Sunil Goutham [Mon, 2 Mar 2020 09:59:00 +0000 (15:29 +0530)]
net: thunderx: Adjust CQE_RX drop levels for better performance
With the current RX RED/DROP levels of 192/184 for CQE_RX, when
packet incoming rate is high, LLC is getting polluted resulting
in more cache misses and higher latency in packet processing. This
slows down the whole process and performance loss. Hence reduced
the levels to 224/216 (ie for a CQ size of 1024, Rx pkts will be
red dropped or dropped when unused CQE are less than 128/160 respectively)
Signed-off-by: Sunil Goutham <sgoutham@marvell.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
David S. Miller [Mon, 2 Mar 2020 19:08:52 +0000 (11:08 -0800)]
Merge branch 'octeontx2-Flow-control-support-and-other-misc-changes'
Sunil Goutham says:
====================
octeontx2: Flow control support and other misc changes
This patch series adds flow control support (802.3 pause frames) and
has other changes wrt generic admin function (AF) driver functionality.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
Sunil Goutham [Mon, 2 Mar 2020 07:19:28 +0000 (12:49 +0530)]
octeontx2-af: Modify rvu_reg_poll() to check reg atleast twice
Currently on the first check if the operation is still not
finished, the poll goes to sleep for 2-5 usecs. But if for
some reason (due to other priority stuff like interrupts etc) by
the time the poll wakes up the 10ms time is expired then we don't
check if operation is finished or not and return failure.
This patch modifies poll logic to check HW operation after sleep so
that the status is checked atleast twice.
Signed-off-by: Sunil Goutham <sgoutham@marvell.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Sunil Goutham [Mon, 2 Mar 2020 07:19:27 +0000 (12:49 +0530)]
octeontx2-af: Enable PCI master
Bus mastering is enabled by firmware, but when this driver
is unbinded bus mastering gets disabled by the PCI subsystem
which results interrupts not working when driver is reloaded.
Hence set bus mastering everytime in probe().
Also
- Converted pci_set_dma_mask() and pci_set_consistent_dma_mask()
to dma_set_mask_and_coherent().
- Cleared transaction pending bit which gets set during
driver unbind due to clearing of bus mastering (ME bit).
Signed-off-by: Sunil Goutham <sgoutham@marvell.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Sunil Goutham [Mon, 2 Mar 2020 07:19:26 +0000 (12:49 +0530)]
octeontx2-af: Set discovery ID for RVUM block
Currently there is no way for AF dependent drivers in
any domain to check if the AF driver is loaded. This
patch sets an ID for RVUM block which will automatically
reflects in PF/VFs discovery register which they can
check and defer their probe until AF is up.
Signed-off-by: Sunil Goutham <sgoutham@marvell.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Linu Cherian [Mon, 2 Mar 2020 07:19:25 +0000 (12:49 +0530)]
octeontx2-af: Optimize data retrieval from firmware
For retrieving info like interface MAC addresses, packet
parser key extraction config etc currently a command
is sent to firmware and firmware which periodically polls
for commands, processes these and returns the info.
This is resulting in interface initialization taking lot
of time. To optimize this a memory region is shared between
firmware and this driver, firmware while booting puts
static info like these into that region for driver to
read directly without using commands.
With this
- Logic for retrieving packet parser extraction config
via commands is removed and repalced with using the
shared 'fwdata' structure.
- Now RVU MSIX vector address is also retrieved from this fwdata struct
instead of from CSR. Otherwise when kexec/kdump crash kernel loads
CSR will have a IOVA setup by primary kernel which impacts
RVU PF/VF's interrupts.
- Also added a mbox handler for PF/VF interfaces to retrieve their MAC
addresses from AF.
Signed-off-by: Linu Cherian <lcherian@marvell.com>
Signed-off-by: Christina Jacob <cjacob@marvell.com>
Signed-off-by: Rakesh Babu <rsaladi2@marvell.com>
Signed-off-by: Sunil Goutham <sgoutham@marvell.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Geetha sowjanya [Mon, 2 Mar 2020 07:19:24 +0000 (12:49 +0530)]
octeontx2-pf: Support to enable/disable pause frames via ethtool
Added mailbox requests to retrieve backpressure IDs from AF and Aura,
CQ contexts are configured with these BPIDs. So that when resource
levels reach configured thresholds they assert backpressure on the
interface which is also mapped to same BPID.
Also added support to enable/disable pause frames generation via ethtool.
Signed-off-by: Geetha sowjanya <gakula@marvell.com>
Signed-off-by: Sunil Goutham <sgoutham@marvell.com>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: David S. Miller <davem@davemloft.net>
Geetha sowjanya [Mon, 2 Mar 2020 07:19:23 +0000 (12:49 +0530)]
octeontx2-af: Pause frame configuration at cgx
CGX LMAC, the physical interface can generate pause frames when
internal resources asserts backpressure due to exhaustion.
This patch configures CGX to generate 802.3 pause frames.
Also enabled processing of received pause frames on the line which
will assert backpressure on the internal transmit path.
Also added mailbox handlers for PF drivers to enable or disable
pause frames anytime.
Signed-off-by: Geetha sowjanya <gakula@marvell.com>
Signed-off-by: Sunil Goutham <sgoutham@marvell.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Geetha sowjanya [Mon, 2 Mar 2020 07:19:22 +0000 (12:49 +0530)]
octeontx2-af: Interface backpressure configuration
Each of the interface receive channels can be backpressured by
resources upon exhaustion or reaching configured threshold levels.
Resources here are receive buffer queues (Auras) and pkt notification
descriptor queues (CQs). Resources and interface channels are mapped
using backpressure IDs (BPIDs).
HW supports upto 512 BPIDs, this patch divides these BPIDs statically
across CGX/LBK/SDP interfaces as follows.
BPIDs 0 - 191 are mapped to LMAC channels, 16 per LMAC.
BPIDs 192 - 255 are mapped to LBK channels.
BPIDs 256 - 511 are mapped to SDP channels.
Also did the needed basic configuration of BPIDs.
Added mbox handlers with which a PF device can request for a BPID which
it will use to configure Auras and CQs.
Signed-off-by: Geetha sowjanya <gakula@marvell.com>
Signed-off-by: Sunil Goutham <sgoutham@marvell.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Heiner Kallweit [Sun, 1 Mar 2020 20:57:08 +0000 (21:57 +0100)]
net: phy: mscc: add constants for used interrupt mask bits
Add constants for the used interrupts bits. This avoids the magic
number for MII_VSC85XX_INT_MASK_MASK.
Signed-off-by: Heiner Kallweit <hkallweit1@gmail.com>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: David S. Miller <davem@davemloft.net>
Gustavo A. R. Silva [Sat, 29 Feb 2020 01:07:01 +0000 (19:07 -0600)]
arcnet: Replace zero-length array with flexible-array member
The current codebase makes use of the zero-length array language
extension to the C90 standard, but the preferred mechanism to declare
variable-length types such as these ones is a flexible array member[1][2],
introduced in C99:
struct foo {
int stuff;
struct boo array[];
};
By making use of the mechanism above, we will get a compiler warning
in case the flexible array does not occur last in the structure, which
will help us prevent some kind of undefined behavior bugs from being
inadvertently introduced[3] to the codebase from now on.
Also, notice that, dynamic memory allocations won't be affected by
this change:
"Flexible array members have incomplete type, and so the sizeof operator
may not be applied. As a quirk of the original implementation of
zero-length arrays, sizeof evaluates to zero."[1]
This issue was found with the help of Coccinelle.
[1] https://gcc.gnu.org/onlinedocs/gcc/Zero-Length.html
[2] https://github.com/KSPP/linux/issues/21
[3] commit
76497732932f ("cxgb3/l2t: Fix undefined behaviour")
Signed-off-by: Gustavo A. R. Silva <gustavo@embeddedor.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Gustavo A. R. Silva [Sat, 29 Feb 2020 01:05:02 +0000 (19:05 -0600)]
neighbour: Replace zero-length array with flexible-array member
The current codebase makes use of the zero-length array language
extension to the C90 standard, but the preferred mechanism to declare
variable-length types such as these ones is a flexible array member[1][2],
introduced in C99:
struct foo {
int stuff;
struct boo array[];
};
By making use of the mechanism above, we will get a compiler warning
in case the flexible array does not occur last in the structure, which
will help us prevent some kind of undefined behavior bugs from being
inadvertently introduced[3] to the codebase from now on.
Also, notice that, dynamic memory allocations won't be affected by
this change:
"Flexible array members have incomplete type, and so the sizeof operator
may not be applied. As a quirk of the original implementation of
zero-length arrays, sizeof evaluates to zero."[1]
This issue was found with the help of Coccinelle.
[1] https://gcc.gnu.org/onlinedocs/gcc/Zero-Length.html
[2] https://github.com/KSPP/linux/issues/21
[3] commit
76497732932f ("cxgb3/l2t: Fix undefined behaviour")
Signed-off-by: Gustavo A. R. Silva <gustavo@embeddedor.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Gustavo A. R. Silva [Sat, 29 Feb 2020 01:02:54 +0000 (19:02 -0600)]
net: flow_offload: Replace zero-length array with flexible-array member
The current codebase makes use of the zero-length array language
extension to the C90 standard, but the preferred mechanism to declare
variable-length types such as these ones is a flexible array member[1][2],
introduced in C99:
struct foo {
int stuff;
struct boo array[];
};
By making use of the mechanism above, we will get a compiler warning
in case the flexible array does not occur last in the structure, which
will help us prevent some kind of undefined behavior bugs from being
inadvertently introduced[3] to the codebase from now on.
Also, notice that, dynamic memory allocations won't be affected by
this change:
"Flexible array members have incomplete type, and so the sizeof operator
may not be applied. As a quirk of the original implementation of
zero-length arrays, sizeof evaluates to zero."[1]
This issue was found with the help of Coccinelle.
