devlink_nl_cmd_region_read_dumpit() misses to return right error code on
most error conditions.
Return the right error code on such errors.
Fixes: 4e54795a27 ("devlink: Add support for region snapshot read command")
Signed-off-by: Parav Pandit <parav@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
This patch is a little improvement. If user use the
command shown as below, we should print the info [1]
instead of [2]. The eth0 exists actually, and it may
confuse user.
$ echo "eth0" > /sys/class/net/bond4/bonding/slaves
[1] "bond4: no command found in slaves file - use +ifname or -ifname"
[2] "write error: No such device"
Signed-off-by: Tonghao Zhang <xiangxia.m.yue@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Ido Schimmel says:
====================
mlxsw: hwmon and thermal extensions
Vadim says:
This patchset contains various improvements to hwmon and thermal code in
mlxsw. The most significant improvement is the ability to read modules'
temperature attributes (input, fault, critical and emergency thresholds)
as well as fans' fault indication. These new attributes will improve the
ability to monitor the system.
Patches #1-#4 add the necessary device registers and APIs to read
modules' temperature attributes and fans' fault indication.
Patches #5-#8 perform small improvements in hwmon and thermal code such
as using a more indicative name for cooling devices.
Patch #9 exposes fans' fault indication via hwmon.
Patch #10 exposes modules' temperature attributes via hwmon.
Patch #11 adds an hwmon label to modules' temperature sensor. This helps
to parse the output of utilities such as "sensors".
Patch #12 allows to bind an external cooling device ("mlxreg-fan") to
mlxsw thermal zone. This will allow the mlxsw thermal zone to change the
cooling level of cooling devices not programmed via switch registers.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
Allow thermal zone binding to an external cooling device from the
cooling devices white list.
It provides support for Mellanox next generation systems on which
cooling device logic is not controlled through the switch registers.
Signed-off-by: Vadim Pasternak <vadimp@mellanox.com>
Reviewed-by: Jiri Pirko <jiri@mellanox.com>
Signed-off-by: Ido Schimmel <idosch@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Add label attribute to hwmon object for exposing QSFP module's
temperature sensor name. Modules are labeled as "front panel xxx". The
label is used by utilities such as "sensors":
front panel 001: +0.0C (crit = +0.0C, emerg = +0.0C)
..
front panel 020: +31.0C (crit = +70.0C, emerg = +80.0C)
..
front panel 056: +41.0C (crit = +70.0C, emerg = +80.0C)
Signed-off-by: Vadim Pasternak <vadimp@mellanox.com>
Reviewed-by: Jiri Pirko <jiri@mellanox.com>
Signed-off-by: Ido Schimmel <idosch@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Add new attributes to hwmon object for exposing QSFP module temperature
input, fault indication, critical and emergency thresholds. Temperature
input and fault indication are read from Management Temperature Bulk
Register. Temperature thresholds are read from Management Cable Info
Access Register.
Signed-off-by: Vadim Pasternak <vadimp@mellanox.com>
Reviewed-by: Jiri Pirko <jiri@mellanox.com>
Signed-off-by: Ido Schimmel <idosch@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Add new fan hwmon attribute for exposing fan faults (fault indication is
read from Fan Out of Range Event Register).
Signed-off-by: Vadim Pasternak <vadimp@mellanox.com>
Reviewed-by: Jiri Pirko <jiri@mellanox.com>
Signed-off-by: Ido Schimmel <idosch@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Rename cooling device from "Fan" to "mlxsw_fan". Name "Fan" is too
common name, and such name is misleading, while it's interpreted by
user. For example name "Fan" could be used by ACPI.
Signed-off-by: Vadim Pasternak <vadimp@mellanox.com>
Reviewed-by: Jiri Pirko <jiri@mellanox.com>
Signed-off-by: Ido Schimmel <idosch@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Replace thermal hardcoded temperature trip values with defines.
Signed-off-by: Vadim Pasternak <vadimp@mellanox.com>
Reviewed-by: Jiri Pirko <jiri@mellanox.com>
Signed-off-by: Ido Schimmel <idosch@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Modify thermal zone trip points setting for better alignment with system
thermal requirement.
Add hysteresis thresholds for thermal trips in order to avoid throttling
around thermal trip point. If hysteresis temperature is not considered,
PWM can have side effect of flip up/down on thermal trip point boundary.
