed-off-by: Matt Fleming <[email protected]>
|
|
We're using edac_mc_workq_setup() both on the init path, when
we load an edac driver and when we change the polling period
(edac_mc_reset_delay_period) through /sys/.../edac_mc_poll_msec.
On that second path we don't need to init the workqueue which has been
initialized already.
Thanks to Tejun for workqueue insights.
Signed-off-by: Borislav Petkov <[email protected]>
Link: http://lkml.kernel.org/r/[email protected]
Cc: <[email protected]>
|
|
Sanitize code even more to accept unsigned longs only and to not allow
polling intervals below 1 second as this is unnecessary and doesn't make
much sense anyway for polling errors.
Signed-off-by: Borislav Petkov <[email protected]>
Link: http://lkml.kernel.org/r/[email protected]
Cc: Doug Thompson <[email protected]>
Cc: <[email protected]>
|
|
We do not enable the new efi memmap on 32-bit and thus we need to run
runtime_code_page_mkexec() unconditionally there. Fix that.
Reported-and-tested-by: Lejun Zhu <[email protected]>
Signed-off-by: Borislav Petkov <[email protected]>
Signed-off-by: Matt Fleming <[email protected]>
|
|
Without that patch, a user can't select the imxfb driver when the i.MX25
and/or the i.MX27 device tree board are selected and that no boards that
selects IMX_HAVE_PLATFORM_IMX_FB are compiled in.
Cc: Eric Bénard <[email protected]>
Cc: Jean-Christophe Plagniol-Villard <[email protected]>
Cc: Sascha Hauer <[email protected]>
Cc: Tomi Valkeinen <[email protected]>
Cc: [email protected]
Cc: [email protected]
Signed-off-by: Denis Carikli <[email protected]>
Acked-by: Shawn Guo <[email protected]>
Acked-by: Jean-Christophe PLAGNIOL-VILLARD <[email protected]>
Acked-by: Sascha Hauer <[email protected]>
Signed-off-by: Tomi Valkeinen <[email protected]>
|
|
Enable S6E8AX0 LCD driver only if LCD_CLASS_DEVICE is a built-in driver.
Else we get the following errors due to missing symbols:
drivers/built-in.o: In function `s6e8ax0_probe':
:(.text+0x51aec): undefined reference to `lcd_device_register'
:(.text+0x51c44): undefined reference to `lcd_device_unregister'
Signed-off-by: Sachin Kamat <[email protected]>
Signed-off-by: Tomi Valkeinen <[email protected]>
|
|
Even if the 'time_before' macro expand with parentheses, the look is bad.
Signed-off-by: Francois-Xavier Le Bail <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
|
|
Remove Kconfig dependency of mlx5_ib/mlx5_core on X86, since there is
no such dependency in reality.
Signed-off-by: Eli Cohen <[email protected]>
Signed-off-by: Michael Neuling <[email protected]>
Signed-off-by: Roland Dreier <[email protected]>
|
|
On some architectures (for example, arm), we don't end up indirectly
pulling in the declaration of kzalloc() and kfree(), and so building
anything that includes <linux/mlx5/driver.h> breaks. Fix this by adding
an explicit include to get the declaration.
Reported-by: kbuild test robot <[email protected]>
Signed-off-by: Roland Dreier <[email protected]>
|
|
The ntc thermistor code was doing math whose temporary result might
have overflowed 32-bits. We need some casts in there to make it safe.
In one example I found:
- pullup_uV: 1800000
- result of iio_read_channel_raw: 3226
- 1800000 * 3226 => 0x15a1cbc80
Signed-off-by: Doug Anderson <[email protected]>
Cc: [email protected] # 3.10+
Signed-off-by: Guenter Roeck <[email protected]>
|
|
phy_init_hw not does a full PHY reset after the driver probe has
finished, so any hw initialization done in the probe will be lost.
Part of the timestamping functionality of the dp83640 is set up in the
probe and with that lost, enabling timestamping will cause a PHY
lockup, requiring a hard reset / power cycle to recover.
This patch moves all the HW initialization in dp83640_probe to
dp83640_config_init.
Signed-off-by: Stefan Sørensen <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
|
|
Commit 0cd8f9cc0654c06adde353c6532114c5f53a18e8 ("drivers: net: cpsw:
enable promiscuous mode support")
Enable promiscuous mode support for CPSW.
Introduced a crash on an am335x based board (similiar to am335x-evm).
Reason is buggy end condition in for loop in cpsw_set_promiscious()
for (i = 0; i <= priv->data.slaves; i++)
should be
for (i = 0; i < priv->data.slaves; i++)
Fix this ...
Signed-off-by: Heiko Schocher <[email protected]>
Cc: Mugunthan V N <[email protected]>
Cc: David S. Miller <[email protected]>
Cc: Sebastian Siewior <[email protected]>
Cc: Daniel Mack <[email protected]>
Cc: Felipe Balbi <[email protected]>
Cc: Markus Pargmann <[email protected]>
Cc: [email protected]
Cc: [email protected]
Acked-by: Mugunthan V N <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
|
|
vhost_zerocopy_callback accesses VQ right after it drops a ubuf
reference. In theory, this could race with device removal which waits
on the ubuf kref, and crash on use after free.