[1] https://gcc.gnu.org/onlinedocs/gcc/Zero-Length.html
[2] https://github.com/KSPP/linux/issues/21
[3] commit
76497732932f ("cxgb3/l2t: Fix undefined behaviour")
Signed-off-by: Gustavo A. R. Silva <gustavo@embeddedor.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Gustavo A. R. Silva [Sat, 29 Feb 2020 01:00:40 +0000 (19:00 -0600)]
net: dn_fib: Replace zero-length array with flexible-array member
The current codebase makes use of the zero-length array language
extension to the C90 standard, but the preferred mechanism to declare
variable-length types such as these ones is a flexible array member[1][2],
introduced in C99:
struct foo {
int stuff;
struct boo array[];
};
By making use of the mechanism above, we will get a compiler warning
in case the flexible array does not occur last in the structure, which
will help us prevent some kind of undefined behavior bugs from being
inadvertently introduced[3] to the codebase from now on.
Also, notice that, dynamic memory allocations won't be affected by
this change:
"Flexible array members have incomplete type, and so the sizeof operator
may not be applied. As a quirk of the original implementation of
zero-length arrays, sizeof evaluates to zero."[1]
This issue was found with the help of Coccinelle.
[1] https://gcc.gnu.org/onlinedocs/gcc/Zero-Length.html
[2] https://github.com/KSPP/linux/issues/21
[3] commit
76497732932f ("cxgb3/l2t: Fix undefined behaviour")
Signed-off-by: Gustavo A. R. Silva <gustavo@embeddedor.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Gustavo A. R. Silva [Sat, 29 Feb 2020 00:53:11 +0000 (18:53 -0600)]
ndisc: Replace zero-length array with flexible-array member
The current codebase makes use of the zero-length array language
extension to the C90 standard, but the preferred mechanism to declare
variable-length types such as these ones is a flexible array member[1][2],
introduced in C99:
struct foo {
int stuff;
struct boo array[];
};
By making use of the mechanism above, we will get a compiler warning
in case the flexible array does not occur last in the structure, which
will help us prevent some kind of undefined behavior bugs from being
inadvertently introduced[3] to the codebase from now on.
Also, notice that, dynamic memory allocations won't be affected by
this change:
"Flexible array members have incomplete type, and so the sizeof operator
may not be applied. As a quirk of the original implementation of
zero-length arrays, sizeof evaluates to zero."[1]
This issue was found with the help of Coccinelle.
[1] https://gcc.gnu.org/onlinedocs/gcc/Zero-Length.html
[2] https://github.com/KSPP/linux/issues/21
[3] commit
76497732932f ("cxgb3/l2t: Fix undefined behaviour")
Signed-off-by: Gustavo A. R. Silva <gustavo@embeddedor.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Gustavo A. R. Silva [Sat, 29 Feb 2020 00:51:41 +0000 (18:51 -0600)]
net: ipv6: mld: Replace zero-length array with flexible-array member
The current codebase makes use of the zero-length array language
extension to the C90 standard, but the preferred mechanism to declare
variable-length types such as these ones is a flexible array member[1][2],
introduced in C99:
struct foo {
int stuff;
struct boo array[];
};
By making use of the mechanism above, we will get a compiler warning
in case the flexible array does not occur last in the structure, which
will help us prevent some kind of undefined behavior bugs from being
inadvertently introduced[3] to the codebase from now on.
Also, notice that, dynamic memory allocations won't be affected by
this change:
"Flexible array members have incomplete type, and so the sizeof operator
may not be applied. As a quirk of the original implementation of
zero-length arrays, sizeof evaluates to zero."[1]
This issue was found with the help of Coccinelle.
[1] https://gcc.gnu.org/onlinedocs/gcc/Zero-Length.html
[2] https://github.com/KSPP/linux/issues/21
[3] commit
76497732932f ("cxgb3/l2t: Fix undefined behaviour")
Signed-off-by: Gustavo A. R. Silva <gustavo@embeddedor.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Gustavo A. R. Silva [Sat, 29 Feb 2020 00:50:23 +0000 (18:50 -0600)]
net: lwtunnel: Replace zero-length array with flexible-array member
The current codebase makes use of the zero-length array language
extension to the C90 standard, but the preferred mechanism to declare
variable-length types such as these ones is a flexible array member[1][2],
introduced in C99:
struct foo {
int stuff;
struct boo array[];
};
By making use of the mechanism above, we will get a compiler warning
in case the flexible array does not occur last in the structure, which
will help us prevent some kind of undefined behavior bugs from being
inadvertently introduced[3] to the codebase from now on.
Also, notice that, dynamic memory allocations won't be affected by
this change:
"Flexible array members have incomplete type, and so the sizeof operator
may not be applied. As a quirk of the original implementation of
zero-length arrays, sizeof evaluates to zero."[1]
This issue was found with the help of Coccinelle.
[1] https://gcc.gnu.org/onlinedocs/gcc/Zero-Length.html
[2] https://github.com/KSPP/linux/issues/21
[3] commit
76497732932f ("cxgb3/l2t: Fix undefined behaviour")
Signed-off-by: Gustavo A. R. Silva <gustavo@embeddedor.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Gustavo A. R. Silva [Sat, 29 Feb 2020 00:44:10 +0000 (18:44 -0600)]
net: ip6_route: Replace zero-length array with flexible-array member
The current codebase makes use of the zero-length array language
extension to the C90 standard, but the preferred mechanism to declare
variable-length types such as these ones is a flexible array member[1][2],
introduced in C99:
struct foo {
int stuff;
struct boo array[];
};
By making use of the mechanism above, we will get a compiler warning
in case the flexible array does not occur last in the structure, which
will help us prevent some kind of undefined behavior bugs from being
inadvertently introduced[3] to the codebase from now on.
Also, notice that, dynamic memory allocations won't be affected by
this change:
"Flexible array members have incomplete type, and so the sizeof operator
may not be applied. As a quirk of the original implementation of
zero-length arrays, sizeof evaluates to zero."[1]
This issue was found with the help of Coccinelle.
[1] https://gcc.gnu.org/onlinedocs/gcc/Zero-Length.html
[2] https://github.com/KSPP/linux/issues/21
[3] commit
76497732932f ("cxgb3/l2t: Fix undefined behaviour")
Signed-off-by: Gustavo A. R. Silva <gustavo@embeddedor.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Gustavo A. R. Silva [Sat, 29 Feb 2020 00:14:11 +0000 (18:14 -0600)]
net: nexthop: Replace zero-length array with flexible-array member
The current codebase makes use of the zero-length array language
extension to the C90 standard, but the preferred mechanism to declare
variable-length types such as these ones is a flexible array member[1][2],
introduced in C99:
struct foo {
int stuff;
struct boo array[];
};
By making use of the mechanism above, we will get a compiler warning
in case the flexible array does not occur last in the structure, which
will help us prevent some kind of undefined behavior bugs from being
inadvertently introduced[3] to the codebase from now on.
Also, notice that, dynamic memory allocations won't be affected by
this change:
"Flexible array members have incomplete type, and so the sizeof operator
may not be applied. As a quirk of the original implementation of
zero-length arrays, sizeof evaluates to zero."[1]
This issue was found with the help of Coccinelle.
[1] https://gcc.gnu.org/onlinedocs/gcc/Zero-Length.html
[2] https://github.com/KSPP/linux/issues/21
[3] commit
76497732932f ("cxgb3/l2t: Fix undefined behaviour")
Signed-off-by: Gustavo A. R. Silva <gustavo@embeddedor.com>
Reviewed-by: David Ahern <dsahern@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Gustavo A. R. Silva [Sat, 29 Feb 2020 00:13:05 +0000 (18:13 -0600)]
net: sctp: Replace zero-length array with flexible-array member
The current codebase makes use of the zero-length array language
extension to the C90 standard, but the preferred mechanism to declare
variable-length types such as these ones is a flexible array member[1][2],
introduced in C99:
struct foo {
int stuff;
struct boo array[];
};
By making use of the mechanism above, we will get a compiler warning
in case the flexible array does not occur last in the structure, which
will help us prevent some kind of undefined behavior bugs from being
inadvertently introduced[3] to the codebase from now on.
Also, notice that, dynamic memory allocations won't be affected by
this change:
"Flexible array members have incomplete type, and so the sizeof operator
may not be applied. As a quirk of the original implementation of
zero-length arrays, sizeof evaluates to zero."[1]
This issue was found with the help of Coccinelle.
[1] https://gcc.gnu.org/onlinedocs/gcc/Zero-Length.html
[2] https://github.com/KSPP/linux/issues/21
[3] commit
76497732932f ("cxgb3/l2t: Fix undefined behaviour")
Signed-off-by: Gustavo A. R. Silva <gustavo@embeddedor.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Gustavo A. R. Silva [Sat, 29 Feb 2020 00:11:02 +0000 (18:11 -0600)]
net: sock_reuseport: Replace zero-length array with flexible-array member
The current codebase makes use of the zero-length array language
extension to the C90 standard, but the preferred mechanism to declare
variable-length types such as these ones is a flexible array member[1][2],
introduced in C99:
struct foo {
int stuff;
struct boo array[];
};
By making use of the mechanism above, we will get a compiler warning
in case the flexible array does not occur last in the structure, which
will help us prevent some kind of undefined behavior bugs from being
inadvertently introduced[3] to the codebase from now on.
Also, notice that, dynamic memory allocations won't be affected by
this change:
"Flexible array members have incomplete type, and so the sizeof operator
may not be applied. As a quirk of the original implementation of
zero-length arrays, sizeof evaluates to zero."[1]
This issue was found with the help of Coccinelle.