Signed-off-by: Vadim Pasternak <vadimp@mellanox.com>
Reviewed-by: Jiri Pirko <jiri@mellanox.com>
Signed-off-by: Ido Schimmel <idosch@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Add low frequency bus capability in order to allow core functionality
separation based on bus type. Driver could run over PCIe, which is
considered as high frequency bus or I2C, which is considered as low
frequency bus. In the last case time setting, for example, for thermal
polling interval, should be increased.
Use different thermal monitoring based on bus type. For I2C bus time is
set to 20 seconds, while for PCIe 1 second polling interval is used.
Signed-off-by: Vadim Pasternak <vadimp@mellanox.com>
Reviewed-by: Jiri Pirko <jiri@mellanox.com>
Signed-off-by: Ido Schimmel <idosch@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Add new API to read QSFP module's temperature thresholds - warning and
critical.
New internal API reads the temperature thresholds from the modules,
which are equipped with the thermal sensor. These thresholds will be
exposed via hwmon subsystem.
Signed-off-by: Vadim Pasternak <vadimp@mellanox.com>
Reviewed-by: Jiri Pirko <jiri@mellanox.com>
Signed-off-by: Ido Schimmel <idosch@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Add FORE (Fan Out of Range Event Register), which is used for fan fault
reading.
Signed-off-by: Vadim Pasternak <vadimp@mellanox.com>
Reviewed-by: Jiri Pirko <jiri@mellanox.com>
Signed-off-by: Ido Schimmel <idosch@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Add MTBR (Management Temperature Bulk Register), which is used for port
temperature reading in a bulk mode.
Signed-off-by: Vadim Pasternak <vadimp@mellanox.com>
Reviewed-by: Jiri Pirko <jiri@mellanox.com>
Signed-off-by: Ido Schimmel <idosch@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Move QSFP EEPROM definitions to common location from the spectrum driver
in order to make them available for other mlxsw modules. They are common
for all kind of chips and have relation to SFF specifications 8024,
8436, 8472, 8636, rather than to chip type.
Signed-off-by: Vadim Pasternak <vadimp@mellanox.com>
Reviewed-by: Jiri Pirko <jiri@mellanox.com>
Signed-off-by: Ido Schimmel <idosch@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
- fix memory leak in in batadv_dat_put_dhcp, by Martin Weinelt
- fix typo, by Sven Eckelmann
- netlink restructuring patch series (part 2), by Sven Eckelmann
(19 patches)
-----BEGIN PGP SIGNATURE-----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=+p14
-----END PGP SIGNATURE-----
Merge tag 'batadv-next-for-davem-20190213' of git://git.open-mesh.org/linux-merge
Simon Wunderlich says:
====================
This feature/cleanup patchset includes the following patches:
- fix memory leak in in batadv_dat_put_dhcp, by Martin Weinelt
- fix typo, by Sven Eckelmann
- netlink restructuring patch series (part 2), by Sven Eckelmann
(19 patches)
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
We need to set the error message on this path otherwise some of the
callers, such as test_hints_case(), print from an uninitialized pointer.
We had a similar bug earlier and set "errmsg" to NULL in the caller,
test_delta_action_item(). That code is no longer required so I have
removed it.
Fixes: 9069a3817d ("lib: objagg: implement optimization hints assembly and use hints for object creation")
Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Acked-by: Jiri Pirko <jiri@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
There is a typo here. We intended to check "objagg2" but we instead
test "objagg" which is not an error pointer.
Fixes: 9069a3817d ("lib: objagg: implement optimization hints assembly and use hints for object creation")
Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Acked-by: Jiri Pirko <jiri@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
We need to set the error code on this path otherwise we return
ERR_PTR(0) which would result in a NULL dereference in the caller.
Fixes: 9069a3817d ("lib: objagg: implement optimization hints assembly and use hints for object creation")
Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Acked-by: Jiri Pirko <jiri@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
CR4_QSFP 10G Speed technology should be 10000baseKR_Full
And also report available FEC modes.
Signed-off-by: Vishal Kulkarni <vishal@chelsio.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Jesper Dangaard Brouer says:
====================
Fix page_pool API and dma address storage
As pointed out by David Miller in [1] the current page_pool implementation
stores dma_addr_t in page->private. This won't work on 32-bit platforms with
64-bit DMA addresses since the page->private is an unsigned long and the
dma_addr_t a u64.