Do all accesses within rcu read side critical section, and synchronize
on release.
Since callbacks are always invoked from bh, synchronize_rcu_bh seems
enough and will help release complete a bit faster.
Signed-off-by: Michael S. Tsirkin <[email protected]>
Acked-by: Jason Wang <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
|
|
vhost checked the counter within the refcnt before decrementing. It
really wanted to know that it is the one that has the last reference, as
a way to batch freeing resources a bit more efficiently.
Note: we only let refcount go to 0 on device release.
This works well but we now access the ref counter twice so there's a
race: all users might see a high count and decide to defer freeing
resources.
In the end no one initiates freeing resources until the last reference
is gone (which is on VM shotdown so might happen after a looooong time).
Let's do what we probably should have done straight away:
switch from kref to plain atomic, documenting the
semantics, return the refcount value atomically after decrement,
then use that to avoid the deadlock.
Reported-by: Qin Chuanyu <[email protected]>
Signed-off-by: Michael S. Tsirkin <[email protected]>
Acked-by: Jason Wang <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
|
|
Signed-off-by: Liu Junliang <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
|
|
Use %zu for size_t in order to avoid the following build
warning in printks.
drivers/net/usb/sr9800.c: In function 'sr9800_bind'
drivers/net/usb/sr9800.c:826:2: warning: format '%ld' expects argument of type 'long int' but argument 5 has type 'size_t'
[-Wformat]
Signed-off-by: Jingoo Han <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
|
|
Fix incorrect comment reported by Norbert Kiesel. Edit another comment to add
more details. Also add references to algorithm (IETF draft and paper) to top of
file.
Signed-off-by: Vijay Subramanian <[email protected]>
CC: Mythili Prabhu <[email protected]>
CC: Norbert Kiesel <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
|
|
It's wrong if the device tree doesn't provide a phy-mode property for
the cpsw slaves as it is documented to be required in
Documentation/devicetree/bindings/net/cpsw.txt.
Anyhow it's nice to catch that problem, still more as it used to work
without this property up to commit 388367a5a9fb (drivers: net: cpsw: use
cpsw-phy-sel driver to configure phy mode) which is in v3.13-rc1.
Signed-off-by: Uwe Kleine-König <[email protected]>
Acked-by: Mugunthan V N <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
|
|
When the PXS8 and PHS8 devices show up with PID 0x0053 they will expose both a
QMI port and a WWAN interface.
CC: Hans-Christoph Schemmel <[email protected]>
CC: Christian Schmiedl <[email protected]>
CC: Nicolaus Colberg <[email protected]>
CC: David McCullough <[email protected]>
Signed-off-by: Aleksander Morgado <[email protected]>
Acked-by: Bjørn Mork <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
|
|
git://gitorious.org/linux-can/linux-can
linux-can-fixes-for-3.14-20140212
Marc Kleine-Budde says:
====================
this is a pull request with one patch for net/master, for the current release
cycle. Olivier Sobrie noticed and fixed that the kvaser_usb driver doesn't
check the number of channels value from the hardware, which may result in
writing over the bounds of an array in the driver.
====================
Signed-off-by: David S. Miller <[email protected]>
|
|
Warning log:
xilinx_axienet_main.c: In function 'axienet_start_xmit_done':
xilinx_axienet_main.c:617:16: warning: operation on 'lp->tx_bd_ci' may be undefined [-Wsequence-point]
xilinx_axienet_main.c: In function 'axienet_start_xmit':
xilinx_axienet_main.c:703:18: warning: operation on 'lp->tx_bd_tail' may be undefined [-Wsequence-point]
xilinx_axienet_main.c:719:17: warning: operation on 'lp->tx_bd_tail' may be undefined [-Wsequence-point]
xilinx_axienet_main.c: In function 'axienet_recv':
xilinx_axienet_main.c:792:16: warning: operation on 'lp->rx_bd_ci' may be undefined [-Wsequence-point]
xilinx_axienet_main.c: In function 'axienet_of_probe':
xilinx_axienet_main.c:1501:21: warning: unused variable 'rc' [-Wunused-variable]
Signed-off-by: Michal Simek <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
|
|
Add missing header to fix compilation error.
drivers/net/ethernet/xilinx/xilinx_axienet_main.c:1575:22:
error: undefined identifier 'irq_of_parse_and_map'
drivers/net/ethernet/xilinx/xilinx_axienet_main.c:1576:22:
error: undefined identifier 'irq_of_parse_and_map'
Signed-off-by: Michal Simek <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
|
|
Research has shown that commit a77fcf895046 ("IB/qib: Use a single
txselect module parameter for serdes tuning") missed a key serdes init
sequence.
This patch add that sequence.