[1] https://gcc.gnu.org/onlinedocs/gcc/Zero-Length.html
[2] https://github.com/KSPP/linux/issues/21
[3] commit
76497732932f ("cxgb3/l2t: Fix undefined behaviour")
Signed-off-by: Gustavo A. R. Silva <gustavo@embeddedor.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
David S. Miller [Sun, 1 Mar 2020 05:48:55 +0000 (21:48 -0800)]
Merge branch 'net-ethtool-Introduce-link_ksettings-API-for-virtual-network-devices'
Cris Forno says:
====================
net/ethtool: Introduce link_ksettings API for virtual network devices
This series provides an API for drivers of virtual network devices that
allows users to alter initial device speed and duplex settings to reflect
the actual capabilities of underlying hardware. The changes made include
a helper function ethtool_virtdev_set_link_ksettings, which is used to
retrieve alterable link settings. In addition, there is a new ethtool
function defined to validate those settings. These changes resolve code
duplication for existing virtual network drivers that have already
implemented this behavior. In the case of the ibmveth driver, this API is
used to provide this capability for the first time.
---
v7: - removed ethtool_validate_cmd function pointer parameter from
ethtool_virtdev_set_link_ksettings since none of the virtual drivers
pass in a custom validate function as suggested by Michal Kubecek.
v6: - removed netvsc_validate_ethtool_ss_cmd(). netvsc_drv now uses
ethtool_virtdev_validate_cmd() instead as suggested by Michal Kubecek
and approved by Haiyang Zhang.
- matched handler argument name of ethtool_virtdev_set_link_ksettings
in declaration and definition as suggested by Michal Kubecek.
- shortened validate variable assignment in
ethtool_virtdev_set_link_ksettings as suggested by Michal Kubecek.
v5: - virtdev_validate_link_ksettings is taken out of the ethtool global
structure and is instead added as an argument to
ethtool_virtdev_set_link_ksettings as suggested by Jakub Kicinski.
v4: - Cleaned up return statement in ethtool_virtdev_validate_cmd based
off of Michal Kubecek's and Thomas Falcon's suggestion.
- If the netvsc driver is using the VF device in order to get
accelerated networking, the real speed and duplex is reported by using
the VF device as suggested by Stephen Hemminger.
- The speed and duplex variables are now passed by value rather than
passed by pointer as suggested by Willem de Bruijin and Michal
Kubecek.
- Removed ethtool_virtdev_get_link_ksettings since it was too simple
to warrant a helper function.
v3: - Factored out duplicated code to core/ethtool to provide API to
virtual drivers
v2: - Updated default driver speed/duplex settings to avoid breaking
existing setups
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
Cris Forno [Fri, 28 Feb 2020 20:12:05 +0000 (14:12 -0600)]
net/ethtool: Introduce link_ksettings API for virtual network devices
With the ethtool_virtdev_set_link_ksettings function in core/ethtool.c,
ibmveth, netvsc, and virtio now use the core's helper function.
Funtionality changes that pertain to ibmveth driver include:
1. Changed the initial hardcoded link speed to 1GB.
2. Added support for allowing a user to change the reported link
speed via ethtool.
Functionality changes to the netvsc driver include:
1. When netvsc_get_link_ksettings is called, it will defer to the VF
device if it exists to pull accelerated networking values, otherwise
pull default or user-defined values.
2. Similarly, if netvsc_set_link_ksettings called and a VF device
exists, the real values of speed and duplex are changed.
Signed-off-by: Cris Forno <cforno12@linux.vnet.ibm.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Cris Forno [Fri, 28 Feb 2020 20:12:04 +0000 (14:12 -0600)]
ethtool: Factored out similar ethtool link settings for virtual devices to core
Three virtual devices (ibmveth, virtio_net, and netvsc) all have
similar code to set link settings and validate ethtool command. To
eliminate duplication of code, it is factored out into core/ethtool.c.
Signed-off-by: Cris Forno <cforno12@linux.vnet.ibm.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
David S. Miller [Sun, 1 Mar 2020 05:37:03 +0000 (21:37 -0800)]
Merge branch 'hsr-several-code-cleanup-for-hsr-module'
Taehee Yoo says:
====================
hsr: several code cleanup for hsr module
This patchset is to clean up hsr module code.
1. The first patch is to use debugfs_remove_recursive().
If it uses debugfs_remove_recursive() instead of debugfs_remove(),
hsr_priv() doesn't need to have "node_tbl_file" pointer variable.
2. The second patch is to use extack error message.
If HSR uses the extack instead of netdev_info(), users can get
error messages immediately without any checking the kernel message.
3. The third patch is to use netdev_err() instead of WARN_ONCE().
When a packet is being sent, hsr_addr_subst_dest() is called and
it tries to find the node with the ethernet destination address.
If it couldn't find a node, it warns with WARN_ONCE().
But, using WARN_ONCE() is a little bit overdoing.
So, in this patch, netdev_err() is used instead.
4. The fourth patch is to remove unnecessary rcu_read_{lock/unlock}().
There are some rcu_read_{lock/unlock}() in hsr module and some of
them are unnecessary. In this patch,
these unnecessary rcu_read_{lock/unlock}() will be removed.
5. The fifth patch is to use upper/lower device infrastructure.
netdev_upper_dev_link() is useful to manage lower/upper interfaces.
And this function internally validates looping, maximum depth.
If hsr module uses upper/lower device infrastructure,
it can prevent these above problems.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
Taehee Yoo [Fri, 28 Feb 2020 18:02:10 +0000 (18:02 +0000)]
hsr: use upper/lower device infrastructure
netdev_upper_dev_link() is useful to manage lower/upper interfaces.
And this function internally validates looping, maximum depth.
All or most virtual interfaces that could have a real interface
(e.g. macsec, macvlan, ipvlan etc.) use lower/upper infrastructure.
Signed-off-by: Taehee Yoo <ap420073@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Taehee Yoo [Fri, 28 Feb 2020 18:01:56 +0000 (18:01 +0000)]
hsr: remove unnecessary rcu_read_lock() in hsr module
In order to access the port list, the hsr_port_get_hsr() is used.
And this is protected by RTNL and RCU.
The hsr_fill_info(), hsr_check_carrier(), hsr_dev_open() and
hsr_get_max_mtu() are protected by RTNL.
So, rcu_read_lock() in these functions are not necessary.
The hsr_handle_frame() also uses rcu_read_lock() but this function
is called by packet path.
It's already protected by RCU.
So, the rcu_read_lock() in hsr_handle_frame() can be removed.
Signed-off-by: Taehee Yoo <ap420073@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Taehee Yoo [Fri, 28 Feb 2020 18:01:46 +0000 (18:01 +0000)]
hsr: use netdev_err() instead of WARN_ONCE()
When HSR interface is sending a frame, it finds a node with
the destination ethernet address from the list.
If there is no node, it calls WARN_ONCE().
But, using WARN_ONCE() for this situation is a little bit overdoing.
So, in this patch, the netdev_err() is used instead.
Signed-off-by: Taehee Yoo <ap420073@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Taehee Yoo [Fri, 28 Feb 2020 18:01:35 +0000 (18:01 +0000)]
hsr: use extack error message instead of netdev_info
If HSR uses the extack instead of netdev_info(), users can get
error messages immediately without any checking the kernel message.
Signed-off-by: Taehee Yoo <ap420073@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Taehee Yoo [Fri, 28 Feb 2020 18:01:20 +0000 (18:01 +0000)]
hsr: use debugfs_remove_recursive() instead of debugfs_remove()
If it uses debugfs_remove_recursive() instead of debugfs_remove(),
hsr_priv() doesn't need to have "node_tbl_file" pointer variable.
Signed-off-by: Taehee Yoo <ap420073@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Oleksij Rempel [Fri, 28 Feb 2020 14:50:49 +0000 (15:50 +0100)]
net: ag71xx: port to phylink
The port to phylink was done as close as possible to initial
functionality.
Signed-off-by: Oleksij Rempel <o.rempel@pengutronix.de>
Acked-by: Russell King <rmk+kernel@armlinux.org.uk>
Signed-off-by: David S. Miller <davem@davemloft.net>
David S. Miller [Sun, 1 Mar 2020 05:30:43 +0000 (21:30 -0800)]
Merge branch 'net-ll_temac-RX-TX-ring-size-and-coalesce-ethtool-parameters'
Esben Haabendal says:
====================
net: ll_temac: RX/TX ring size and coalesce ethtool parameters
This series adds support for RX/TX ring size and irq coalesce ethtool
parameters to ll_temac driver.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
Esben Haabendal [Fri, 28 Feb 2020 07:57:41 +0000 (08:57 +0100)]
net: ll_temac: Add ethtool support for coalesce parameters
Please note that the delays are calculated based on typical
parameters. But as TEMAC is an HDL IP, designs may vary, and future
work might be needed to make this calculation configurable.
Signed-off-by: Esben Haabendal <esben@geanix.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Esben Haabendal [Fri, 28 Feb 2020 07:57:26 +0000 (08:57 +0100)]
net: ll_temac: Make RX/TX ring sizes configurable
Add support for setting the RX and TX ring sizes for this driver using
ethtool. Also increase the default RX ring size as the previous default
was far too low for good performance in some configurations.
Signed-off-by: Esben Haabendal <esben@geanix.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Esben Haabendal [Fri, 28 Feb 2020 07:57:12 +0000 (08:57 +0100)]
net: ll_temac: Remove unused start_p variable
The start_p variable was included in the initial commit,
commit
92744989533c ("net: add Xilinx ll_temac device driver"),
but has never had any real use.
Signed-off-by: Esben Haabendal <esben@geanix.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Esben Haabendal [Fri, 28 Feb 2020 07:56:57 +0000 (08:56 +0100)]
net: ll_temac: Remove unused tx_bd_next struct field
The tx_bd_next field was included in the initial commit,
commit
92744989533c ("net: add Xilinx ll_temac device driver"),
but has never had any real use.