Since no driver is yet using the DMA mapping capabilities of the API let's
fix this by storing the information in 'struct page' and use that to store
and retrieve DMA addresses from network drivers.
As long as the addresses returned from dma_map_page() are aligned the first
bit, used by the compound pages code should not be set.
Ilias tested the first two patches on Espressobin driver mvneta, for which
we have patches for using the DMA API of page_pool.
[1]: https://lore.kernel.org/netdev/20181207.230655.1261252486319967024.davem@davemloft.net/
====================
Signed-off-by: Jesper Dangaard Brouer <brouer@redhat.com>
Signed-off-by: Ilias Apalodimas <ilias.apalodimas@linaro.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
As pointed out by Alexander Duyck, the DMA mapping done in page_pool needs
to use the DMA attribute DMA_ATTR_SKIP_CPU_SYNC.
As the principle behind page_pool keeping the pages mapped is that the
driver takes over the DMA-sync steps.
Reported-by: Alexander Duyck <alexander.duyck@gmail.com>
Signed-off-by: Jesper Dangaard Brouer <brouer@redhat.com>
Signed-off-by: Ilias Apalodimas <ilias.apalodimas@linaro.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
As pointed out by David Miller the current page_pool implementation
stores dma_addr_t in page->private.
This won't work on 32-bit platforms with 64-bit DMA addresses since the
page->private is an unsigned long and the dma_addr_t a u64.
A previous patch is adding dma_addr_t on struct page to accommodate this.
This patch adapts the page_pool related functions to use the newly added
struct for storing and retrieving DMA addresses from network drivers.
Signed-off-by: Ilias Apalodimas <ilias.apalodimas@linaro.org>
Signed-off-by: Jesper Dangaard Brouer <brouer@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
The page_pool API is using page->private to store DMA addresses.
As pointed out by David Miller we can't use that on 32-bit architectures
with 64-bit DMA
This patch adds a new dma_addr_t struct to allow storing DMA addresses
Signed-off-by: Jesper Dangaard Brouer <brouer@redhat.com>
Signed-off-by: Ilias Apalodimas <ilias.apalodimas@linaro.org>
Acked-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
With the introduction of flow_stats_update(), drivers now update the stats
fields of the passed tc_cls_flower_offload struct, rather than call
tcf_exts_stats_update() directly to update the stats of offloaded TC
flower rules. However, if multiple qdiscs are registered to a TC shared
block and a flower rule is applied, then, when getting stats for the rule,
multiple callbacks may be made.
Take this into consideration by modifying flow_stats_update to gather the
stats from all callbacks. Currently, the values in tc_cls_flower_offload
only account for the last stats callback in the list.
Fixes: 3b1903ef97 ("flow_offload: add statistics retrieval infrastructure and use it")
Signed-off-by: John Hurley <john.hurley@netronome.com>
Reviewed-by: Jakub Kicinski <jakub.kicinski@netronome.com>
Acked-by: Pablo Neira Ayuso <pablo@netfilter.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
Recently introduced tc_setup_flow_action() can fail when parsing tcf_exts
on some unsupported action commands. However, this should not affect the
case when user did not explicitly request hw offload by setting skip_sw
flag. Modify tc_setup_flow_action() callers to only propagate the error if
skip_sw flag is set for filter that is being offloaded, and set extack
error message in that case.
Signed-off-by: Vlad Buslov <vladbu@mellanox.com>
Fixes: 3a7b68617d ("cls_api: add translator to flow_action representation")
Signed-off-by: David S. Miller <davem@davemloft.net>
There are some statements that are indented incorrectly. Fix these.
Signed-off-by: Colin Ian King <colin.king@canonical.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
There are some statements in an if-block that are not correctly
indented. Fix these.
Signed-off-by: Colin Ian King <colin.king@canonical.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
dev_consume_skb_irq() should be called in eth_txdone_irq() when skb
xmit done. It makes drop profiles(dropwatch, perf) more friendly.
Signed-off-by: Yang Wei <yang.wei9@zte.com.cn>
Signed-off-by: David S. Miller <davem@davemloft.net>
dev_consume_skb_irq() should be called in at91ether_interrupt() when
skb xmit done. It makes drop profiles(dropwatch, perf) more friendly.
Signed-off-by: Yang Wei <yang.wei9@zte.com.cn>
Reviewed-by: Claudiu Beznea <claudiu.beznea@microchip.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
dev_consume_skb_irq() should be called when skb xmit done. It makes
drop profiles(dropwatch, perf) more friendly.