Cc: <[email protected]>
Reviewed-by: Dennis Dalessandro <[email protected]>
Signed-off-by: Mike Marciniszyn <[email protected]>
Signed-off-by: Roland Dreier <[email protected]>
|
|
Signed-off-by: Kumar Sanghvi <[email protected]>
Signed-off-by: Hariprasad Shenai <[email protected]>
Signed-off-by: Roland Dreier <[email protected]>
|
|
For userspace RoCE UD QPs we need to know the GID format that the
kernel uses, e.g when working over older kernels. For that end, add a
new port capability IB_PORT_IP_BASED_GIDS and report it when query
port is issued.
Signed-off-by: Moni Shoua <[email protected]>
Signed-off-by: Matan Barak <[email protected]>
Signed-off-by: Or Gerlitz <[email protected]>
Signed-off-by: Roland Dreier <[email protected]>
|
|
When scanning netdevices we need to check a few more conditions and
cases to build the IBoE GID table properly. For example, under
bonding we must make sure that when a port is down, the bond IP
address isn't programmed as a GID, since doing so will cause failure
with IB core flows that selects ports by GID.
Signed-off-by: Moni Shoua <[email protected]>
Signed-off-by: Or Gerlitz <[email protected]>
Signed-off-by: Roland Dreier <[email protected]>
|
|
The IBoE code used to reset the GID table did it for all Ethernet
ports of the device. Since the whole architecture of generating GIDs
and responding to events is port-based, this is inefficient and can
lead to wrong content in the GID table. Change the reset flow to be
per-port.
Signed-off-by: Moni Shoua <[email protected]>
Signed-off-by: Or Gerlitz <[email protected]>
Signed-off-by: Roland Dreier <[email protected]>
|
|
Updating the GID table under IBoE requires read/write from/to shared
data structures. These data structures are protected with the device
iboe lock. The flows that modify the GID table start from
1. Initializing the GID table
2. NETDEV events
3. INET or INET6 events
This patch makes sure that the flow of initializing the GID table is
consistent with the other two flows w.r.t on what step the lock is taken.
Signed-off-by: Moni Shoua <[email protected]>
Signed-off-by: Or Gerlitz <[email protected]>
Signed-off-by: Roland Dreier <[email protected]>
|
|
On the one hand, the invocation of netdev_master_upper_dev_get()
within mlx4_ib_scan_netdevs() must be done with rtnl lock held. On
the other hand, it's wrong to call rtnl_lock() from within this
function since it's also called by our netdev notifier callback.
Therefore move the locking to mlx4_ib_add() so that both cases are
covered.
Signed-off-by: Moni Shoua <[email protected]>
Signed-off-by: Or Gerlitz <[email protected]>
Signed-off-by: Roland Dreier <[email protected]>
|
|
Make sure that for Ethernet ports, the port GID table index 0 is always
occupied with a default GID of the relevant IPv6 link-local adderss.
This provides better user experience for legacy applications that don't use
the RDMA CM and were working on index 0 prior to the IP addressing change.
Also, as GIDs are generated from IP addresses of the network devices that
are associated with the port, it's basically possible that the GID table
will be empty if no IP address was assigned. This doesn't comply with the
IB spec section 4.1.1 "GID usage and properties".
Signed-off-by: Moni Shoua <[email protected]>
Signed-off-by: Or Gerlitz <[email protected]>
Signed-off-by: Roland Dreier <[email protected]>
|
|
Marcelo Ricardo Leitner reported problems when the forwarding link path
has a lower mtu than the incoming one if the inbound interface supports GRO.
Given:
Host <mtu1500> R1 <mtu1200> R2
Host sends tcp stream which is routed via R1 and R2. R1 performs GRO.
In this case, the kernel will fail to send ICMP fragmentation needed
messages (or pkt too big for ipv6), as GSO packets currently bypass dstmtu
checks in forward path. Instead, Linux tries to send out packets exceeding
the mtu.
When locking route MTU on Host (i.e., no ipv4 DF bit set), R1 does
not fragment the packets when forwarding, and again tries to send out
packets exceeding R1-R2 link mtu.
This alters the forwarding dstmtu checks to take the individual gso
segment lengths into account.
For ipv6, we send out pkt too big error for gso if the individual
segments are too big.
For ipv4, we either send icmp fragmentation needed, or, if the DF bit
is not set, perform software segmentation and let the output path
create fragments when the packet is leaving the machine.
It is not 100% correct as the error message will contain the headers of
the GRO skb instead of the original/segmented one, but it seems to
work fine in my (limited) tests.
Eric Dumazet suggested to simply shrink mss via ->gso_size to avoid
sofware segmentation.
However it turns out that skb_segment() assumes skb nr_frags is related
to mss size so we would BUG there. I don't want to mess with it considering
Herbert and Eric disagree on what the correct behavior should be.
Hannes Frederic Sowa notes that when we would shrink gso_size
skb_segment would then also need to deal with the case where
SKB_MAX_FRAGS would be exceeded.
This uses sofware segmentation in the forward path when we hit ipv4
non-DF packets and the outgoing link mtu is too small. Its not perfect,
but given the lack of bug reports wrt. GRO fwd being broken this is a
rare case anyway. Also its not like this could not be improved later
once the dust settles.
Acked-by: Herbert Xu <[email protected]>
Reported-by: Marcelo Ricardo Leitner <[email protected]>
Signed-off-by: Florian Westphal <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
|