Signed-off-by: Esben Haabendal <esben@geanix.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Gustavo A. R. Silva [Thu, 27 Feb 2020 20:58:44 +0000 (14:58 -0600)]
net: sched: Replace zero-length array with flexible-array member
The current codebase makes use of the zero-length array language
extension to the C90 standard, but the preferred mechanism to declare
variable-length types such as these ones is a flexible array member[1][2],
introduced in C99:
struct foo {
int stuff;
struct boo array[];
};
By making use of the mechanism above, we will get a compiler warning
in case the flexible array does not occur last in the structure, which
will help us prevent some kind of undefined behavior bugs from being
inadvertently introduced[3] to the codebase from now on.
Also, notice that, dynamic memory allocations won't be affected by
this change:
"Flexible array members have incomplete type, and so the sizeof operator
may not be applied. As a quirk of the original implementation of
zero-length arrays, sizeof evaluates to zero."[1]
This issue was found with the help of Coccinelle.
[1] https://gcc.gnu.org/onlinedocs/gcc/Zero-Length.html
[2] https://github.com/KSPP/linux/issues/21
[3] commit
76497732932f ("cxgb3/l2t: Fix undefined behaviour")
Signed-off-by: Gustavo A. R. Silva <gustavo@embeddedor.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
David S. Miller [Sat, 29 Feb 2020 23:53:35 +0000 (15:53 -0800)]
Merge git://git./linux/kernel/git/bpf/bpf-next
Alexei Starovoitov says:
====================
pull-request: bpf-next 2020-02-28
The following pull-request contains BPF updates for your *net-next* tree.
We've added 41 non-merge commits during the last 7 day(s) which contain
a total of 49 files changed, 1383 insertions(+), 499 deletions(-).
The main changes are:
1) BPF and Real-Time nicely co-exist.
2) bpftool feature improvements.
3) retrieve bpf_sk_storage via INET_DIAG.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
David S. Miller [Fri, 28 Feb 2020 20:13:02 +0000 (12:13 -0800)]
Merge branch 'net-cleanup-datagram-receive-helpers'
Paolo Abeni says:
====================
net: cleanup datagram receive helpers
Several receive helpers have an optional destructor argument, which uglify
the code a bit and is taxed by retpoline overhead.
This series refactor the code so that we can drop such optional argument,
cleaning the helpers a bit and avoiding an indirect call in fast path.
The first patch refactor a bit the caller, so that the second patch
actually dropping the argument is more straight-forward
v1 -> v2:
- call scm_stat_del() only when not peeking - Kirill
- fix build issue with CONFIG_INET_ESPINTCP
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
Paolo Abeni [Fri, 28 Feb 2020 13:45:22 +0000 (14:45 +0100)]
net: datagram: drop 'destructor' argument from several helpers
The only users for such argument are the UDP protocol and the UNIX
socket family. We can safely reclaim the accounted memory directly
from the UDP code and, after the previous patch, we can do scm
stats accounting outside the datagram helpers.
Overall this cleans up a bit some datagram-related helpers, and
avoids an indirect call per packet in the UDP receive path.
v1 -> v2:
- call scm_stat_del() only when not peeking - Kirill
- fix build issue with CONFIG_INET_ESPINTCP
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Reviewed-by: Kirill Tkhai <ktkhai@virtuozzo.com>
Reviewed-by: Willem de Bruijn <willemb@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Paolo Abeni [Fri, 28 Feb 2020 13:45:21 +0000 (14:45 +0100)]
unix: uses an atomic type for scm files accounting
So the scm_stat_{add,del} helper can be invoked with no
additional lock held.
This clean-up the code a bit and will make the next
patch easier.
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Reviewed-by: Kirill Tkhai <ktkhai@virtuozzo.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Gustavo A. R. Silva [Fri, 28 Feb 2020 14:01:43 +0000 (08:01 -0600)]
af_unix: Replace zero-length array with flexible-array member
The current codebase makes use of the zero-length array language
extension to the C90 standard, but the preferred mechanism to declare
variable-length types such as these ones is a flexible array member[1][2],
introduced in C99:
struct foo {
int stuff;
struct boo array[];
};
By making use of the mechanism above, we will get a compiler warning
in case the flexible array does not occur last in the structure, which
will help us prevent some kind of undefined behavior bugs from being
inadvertently introduced[3] to the codebase from now on.
Also, notice that, dynamic memory allocations won't be affected by
this change:
"Flexible array members have incomplete type, and so the sizeof operator
may not be applied. As a quirk of the original implementation of
zero-length arrays, sizeof evaluates to zero."[1]
This issue was found with the help of Coccinelle.
[1] https://gcc.gnu.org/onlinedocs/gcc/Zero-Length.html
[2] https://github.com/KSPP/linux/issues/21
[3] commit
76497732932f ("cxgb3/l2t: Fix undefined behaviour")
Signed-off-by: Gustavo A. R. Silva <gustavo@embeddedor.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Gustavo A. R. Silva [Fri, 28 Feb 2020 13:56:29 +0000 (07:56 -0600)]
bonding: Replace zero-length array with flexible-array member
The current codebase makes use of the zero-length array language
extension to the C90 standard, but the preferred mechanism to declare
variable-length types such as these ones is a flexible array member[1][2],
introduced in C99:
struct foo {
int stuff;
struct boo array[];
};
By making use of the mechanism above, we will get a compiler warning
in case the flexible array does not occur last in the structure, which
will help us prevent some kind of undefined behavior bugs from being
inadvertently introduced[3] to the codebase from now on.
Also, notice that, dynamic memory allocations won't be affected by
this change:
"Flexible array members have incomplete type, and so the sizeof operator
may not be applied. As a quirk of the original implementation of
zero-length arrays, sizeof evaluates to zero."[1]
This issue was found with the help of Coccinelle.
[1] https://gcc.gnu.org/onlinedocs/gcc/Zero-Length.html
[2] https://github.com/KSPP/linux/issues/21
[3] commit
76497732932f ("cxgb3/l2t: Fix undefined behaviour")
Signed-off-by: Gustavo A. R. Silva <gustavo@embeddedor.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Gustavo A. R. Silva [Fri, 28 Feb 2020 13:43:24 +0000 (07:43 -0600)]
net: core: Replace zero-length array with flexible-array member
The current codebase makes use of the zero-length array language
extension to the C90 standard, but the preferred mechanism to declare
variable-length types such as these ones is a flexible array member[1][2],
introduced in C99:
struct foo {
int stuff;
struct boo array[];
};
By making use of the mechanism above, we will get a compiler warning
in case the flexible array does not occur last in the structure, which
will help us prevent some kind of undefined behavior bugs from being
inadvertently introduced[3] to the codebase from now on.
Also, notice that, dynamic memory allocations won't be affected by
this change:
"Flexible array members have incomplete type, and so the sizeof operator
may not be applied. As a quirk of the original implementation of
zero-length arrays, sizeof evaluates to zero."[1]
This issue was found with the help of Coccinelle.
[1] https://gcc.gnu.org/onlinedocs/gcc/Zero-Length.html
[2] https://github.com/KSPP/linux/issues/21
[3] commit
76497732932f ("cxgb3/l2t: Fix undefined behaviour")
Signed-off-by: Gustavo A. R. Silva <gustavo@embeddedor.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Gustavo A. R. Silva [Fri, 28 Feb 2020 13:36:41 +0000 (07:36 -0600)]
ipv6: Replace zero-length array with flexible-array member
The current codebase makes use of the zero-length array language
extension to the C90 standard, but the preferred mechanism to declare
variable-length types such as these ones is a flexible array member[1][2],
introduced in C99:
struct foo {
int stuff;
struct boo array[];
};
By making use of the mechanism above, we will get a compiler warning
in case the flexible array does not occur last in the structure, which
will help us prevent some kind of undefined behavior bugs from being
inadvertently introduced[3] to the codebase from now on.
Also, notice that, dynamic memory allocations won't be affected by
this change:
"Flexible array members have incomplete type, and so the sizeof operator
may not be applied. As a quirk of the original implementation of
zero-length arrays, sizeof evaluates to zero."[1]
This issue was found with the help of Coccinelle.
[1] https://gcc.gnu.org/onlinedocs/gcc/Zero-Length.html
[2] https://github.com/KSPP/linux/issues/21
[3] commit
76497732932f ("cxgb3/l2t: Fix undefined behaviour")
Signed-off-by: Gustavo A. R. Silva <gustavo@embeddedor.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Gustavo A. R. Silva [Fri, 28 Feb 2020 13:33:37 +0000 (07:33 -0600)]
net: dccp: Replace zero-length array with flexible-array member
The current codebase makes use of the zero-length array language
extension to the C90 standard, but the preferred mechanism to declare
variable-length types such as these ones is a flexible array member[1][2],
introduced in C99:
struct foo {
int stuff;
struct boo array[];
};
By making use of the mechanism above, we will get a compiler warning
in case the flexible array does not occur last in the structure, which
will help us prevent some kind of undefined behavior bugs from being
inadvertently introduced[3] to the codebase from now on.
Also, notice that, dynamic memory allocations won't be affected by
this change:
"Flexible array members have incomplete type, and so the sizeof operator
may not be applied. As a quirk of the original implementation of
zero-length arrays, sizeof evaluates to zero."[1]
This issue was found with the help of Coccinelle.
[1] https://gcc.gnu.org/onlinedocs/gcc/Zero-Length.html
[2] https://github.com/KSPP/linux/issues/21
[3] commit
76497732932f ("cxgb3/l2t: Fix undefined behaviour")
Signed-off-by: Gustavo A. R. Silva <gustavo@embeddedor.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Gustavo A. R. Silva [Fri, 28 Feb 2020 13:30:45 +0000 (07:30 -0600)]
l2tp: Replace zero-length array with flexible-array member
The current codebase makes use of the zero-length array language
extension to the C90 standard, but the preferred mechanism to declare
variable-length types such as these ones is a flexible array member[1][2],
introduced in C99:
struct foo {
int stuff;
struct boo array[];
};
By making use of the mechanism above, we will get a compiler warning
in case the flexible array does not occur last in the structure, which
will help us prevent some kind of undefined behavior bugs from being
inadvertently introduced[3] to the codebase from now on.