Signed-off-by: Yang Wei <yang.wei9@zte.com.cn>
Signed-off-by: David S. Miller <davem@davemloft.net>
dev_consume_skb_irq() should be called in intr_handler() when skb
xmit done. It makes drop profiles(dropwatch, perf) more friendly.
Signed-off-by: Yang Wei <yang.wei9@zte.com.cn>
Signed-off-by: David S. Miller <davem@davemloft.net>
dev_consume_skb_irq() should be called in moxart_tx_finished() when
skb xmit done. It makes drop profiles(dropwatch, perf) more friendly.
Signed-off-by: Yang Wei <yang.wei9@zte.com.cn>
Signed-off-by: David S. Miller <davem@davemloft.net>
dev_consume_skb_irq() should be called in mace_interrupt() when skb
xmit done. It makes drop profiles(dropwatch, perf) more friendly.
Signed-off-by: Yang Wei <yang.wei9@zte.com.cn>
Signed-off-by: David S. Miller <davem@davemloft.net>
dev_consume_skb_irq() should be called when skb xmit done. It makes
drop profiles(dropwatch, perf) more friendly.
Signed-off-by: Yang Wei <yang.wei9@zte.com.cn>
Signed-off-by: David S. Miller <davem@davemloft.net>
dev_consume_skb_irq() should be called in emac_mac_tx_process() when
skb xmit done. It makes drop profiles(dropwatch, perf) more friendly.
Signed-off-by: Yang Wei <yang.wei9@zte.com.cn>
Signed-off-by: David S. Miller <davem@davemloft.net>
dev_consume_skb_irq() should be called when skb xmit done. It makes
drop profiles(dropwatch, perf) more friendly.
Signed-off-by: Yang Wei <yang.wei9@zte.com.cn>
Signed-off-by: David S. Miller <davem@davemloft.net>
Maxime Chevallier says:
====================
net: phy: Add 2.5G/5GBASET PHYs support
The 802.3bz standard defines 2 modes based on the NBASET alliance work
that allow to use 2.5Gbps and 5Gbps speeds on Cat 5e, 6 and 7 cables.
This series adds the necessary infrastructure to handle these modes with
C45 PHYs. This series was originally part of a bigger one, that has
seen 2 iterations [1] [2] that added support for these modes on Marvell
Alaska PHYs.
Following some discussions with Heiner and Andrew [3], we decided to
split-out the generic parts so that we can work together on the
following steps to get these mode fully working with Aquantia and
Marvell PHYS.
The first 3 patches are reworking some of the internal network phy
infrastructure to handle the new modes in a more generic way.
The 4th patch adds all the C45 register definition and accesses that
follows the 802.3bz standard to support 2.5GBASET and 5GBASET.
[1] : https://lore.kernel.org/netdev/20190118152352.26417-1-maxime.chevallier@bootlin.com/
[2] : https://lore.kernel.org/netdev/20190207094939.27369-1-maxime.chevallier@bootlin.com/
[3] : https://lore.kernel.org/netdev/81c340ea-54b0-1abf-94af-b8dc4ee83e3a@gmail.com/
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
The 802.3bz specification, based on previous by the NBASET alliance,
defines the 2.5GBaseT and 5GBaseT link modes for ethernet traffic on
cat5e, cat6 and cat7 cables.
These mode integrate with the already defined C45 MDIO PMA/PMD registers
set that added 10G support, by defining some previously reserved bits,
and adding a new register (2.5G/5G Extended abilities).
This commit adds the required definitions in include/uapi/linux/mdio.h
to support these modes, and detect when a link-partner advertises them.
It also adds support for these mode in the generic C45 PHY
infrastructure.
Signed-off-by: Maxime Chevallier <maxime.chevallier@bootlin.com>
Reviewed-by: Heiner Kallweit <hkallweit1@gmail.com>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: David S. Miller <davem@davemloft.net>
Marvell 10G PHY driver has a generic way of initializing the supported
link modes by reading the PHY's C45 PMA abilities. This can be made
generic, since these registers are part of the 802.3 specifications.
This commit extracts the config_init link_mode initialization code from
marvell10g and uses it to introduce the genphy_c45_pma_read_abilities
function.
Only PMA modes are read, it's still up to the caller to set the Pause
parameters.