Also, notice that, dynamic memory allocations won't be affected by
this change:
"Flexible array members have incomplete type, and so the sizeof operator
may not be applied. As a quirk of the original implementation of
zero-length arrays, sizeof evaluates to zero."[1]
Lastly, fix the following checkpatch warning:
CHECK: Prefer kernel type 'u8' over 'uint8_t'
#50: FILE: net/l2tp/l2tp_core.h:119:
+ uint8_t priv[]; /* private data */
This issue was found with the help of Coccinelle.
[1] https://gcc.gnu.org/onlinedocs/gcc/Zero-Length.html
[2] https://github.com/KSPP/linux/issues/21
[3] commit
76497732932f ("cxgb3/l2t: Fix undefined behaviour")
Signed-off-by: Gustavo A. R. Silva <gustavo@embeddedor.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Gustavo A. R. Silva [Fri, 28 Feb 2020 13:26:15 +0000 (07:26 -0600)]
net: mpls: Replace zero-length array with flexible-array member
The current codebase makes use of the zero-length array language
extension to the C90 standard, but the preferred mechanism to declare
variable-length types such as these ones is a flexible array member[1][2],
introduced in C99:
struct foo {
int stuff;
struct boo array[];
};
By making use of the mechanism above, we will get a compiler warning
in case the flexible array does not occur last in the structure, which
will help us prevent some kind of undefined behavior bugs from being
inadvertently introduced[3] to the codebase from now on.
Also, notice that, dynamic memory allocations won't be affected by
this change:
"Flexible array members have incomplete type, and so the sizeof operator
may not be applied. As a quirk of the original implementation of
zero-length arrays, sizeof evaluates to zero."[1]
This issue was found with the help of Coccinelle.
[1] https://gcc.gnu.org/onlinedocs/gcc/Zero-Length.html
[2] https://github.com/KSPP/linux/issues/21
[3] commit
76497732932f ("cxgb3/l2t: Fix undefined behaviour")
Signed-off-by: Gustavo A. R. Silva <gustavo@embeddedor.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Gustavo A. R. Silva [Fri, 28 Feb 2020 13:19:07 +0000 (07:19 -0600)]
xdp: Replace zero-length array with flexible-array member
The current codebase makes use of the zero-length array language
extension to the C90 standard, but the preferred mechanism to declare
variable-length types such as these ones is a flexible array member[1][2],
introduced in C99:
struct foo {
int stuff;
struct boo array[];
};
By making use of the mechanism above, we will get a compiler warning
in case the flexible array does not occur last in the structure, which
will help us prevent some kind of undefined behavior bugs from being
inadvertently introduced[3] to the codebase from now on.
Also, notice that, dynamic memory allocations won't be affected by
this change:
"Flexible array members have incomplete type, and so the sizeof operator
may not be applied. As a quirk of the original implementation of
zero-length arrays, sizeof evaluates to zero."[1]
This issue was found with the help of Coccinelle.
[1] https://gcc.gnu.org/onlinedocs/gcc/Zero-Length.html
[2] https://github.com/KSPP/linux/issues/21
[3] commit
76497732932f ("cxgb3/l2t: Fix undefined behaviour")
Signed-off-by: Gustavo A. R. Silva <gustavo@embeddedor.com>
Acked-by: Jonathan Lemon <jonathan.lemon@gmail.com>
Acked-by: Björn Töpel <bjorn.topel@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
David S. Miller [Fri, 28 Feb 2020 19:59:53 +0000 (11:59 -0800)]
Merge tag 'mlx5-updates-2020-02-27' of git://git./linux/kernel/git/saeed/linux
Saeed Mahameed says:
====================
mlx5-updates-2020-02-27
mlx5 misc updates and minor cleanups:
1) Use per vport tables for mirroring
2) Improve log messages for SW steering (DR)
3) Add devlink fdb_large_groups parameter
4) E-Switch, Allow goto earlier chain
5) Don't allow forwarding between uplink representors
6) Add support for devlink-port in non-representors mode
7) Minor misc cleanups
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
Alexei Starovoitov [Fri, 28 Feb 2020 02:50:20 +0000 (18:50 -0800)]
Merge branch 'bpf_sk_storage_via_inet_diag'
Martin KaFai Lau says:
====================
The bpf_prog can store specific info to a sk by using bpf_sk_storage.
In other words, a sk can be extended by a bpf_prog.
This series is to support providing bpf_sk_storage data during inet_diag's
dump. The primary target is the usage like iproute2's "ss".
The first two patches are refactoring works in inet_diag to make
adding bpf_sk_storage support easier. The next two patches do
the actual work.
Please see individual patch for details.
v2:
- Add commit message for u16 to u32 change in min_dump_alloc in Patch 4 (Song)
- Add comment to explain the !skb->len check in __inet_diag_dump in Patch 4.
- Do the map->map_type check earlier in Patch 3 for readability.
====================
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Martin KaFai Lau [Tue, 25 Feb 2020 23:04:27 +0000 (15:04 -0800)]
bpf: inet_diag: Dump bpf_sk_storages in inet_diag_dump()
This patch will dump out the bpf_sk_storages of a sk
if the request has the INET_DIAG_REQ_SK_BPF_STORAGES nlattr.
An array of SK_DIAG_BPF_STORAGE_REQ_MAP_FD can be specified in
INET_DIAG_REQ_SK_BPF_STORAGES to select which bpf_sk_storage to dump.
If no map_fd is specified, all bpf_sk_storages of a sk will be dumped.
bpf_sk_storages can be added to the system at runtime. It is difficult
to find a proper static value for cb->min_dump_alloc.
This patch learns the nlattr size required to dump the bpf_sk_storages
of a sk. If it happens to be the very first nlmsg of a dump and it
cannot fit the needed bpf_sk_storages, it will try to expand the
skb by "pskb_expand_head()".
Instead of expanding it in inet_sk_diag_fill(), it is expanded at a
sleepable context in __inet_diag_dump() so __GFP_DIRECT_RECLAIM can
be used. In __inet_diag_dump(), it will retry as long as the
skb is empty and the cb->min_dump_alloc becomes larger than before.
cb->min_dump_alloc is bounded by KMALLOC_MAX_SIZE. The min_dump_alloc
is also changed from 'u16' to 'u32' to accommodate a sk that may have
a few large bpf_sk_storages.
The updated cb->min_dump_alloc will also be used to allocate the skb in
the next dump. This logic already exists in netlink_dump().
Here is the sample output of a locally modified 'ss' and it could be made
more readable by using BTF later:
[root@arch-fb-vm1 ~]# ss --bpf-map-id 14 --bpf-map-id 13 -t6an 'dst [::1]:8989'
State Recv-Q Send-Q Local Address:Port Peer Address:PortProcess
ESTAB 0 0 [::1]:51072 [::1]:8989
bpf_map_id:14 value:[ 3feb ]
bpf_map_id:13 value:[ 3f ]
ESTAB 0 0 [::1]:51070 [::1]:8989
bpf_map_id:14 value:[ 3feb ]
bpf_map_id:13 value:[ 3f ]
[root@arch-fb-vm1 ~]# ~/devshare/github/iproute2/misc/ss --bpf-maps -t6an 'dst [::1]:8989'
State Recv-Q Send-Q Local Address:Port Peer Address:Port Process
ESTAB 0 0 [::1]:51072 [::1]:8989
bpf_map_id:14 value:[ 3feb ]
bpf_map_id:13 value:[ 3f ]
bpf_map_id:12 value:[ 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000... total:65407 ]
ESTAB 0 0 [::1]:51070 [::1]:8989
bpf_map_id:14 value:[ 3feb ]
bpf_map_id:13 value:[ 3f ]
bpf_map_id:12 value:[ 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000... total:65407 ]
Signed-off-by: Martin KaFai Lau <kafai@fb.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Acked-by: Song Liu <songliubraving@fb.com>
Link: https://lore.kernel.org/bpf/20200225230427.1976129-1-kafai@fb.com
Martin KaFai Lau [Tue, 25 Feb 2020 23:04:21 +0000 (15:04 -0800)]
bpf: INET_DIAG support in bpf_sk_storage
This patch adds INET_DIAG support to bpf_sk_storage.
1. Although this series adds bpf_sk_storage diag capability to inet sk,
bpf_sk_storage is in general applicable to all fullsock. Hence, the
bpf_sk_storage logic will operate on SK_DIAG_* nlattr. The caller
will pass in its specific nesting nlattr (e.g. INET_DIAG_*) as
the argument.
2. The request will be like:
INET_DIAG_REQ_SK_BPF_STORAGES (nla_nest) (defined in latter patch)
SK_DIAG_BPF_STORAGE_REQ_MAP_FD (nla_put_u32)
SK_DIAG_BPF_STORAGE_REQ_MAP_FD (nla_put_u32)
......
Considering there could have multiple bpf_sk_storages in a sk,
instead of reusing INET_DIAG_INFO ("ss -i"), the user can select
some specific bpf_sk_storage to dump by specifying an array of
SK_DIAG_BPF_STORAGE_REQ_MAP_FD.
If no SK_DIAG_BPF_STORAGE_REQ_MAP_FD is specified (i.e. an empty
INET_DIAG_REQ_SK_BPF_STORAGES), it will dump all bpf_sk_storages
of a sk.
3. The reply will be like:
INET_DIAG_BPF_SK_STORAGES (nla_nest) (defined in latter patch)
SK_DIAG_BPF_STORAGE (nla_nest)
SK_DIAG_BPF_STORAGE_MAP_ID (nla_put_u32)
SK_DIAG_BPF_STORAGE_MAP_VALUE (nla_reserve_64bit)
SK_DIAG_BPF_STORAGE (nla_nest)
SK_DIAG_BPF_STORAGE_MAP_ID (nla_put_u32)
SK_DIAG_BPF_STORAGE_MAP_VALUE (nla_reserve_64bit)
......
4. Unlike other INET_DIAG info of a sk which is pretty static, the size
required to dump the bpf_sk_storage(s) of a sk is dynamic as the
system adding more bpf_sk_storage_map. It is hard to set a static
min_dump_alloc size.
Hence, this series learns it at the runtime and adjust the
cb->min_dump_alloc as it iterates all sk(s) of a system. The
"unsigned int *res_diag_size" in bpf_sk_storage_diag_put()
is for this purpose.
The next patch will update the cb->min_dump_alloc as it
iterates the sk(s).
Signed-off-by: Martin KaFai Lau <kafai@fb.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Acked-by: Song Liu <songliubraving@fb.com>
Link: https://lore.kernel.org/bpf/20200225230421.1975729-1-kafai@fb.com
Martin KaFai Lau [Tue, 25 Feb 2020 23:04:15 +0000 (15:04 -0800)]
inet_diag: Move the INET_DIAG_REQ_BYTECODE nlattr to cb->data
The INET_DIAG_REQ_BYTECODE nlattr is currently re-found every time when
the "dump()" is re-started.
In a latter patch, it will also need to parse the new
INET_DIAG_REQ_SK_BPF_STORAGES nlattr to learn the map_fds. Thus, this
patch takes this chance to store the parsed nlattr in cb->data
during the "start" time of a dump.
By doing this, the "bc" argument also becomes unnecessary
and is removed. Also, the two copies of the INET_DIAG_REQ_BYTECODE
parsing-audit logic between compat/current version can be
consolidated to one.
Signed-off-by: Martin KaFai Lau <kafai@fb.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Acked-by: Song Liu <songliubraving@fb.com>
Link: https://lore.kernel.org/bpf/20200225230415.1975555-1-kafai@fb.com
Martin KaFai Lau [Tue, 25 Feb 2020 23:04:09 +0000 (15:04 -0800)]
inet_diag: Refactor inet_sk_diag_fill(), dump(), and dump_one()
In a latter patch, there is a need to update "cb->min_dump_alloc"
in inet_sk_diag_fill() as it learns the diffierent bpf_sk_storages
stored in a sk while dumping all sk(s) (e.g. tcp_hashinfo).
The inet_sk_diag_fill() currently does not take the "cb" as an argument.
One of the reason is inet_sk_diag_fill() is used by both dump_one()
and dump() (which belong to the "struct inet_diag_handler". The dump_one()
interface does not pass the "cb" along.
This patch is to make dump_one() pass a "cb". The "cb" is created in
inet_diag_cmd_exact(). The "nlh" and "in_skb" are stored in "cb" as
the dump() interface does. The total number of args in
inet_sk_diag_fill() is also cut from 10 to 7 and
that helps many callers to pass fewer args.
In particular,
"struct user_namespace *user_ns", "u32 pid", and "u32 seq"
can be replaced by accessing "cb->nlh" and "cb->skb".
A similar argument reduction is also made to
inet_twsk_diag_fill() and inet_req_diag_fill().
inet_csk_diag_dump() and inet_csk_diag_fill() are also removed.
They are mostly equivalent to inet_sk_diag_fill(). Their repeated
usages are very limited. Thus, inet_sk_diag_fill() is directly used
in those occasions.
Signed-off-by: Martin KaFai Lau <kafai@fb.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Acked-by: Song Liu <songliubraving@fb.com>
Link: https://lore.kernel.org/bpf/20200225230409.1975173-1-kafai@fb.com
David S. Miller [Fri, 28 Feb 2020 02:31:39 +0000 (18:31 -0800)]
Merge git://git./linux/kernel/git/netdev/net
The mptcp conflict was overlapping additions.
The SMC conflict was an additional and removal happening at the same
time.
Signed-off-by: David S. Miller <davem@davemloft.net>
Roi Dayan [Thu, 13 Feb 2020 12:19:50 +0000 (14:19 +0200)]
net/mlx5e: Remove redundant comment about goto slow path
The code is self explanatory and makes the comment redundant.
Signed-off-by: Roi Dayan <roid@mellanox.com>
Reviewed-by: Eli Cohen <eli@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
Eli Cohen [Thu, 13 Feb 2020 09:12:16 +0000 (11:12 +0200)]
net/mlx5e: Reduce number of arguments in slow path handling
mlx5e_tc_offload_to_slow_path() and mlx5e_tc_unoffload_from_slow_path()
take an extra argument allocated on the stack of the caller but not used
by the caller. Avoid the extra argument and use local variable in the
function itself.
Signed-off-by: Eli Cohen <eli@mellanox.com>
Reviewed-by: Roi Dayan <roid@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
Eli Cohen [Thu, 13 Feb 2020 13:18:51 +0000 (15:18 +0200)]
net/mlx5e: Remove unused argument from parse_tc_pedit_action()
parse_attr is not used by parse_tc_pedit_action() so revmove it.
Signed-off-by: Eli Cohen <eli@mellanox.com>
Reviewed-by: Roi Dayan <roid@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
Roi Dayan [Tue, 18 Feb 2020 13:30:58 +0000 (15:30 +0200)]
net/mlx5e: Use NL_SET_ERR_MSG_MOD() extack for errors
This to be consistent and adds the module name to the error message.
Signed-off-by: Roi Dayan <roid@mellanox.com>
Reviewed-by: Eli Cohen <eli@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
Roi Dayan [Tue, 18 Feb 2020 13:24:39 +0000 (15:24 +0200)]
net/mlx5e: Use netdev_warn() instead of pr_err() for errors
This is for added netdev prefix that helps identify
the source of the message.
Signed-off-by: Roi Dayan <roid@mellanox.com>
Reviewed-by: Eli Cohen <eli@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
Roi Dayan [Wed, 25 Dec 2019 15:03:35 +0000 (17:03 +0200)]
net/mlx5e: Use netdev_warn() for errors for added prefix
This helps identify the source of the message.
If netdev still doesn't exists use mlx5_core_warn().
Signed-off-by: Roi Dayan <roid@mellanox.com>
Reviewed-by: Vlad Buslov <vladbu@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
Erez Shitrit [Sun, 12 Jan 2020 08:55:54 +0000 (10:55 +0200)]
net/mlx5: DR, Improve log messages
Few print messages are in debug level where they should be in error, and
few messages are missing.
Signed-off-by: Erez Shitrit <erezsh@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
Hamdan Igbaria [Mon, 17 Feb 2020 13:53:20 +0000 (15:53 +0200)]
net/mlx5: DR, Change matcher priority parameter type
Change matcher priority parameter type from u16 to u32,
this change is needed since sometimes upper levels
create a matcher with priority bigger than 2^16.
Signed-off-by: Hamdan Igbaria <hamdani@mellanox.com>
Reviewed-by: Alex Vesker <valex@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
Jianbo Liu [Fri, 27 Dec 2019 06:37:07 +0000 (06:37 +0000)]
net/mlx5e: Add devlink fdb_large_groups parameter
Add a devlink parameter to control the number of large groups in a
autogrouped flow table. The default value is 15, and the range is between 1
and 1024.
The size of each large group can be calculated according to the following
formula: size = 4M / (fdb_large_groups + 1).
Examples:
- Set the number of large groups to 20.
$ devlink dev param set pci/0000:82:00.0 name fdb_large_groups \
cmode driverinit value 20
Then run devlink reload command to apply the new value.
$ devlink dev reload pci/0000:82:00.0
- Read the number of large groups in flow table.
$ devlink dev param show pci/0000:82:00.0 name fdb_large_groups
pci/0000:82:00.0:
name fdb_large_groups type driver-specific
values:
cmode driverinit value 20
Signed-off-by: Jianbo Liu <jianbol@mellanox.com>
Reviewed-by: Vlad Buslov <vladbu@mellanox.com>
Reviewed-by: Roi Dayan <roid@mellanox.com>
Acked-by: Jiri Pirko <jiri@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
Jianbo Liu [Tue, 7 Jan 2020 08:48:05 +0000 (08:48 +0000)]
net/mlx5: Change the name of steering mode param id
The prefix should be "MLX5_DEVLINK_PARAM_ID_" for all in
mlx5_devlink_param_id enum.
Signed-off-by: Jianbo Liu <jianbol@mellanox.com>
Acked-by: Jiri Pirko <jiri@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
Vladyslav Tarasiuk [Wed, 13 Nov 2019 15:19:47 +0000 (17:19 +0200)]
net/mlx5e: Add support for devlink-port in non-representors mode
Added devlink_port field to mlx5e_priv structure and a callback to
netdev ops to enable devlink to get info about the port. The port
registration happens at driver initialization.
Signed-off-by: Vladyslav Tarasiuk <vladyslavt@mellanox.com>
Reviewed-by: Moshe Shemesh <moshe@mellanox.com>
Reviewed-by: Jiri Pirko <jiri@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
Vladyslav Tarasiuk [Fri, 17 Jan 2020 10:42:53 +0000 (12:42 +0200)]
net/mlx5e: Rename representor get devlink port function
Rename representor's mlx5e_get_devlink_port() to
mlx5e_rep_get_devlink_port().
The downstream patch will add a non-representor mlx5e function called
mlx5e_get_devlink_phy_port().
Signed-off-by: Vladyslav Tarasiuk <vladyslavt@mellanox.com>
Reviewed-by: Moshe Shemesh <moshe@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
Roi Dayan [Thu, 6 Feb 2020 14:06:58 +0000 (16:06 +0200)]
net/mlx5: E-Switch, Allow goto earlier chain if FW supports it
Mellanox FW can support this if ignore_flow_level capability exists.