Signed-off-by: Maxime Chevallier <maxime.chevallier@bootlin.com>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: David S. Miller <davem@davemloft.net>
Since of_set_phy_supported was moved to phy-core.c, we can also move
of_set_phy_eee_broken to the same location, so that we have all OF
functions in the same place.
This patch doesn't intend to introduce any change in behaviour.
Signed-off-by: Maxime Chevallier <maxime.chevallier@bootlin.com>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: David S. Miller <davem@davemloft.net>
When setting a PHY's max speed using either the max-speed DT property
or ethtool, we should mask-out all non-compatible modes according to the
settings table, instead of just the 10/100BASET modes.
Signed-off-by: Maxime Chevallier <maxime.chevallier@bootlin.com>
Suggested-by: Russell King <rmk+kernel@armlinux.org.uk>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: David S. Miller <davem@davemloft.net>
Florian Fainelli says:
====================
Remove unused variables
This removes unused variables from mlxsw and ethsw after the recent
removal of SWITCHDEV_ATTR_ID_PORT_BRIDGE_FLAGS, build scripts are now
fixed to take care of those warnings :).
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
After 1b8b589d91 ("staging: fsl-dpaa2: ethsw: Remove getting
PORT_BRIDGE_FLAGS") we are not accessing any driver private data
anymore, remove port_priv which is unused.
Fixes: 1b8b589d91 ("staging: fsl-dpaa2: ethsw: Remove getting PORT_BRIDGE_FLAGS")
Signed-off-by: Florian Fainelli <f.fainelli@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
After 1ecb195753 ("mlxsw: spectrum_switchdev: Remove getting
PORT_BRIDGE_FLAGS") we are not accessing any driver private data
structure, so the mlxsw_sp_port and mlxsw_sp variables are unused.
Fixes: 1ecb195753 ("mlxsw: spectrum_switchdev: Remove getting PORT_BRIDGE_FLAGS")
Signed-off-by: Florian Fainelli <f.fainelli@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
After 610d2b601b ("rocker: Remove getting PORT_BRIDGE_FLAGS") we no
longer have a port_attr_bridge_flags_get member in the rocker_world_ops
structre, fix that.
Fixes: 610d2b601b ("rocker: Remove getting PORT_BRIDGE_FLAGS")
Signed-off-by: Florian Fainelli <f.fainelli@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Vlad Buslov says:
====================
Refactor classifier API to work with chain/classifiers without rtnl lock
Currently, all netlink protocol handlers for updating rules, actions and
qdiscs are protected with single global rtnl lock which removes any
possibility for parallelism. This patch set is a third step to remove
rtnl lock dependency from TC rules update path.
Recently, new rtnl registration flag RTNL_FLAG_DOIT_UNLOCKED was added.
Handlers registered with this flag are called without RTNL taken. End
goal is to have rule update handlers(RTM_NEWTFILTER, RTM_DELTFILTER,
etc.) to be registered with UNLOCKED flag to allow parallel execution.
However, there is no intention to completely remove or split rtnl lock
itself. This patch set addresses specific problems in implementation of
classifiers API that prevent its control path from being executed
concurrently, and completes refactoring of cls API rules update handlers
by removing rtnl lock dependency from code that handles chains and
classifiers. Rules update handlers are registered with
RTNL_FLAG_DOIT_UNLOCKED flag.
This patch set substitutes global rtnl lock dependency on rules update
path in cls API by extending its data structures with following locks:
- tcf_block with 'lock' mutex. It is used to protect block state and
life-time management fields of chains on the block (chain->refcnt,
chain->action_refcnt, chain->explicitly crated, etc.).
- tcf_chain with 'filter_chain_lock' mutex, that is used to protect list
of classifier instances attached to chain. chain0->filter_chain_lock
serializes calls to head change callbacks and allows them to rely on
filter_chain_lock for serialization instead of rtnl lock.
- tcf_proto with 'lock' spinlock that is intended to be used to
synchronize access to classifiers that support unlocked execution.
Classifiers are extended with reference counting to accommodate parallel
access by unlocked cls API. Classifier ops structure is extended with
additional 'put' function to allow reference counting of filters and
intended to be used by classifiers that implement rtnl-unlocked API.
Users of classifiers and individual filter instances are modified to
always hold reference while working with them.
Classifiers that support unlocked execution still need to know the
status of rtnl lock, so their API is extended with additional
'rtnl_held' argument that is used to indicate that caller holds rtnl
lock. Cls API propagates rtnl lock status across its helper functions
and passes it to classifier.
Changes from V3 to V4:
- Patch 1:
- Extract code that manages chain 'explicitly_created' flag into
standalone patch.
- Patch 2 - new.
Changes from V2 to V3:
- Change block->lock and chain->filter_chain_lock type to mutex. This
removes the need for async miniqp refactoring and allows calling
sleeping functions while holding the block->lock and
chain->filter_chain_lock locks.
- Previous patch 1 - async miniqp is no longer needed, remove the patch.
- Patch 1:
- Change block->lock type to mutex.
- Implement tcf_block_destroy() helper function that destroys
block->lock mutex before deallocating the block.
- Revert GFP_KERNEL->GFP_ATOMIC memory allocation flags of tcf_chain
which is no longer needed after block->lock type change.
- Patch 6:
- Change chain->filter_chain_lock type to mutex.
- Assume chain0->filter_chain_lock synchronizations instead of rtnl
lock in mini_qdisc_pair_swap() function that is called from head
change callback of ingress Qdisc. With filter_chain_lock type
changed to mutex it is now possible to call sleeping function while
holding it, so it is now used instead of async implementation from
previous versions of this patch set.
- Patch 7:
- Add local tp_next var to tcf_chain_flush() and use it to store
tp->next pointer dereferenced with rcu_dereference_protected() to
satisfy kbuild test robot.
- Reset tp pointer to NULL at the beginning of tc_new_tfilter() to
prevent its uninitialized usage in error handling code. This code
was already implemented in patch 10, but must be in patch 8 to
preserve code bisectability.
- Put parent chain in tcf_proto_destroy(). In previous version this
code was implemented in patch 1 which was removed in V3.
Changes from V1 to V2:
- Patch 1:
- Use smp_store_release() instead of xchg() for setting
miniqp->tp_head.
- Move Qdisc deallocation to tc_proto_wq ordered workqueue that is
used to destroy tcf proto instances. This is necessary to ensure
that Qdisc is destroyed after all instances of chain/proto that it
contains in order to prevent use-after-free error in
tc_chain_notify_delete().
- Cache parent net device ifindex in block to prevent use-after-free
of dev queue in tc_chain_notify_delete().
- Patch 2:
- Use lockdep_assert_held() instead of spin_is_locked() for assertion.
- Use 'free_block' argument in tcf_chain_destroy() instead of checking
block's reference count and chain_list for second time.
- Patch 7:
- Refactor tcf_chain0_head_change_cb_add() to not take block->lock and
chain0->filter_chain_lock in correct order.
- Patch 10:
- Always set 'tp_created' flag when creating tp to prevent releasing
the chain twice when tp with same priority was inserted
concurrently.
- Patch 11:
- Add additional check to prevent creation of new proto instance when
parent chain is being flushed to reduce CPU usage.
- Don't call tcf_chain_delete_empty() if tp insertion failed.
- Patch 16 - new.
- Patch 17:
- Refactor to only lock take rtnl lock once (at the beginning of rule
update handlers).
- Always release rtnl mutex in the same function that locked it.
Remove unlock code from tcf_block_release().
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
Register netlink protocol handlers for message types RTM_NEWTFILTER,
RTM_DELTFILTER, RTM_GETTFILTER as unlocked. Set rtnl_held variable that
tracks rtnl mutex state to be false by default.
Introduce tcf_proto_is_unlocked() helper that is used to check
tcf_proto_ops->flag to determine if ops can be called without taking rtnl
lock. Manually lookup Qdisc, class and block in rule update handlers.
Verify that both Qdisc ops and proto ops are unlocked before using any of
their callbacks, and obtain rtnl lock otherwise.
Signed-off-by: Vlad Buslov <vladbu@mellanox.com>
Acked-by: Jiri Pirko <jiri@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Refactor tcf_block_find() code into three standalone functions:
- __tcf_qdisc_find() to lookup Qdisc and increment its reference counter.
- __tcf_qdisc_cl_find() to lookup class.
- __tcf_block_find() to lookup block and increment its reference counter.
This change is necessary to allow netlink tc rule update handlers to call
these functions directly in order to conditionally take rtnl lock
according to Qdisc class ops flags before calling any of class ops
functions.
Signed-off-by: Vlad Buslov <vladbu@mellanox.com>
Acked-by: Jiri Pirko <jiri@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