Signed-off-by: Roi Dayan <roid@mellanox.com>
Reviewed-by: Oz Shlomo <ozsh@mellanox.com>
Reviewed-by: Paul Blakey <paulb@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
Eli Cohen [Tue, 14 Jan 2020 15:30:41 +0000 (17:30 +0200)]
net/mlx5e: Eswitch, Use per vport tables for mirroring
When using port mirroring, we forward the traffic to another table and
use that table to forward to the mirrored vport. Since the hardware
loses the values of reg c, and in particular reg c0, we fail the match
on the input vport which previously existed in reg c0. To overcome this
situation, we use a set of per vport tables, positioned at the lowest
priority, and forward traffic to those tables. Since these tables are
per vport, we can avoid matching on reg c0.
Fixes:
c01cfd0f1115 ("net/mlx5: E-Switch, Add match on vport metadata for rule in fast path")
Signed-off-by: Eli Cohen <eli@mellanox.com>
Reviewed-by: Mark Bloch <markb@mellanox.com>
Reviewed-by: Paul Blakey <paulb@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
Eli Cohen [Thu, 6 Feb 2020 13:13:36 +0000 (15:13 +0200)]
net/mlx5: Eswitch, avoid redundant mask
misc_params.source_port is a 16 bit field already so no need for
redundant masking against 0xffff. Also change local variables type to
u16.
Signed-off-by: Eli Cohen <eli@mellanox.com>
Reviewed-by: Roi Dayan <roid@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
Tonghao Zhang [Mon, 17 Feb 2020 14:08:50 +0000 (22:08 +0800)]
net/mlx5e: Don't allow forwarding between uplink
We can install forwarding packets rule between uplink
in switchdev mode, as show below. But the hardware does
not do that as expected (mlnx_perf -i $PF1, we can't get
the counter of the PF1). By the way, if we add the uplink
PF0, PF1 to Open vSwitch and enable hw-offload, the rules
can be offloaded but not work fine too. This patch add a
check and if so return -EOPNOTSUPP.
$ tc filter add dev $PF0 protocol all parent ffff: prio 1 handle 1 \
flower skip_sw action mirred egress redirect dev $PF1
$ tc -d -s filter show dev $PF0 ingress
skip_sw
in_hw in_hw_count 1
action order 1: mirred (Egress Redirect to device enp130s0f1) stolen
...
Sent hardware 408954 bytes 4173 pkt
...
Signed-off-by: Tonghao Zhang <xiangxia.m.yue@gmail.com>
Reviewed-by: Roi Dayan <roid@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
Linus Torvalds [Fri, 28 Feb 2020 00:34:41 +0000 (16:34 -0800)]
Merge git://git./linux/kernel/git/netdev/net
Pull networking fixes from David Miller:
1) Fix leak in nl80211 AP start where we leak the ACL memory, from
Johannes Berg.
2) Fix double mutex unlock in mac80211, from Andrei Otcheretianski.
3) Fix RCU stall in ipset, from Jozsef Kadlecsik.
4) Fix devlink locking in devlink_dpipe_table_register, from Madhuparna
Bhowmik.
5) Fix race causing TX hang in ll_temac, from Esben Haabendal.
6) Stale eth hdr pointer in br_dev_xmit(), from Nikolay Aleksandrov.
7) Fix TX hash calculation bounds checking wrt. tc rules, from Amritha
Nambiar.
8) Size netlink responses properly in schedule action code to take into
consideration TCA_ACT_FLAGS. From Jiri Pirko.
9) Fix firmware paths for mscc PHY driver, from Antoine Tenart.
10) Don't register stmmac notifier multiple times, from Aaro Koskinen.
11) Various rmnet bug fixes, from Taehee Yoo.
12) Fix vsock deadlock in vsock transport release, from Stefano
Garzarella.
* git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net: (61 commits)
net: dsa: mv88e6xxx: Fix masking of egress port
mlxsw: pci: Wait longer before accessing the device after reset
sfc: fix timestamp reconstruction at 16-bit rollover points
vsock: fix potential deadlock in transport->release()
unix: It's CONFIG_PROC_FS not CONFIG_PROCFS
net: rmnet: fix packet forwarding in rmnet bridge mode
net: rmnet: fix bridge mode bugs
net: rmnet: use upper/lower device infrastructure
net: rmnet: do not allow to change mux id if mux id is duplicated
net: rmnet: remove rcu_read_lock in rmnet_force_unassociate_device()
net: rmnet: fix suspicious RCU usage
net: rmnet: fix NULL pointer dereference in rmnet_changelink()
net: rmnet: fix NULL pointer dereference in rmnet_newlink()
net: phy: marvell: don't interpret PHY status unless resolved
mlx5: register lag notifier for init network namespace only
unix: define and set show_fdinfo only if procfs is enabled
hinic: fix a bug of rss configuration
hinic: fix a bug of setting hw_ioctxt
hinic: fix a irq affinity bug
net/smc: check for valid ib_client_data
...
Gustavo A. R. Silva [Thu, 27 Feb 2020 00:17:44 +0000 (18:17 -0600)]
bpf: Replace zero-length array with flexible-array member
The current codebase makes use of the zero-length array language
extension to the C90 standard, but the preferred mechanism to declare
variable-length types such as these ones is a flexible array member[1][2],
introduced in C99:
struct foo {
int stuff;
struct boo array[];
};
By making use of the mechanism above, we will get a compiler warning
in case the flexible array does not occur last in the structure, which
will help us prevent some kind of undefined behavior bugs from being
inadvertently introduced[3] to the codebase from now on.
Also, notice that, dynamic memory allocations won't be affected by
this change:
"Flexible array members have incomplete type, and so the sizeof operator
may not be applied. As a quirk of the original implementation of
zero-length arrays, sizeof evaluates to zero."[1]
This issue was found with the help of Coccinelle.
[1] https://gcc.gnu.org/onlinedocs/gcc/Zero-Length.html
[2] https://github.com/KSPP/linux/issues/21
[3] commit
76497732932f ("cxgb3/l2t: Fix undefined behaviour")
Signed-off-by: Gustavo A. R. Silva <gustavo@embeddedor.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Song Liu <songliubraving@fb.com>
Link: https://lore.kernel.org/bpf/20200227001744.GA3317@embeddedor
Andrew Lunn [Thu, 27 Feb 2020 20:20:49 +0000 (21:20 +0100)]
net: dsa: mv88e6xxx: Fix masking of egress port
Add missing ~ to the usage of the mask.
Reported-by: Kevin Benson <Kevin.Benson@zii.aero>
Reported-by: Chris Healy <Chris.Healy@zii.aero>
Fixes:
5c74c54ce6ff ("net: dsa: mv88e6xxx: Split monitor port configuration")
Signed-off-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: David S. Miller <davem@davemloft.net>
Amit Cohen [Thu, 27 Feb 2020 20:07:53 +0000 (21:07 +0100)]
mlxsw: pci: Wait longer before accessing the device after reset
During initialization the driver issues a reset to the device and waits
for 100ms before checking if the firmware is ready. The waiting is
necessary because before that the device is irresponsive and the first
read can result in a completion timeout.
While 100ms is sufficient for Spectrum-1 and Spectrum-2, it is
insufficient for Spectrum-3.
Fix this by increasing the timeout to 200ms.
Fixes:
da382875c616 ("mlxsw: spectrum: Extend to support Spectrum-3 ASIC")
Signed-off-by: Amit Cohen <amitc@mellanox.com>
Signed-off-by: Ido Schimmel <idosch@mellanox.com>
Signed-off-by: Jiri Pirko <jiri@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Jiri Pirko [Thu, 27 Feb 2020 19:59:26 +0000 (20:59 +0100)]
mlxsw: reg: Update module_type values in PMTM register and map them to width
There are couple new values that PMTM register can return
in module_type field. Add them and map them to module width in
mlxsw_core_module_max_width(). Fix the existing names on the way.
Signed-off-by: Jiri Pirko <jiri@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Gustavo A. R. Silva [Wed, 26 Feb 2020 21:23:17 +0000 (15:23 -0600)]
WAN: Replace zero-length array with flexible-array member
The current codebase makes use of the zero-length array language
extension to the C90 standard, but the preferred mechanism to declare
variable-length types such as these ones is a flexible array member[1][2],
introduced in C99:
struct foo {
int stuff;
struct boo array[];
};
By making use of the mechanism above, we will get a compiler warning
in case the flexible array does not occur last in the structure, which
will help us prevent some kind of undefined behavior bugs from being
inadvertently introduced[3] to the codebase from now on.
Also, notice that, dynamic memory allocations won't be affected by
this change:
"Flexible array members have incomplete type, and so the sizeof operator
may not be applied. As a quirk of the original implementation of
zero-length arrays, sizeof evaluates to zero."[1]
This issue was found with the help of Coccinelle.
[1] https://gcc.gnu.org/onlinedocs/gcc/Zero-Length.html
[2] https://github.com/KSPP/linux/issues/21
[3] commit
76497732932f ("cxgb3/l2t: Fix undefined behaviour")
Signed-off-by: Gustavo A. R. Silva <gustavo@embeddedor.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Gustavo A. R. Silva [Wed, 26 Feb 2020 21:20:00 +0000 (15:20 -0600)]
NFC: Replace zero-length array with flexible-array member
The current codebase makes use of the zero-length array language
extension to the C90 standard, but the preferred mechanism to declare
variable-length types such as these ones is a flexible array member[1][2],
introduced in C99:
struct foo {
int stuff;
struct boo array[];
};
By making use of the mechanism above, we will get a compiler warning
in case the flexible array does not occur last in the structure, which
will help us prevent some kind of undefined behavior bugs from being
inadvertently introduced[3] to the codebase from now on.
Also, notice that, dynamic memory allocations won't be affected by
this change:
"Flexible array members have incomplete type, and so the sizeof operator
may not be applied. As a quirk of the original implementation of
zero-length arrays, sizeof evaluates to zero."[1]
This issue was found with the help of Coccinelle.
[1] https://gcc.gnu.org/onlinedocs/gcc/Zero-Length.html
[2] https://github.com/KSPP/linux/issues/21
[3] commit
76497732932f ("cxgb3/l2t: Fix undefined behaviour")
Signed-off-by: Gustavo A. R. Silva <gustavo@embeddedor.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Alex Maftei (amaftei) [Wed, 26 Feb 2020 17:33:19 +0000 (17:33 +0000)]
sfc: fix timestamp reconstruction at 16-bit rollover points
We can't just use the top bits of the last sync event as they could be
off-by-one every 65,536 seconds, giving an error in reconstruction of
65,536 seconds.
This patch uses the difference in the bottom 16 bits (mod 2^16) to
calculate an offset that needs to be applied to the last sync event to
get to the current time.
Signed-off-by: Alexandru-Mihai Maftei <amaftei@solarflare.com>
Acked-by: Martin Habets <mhabets@solarflare.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Stefano Garzarella [Wed, 26 Feb 2020 10:58:18 +0000 (11:58 +0100)]
vsock: fix potential deadlock in transport->release()
Some transports (hyperv, virtio) acquire the sock lock during the
.release() callback.
In the vsock_stream_connect() we call vsock_assign_transport(); if
the socket was previously assigned to another transport, the
vsk->transport->release() is called, but the sock lock is already
held in the vsock_stream_connect(), causing a deadlock reported by
syzbot:
INFO: task syz-executor280:9768 blocked for more than 143 seconds.
Not tainted 5.6.0-rc1-syzkaller #0
"echo 0 > /proc/sys/kernel/hung_task_timeout_secs" disables this message.
syz-executor280 D27912 9768 9766 0x00000000
Call Trace:
context_switch kernel/sched/core.c:3386 [inline]
__schedule+0x934/0x1f90 kernel/sched/core.c:4082
schedule+0xdc/0x2b0 kernel/sched/core.c:4156
__lock_sock+0x165/0x290 net/core/sock.c:2413
lock_sock_nested+0xfe/0x120 net/core/sock.c:2938
virtio_transport_release+0xc4/0xd60 net/vmw_vsock/virtio_transport_common.c:832
vsock_assign_transport+0xf3/0x3b0 net/vmw_vsock/af_vsock.c:454
vsock_stream_connect+0x2b3/0xc70 net/vmw_vsock/af_vsock.c:1288
__sys_connect_file+0x161/0x1c0 net/socket.c:1857
__sys_connect+0x174/0x1b0 net/socket.c:1874
__do_sys_connect net/socket.c:1885 [inline]
__se_sys_connect net/socket.c:1882 [inline]
__x64_sys_connect+0x73/0xb0 net/socket.c:1882
do_syscall_64+0xfa/0x790 arch/x86/entry/common.c:294
entry_SYSCALL_64_after_hwframe+0x49/0xbe
To avoid this issue, this patch remove the lock acquiring in the
.release() callback of hyperv and virtio transports, and it holds
the lock when we call vsk->transport->release() in the vsock core.
Reported-by: syzbot+731710996d79d0d58fbc@syzkaller.appspotmail.com
Fixes:
408624af4c89 ("vsock: use local transport when it is loaded")
Signed-off-by: Stefano Garzarella <sgarzare@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
David S. Miller [Thu, 27 Feb 2020 20:02:14 +0000 (12:02 -0800)]
Merge branch 'rework-phylink-interface-for-split-MAC-PCS-support'
Russell King says:
====================
rework phylink interface for split MAC/PCS support
The following series changes the phylink interface to allow us to
better support split MAC / MAC PCS setups. The fundamental change
required for this turns out to be quite simple.
Today, mac_config() is used for everything to do with setting the
parameters for the MAC, and mac_link_up() is used to inform the
MAC driver that the link is now up (and so to allow packet flow.)
mac_config() also has had a few implementation issues, with folk
who believe that members such as "speed" and "duplex" are always
valid, where "link" gets used inappropriately, etc.
With the proposed patches, all this changes subtly - but in a
backwards compatible way at this stage.
We pass the the full resolved link state (speed, duplex, pause) to
mac_link_up(), and it is now guaranteed that these parameters to
this function will always be valid (no more SPEED_UNKNOWN or
DUPLEX_UNKNOWN here - unless phylink is fed with such things.)
Drivers should convert over to using the state in mac_link_up()
rather than configuring the speed, duplex and pause in the
mac_config() method. The patch series includes a number of MAC
drivers which I've thought have been easy targets - I've left the
remainder as I think they need maintainer input. However, *all*
drivers will need conversion for future phylink development.
v2: add ocelot/felix and qca/ar9331 DSA drivers to patch 2, add
received tested-by so far.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
Russell King [Wed, 26 Feb 2020 10:24:17 +0000 (10:24 +0000)]
net: mvpp2: use resolved link config in mac_link_up()
Convert the Marvell mvpp2 ethernet driver to use the finalised link
parameters in mac_link_up() rather than the parameters in mac_config().
Signed-off-by: Russell King <rmk+kernel@armlinux.org.uk>
Signed-off-by: David S. Miller <davem@davemloft.net>
Russell King [Wed, 26 Feb 2020 10:24:12 +0000 (10:24 +0000)]
net: mvneta: use resolved link config in mac_link_up()
Convert the Marvell mvneta ethernet driver to use the finalised link
parameters in mac_link_up() rather than the parameters in mac_config().
Signed-off-by: Russell King <rmk+kernel@armlinux.org.uk>
Signed-off-by: David S. Miller <davem@davemloft.net>
Russell King [Wed, 26 Feb 2020 10:24:06 +0000 (10:24 +0000)]
net: macb: use resolved link config in mac_link_up()
Convert the macb ethernet driver to use the finalised link
parameters in mac_link_up() rather than the parameters in mac_config().
Tested-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Signed-off-by: Russell King <rmk+kernel@armlinux.org.uk>
Signed-off-by: David S. Miller <davem@davemloft.net>
Russell King [Wed, 26 Feb 2020 10:24:01 +0000 (10:24 +0000)]
net: dpaa2-mac: use resolved link config in mac_link_up()
Convert the DPAA2 ethernet driver to use the finalised link parameters
in mac_link_up() rather than the parameters in mac_config(), which are
more suited to the needs of the DPAA2 MC firmware than those available
via mac_config().
Tested-by: Ioana Ciornei <ioana.ciornei@nxp.com>
Signed-off-by: Russell King <rmk+kernel@armlinux.org.uk>
Signed-off-by: David S. Miller <davem@davemloft.net>
Russell King [Wed, 26 Feb 2020 10:23:56 +0000 (10:23 +0000)]
net: axienet: use resolved link config in mac_link_up()
Convert the Xilinx AXI ethernet driver to use the finalised link
parameters in mac_link_up() rather than the parameters in mac_config().
Tested-by: Andre Przywara <andre.przywara@arm.com>
Signed-off-by: Russell King <rmk+kernel@armlinux.org.uk>
Signed-off-by: David S. Miller <davem@davemloft.net>
Russell King [Wed, 26 Feb 2020 10:23:51 +0000 (10:23 +0000)]
net: mv88e6xxx: use resolved link config in mac_link_up()
Use the resolved link configuration to set the MAC configuration when
mac_link_up() for non-internal-PHY ports.
Signed-off-by: Russell King <rmk+kernel@armlinux.org.uk>
Signed-off-by: David S. Miller <davem@davemloft.net>
Russell King [Wed, 26 Feb 2020 10:23:46 +0000 (10:23 +0000)]
net: dsa: propagate resolved link config via mac_link_up()
Propagate the resolved link configuration down via DSA's
phylink_mac_link_up() operation to allow split PCS/MAC to work.
Tested-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Signed-off-by: Russell King <rmk+kernel@armlinux.org.uk>
Signed-off-by: David S. Miller <davem@davemloft.net>
Russell King [Wed, 26 Feb 2020 10:23:41 +0000 (10:23 +0000)]
net: phylink: propagate resolved link config via mac_link_up()
Propagate the resolved link parameters via the mac_link_up() call for
MACs that do not automatically track their PCS state. We propagate the
link parameters via function arguments so that inappropriate members
of struct phylink_link_state can't be accessed, and creating a new
structure just for this adds needless complexity to the API.
Tested-by: Andre Przywara <andre.przywara@arm.com>
Tested-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Tested-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Signed-off-by: Russell King <rmk+kernel@armlinux.org.uk>
Signed-off-by: David S. Miller <davem@davemloft.net>
David S. Miller [Thu, 27 Feb 2020 19:52:35 +0000 (11:52 -0800)]
unix: It's CONFIG_PROC_FS not CONFIG_PROCFS
Fixes:
3a12500ed5dd ("unix: define and set show_fdinfo only if procfs is enabled")
Signed-off-by: David S. Miller <davem@davemloft.net>
David S. Miller [Thu, 27 Feb 2020 19:45:07 +0000 (11:45 -0800)]
Merge branch 'net-rmnet-fix-several-bugs'
Taehee Yoo says:
====================
net: rmnet: fix several bugs
This patchset is to fix several bugs in RMNET module.
1. The first patch fixes NULL-ptr-deref in rmnet_newlink().
When rmnet interface is being created, it uses IFLA_LINK
without checking NULL.
So, if userspace doesn't set IFLA_LINK, panic will occur.
In this patch, checking NULL pointer code is added.
2. The second patch fixes NULL-ptr-deref in rmnet_changelink().
To get real device in rmnet_changelink(), it uses IFLA_LINK.
But, IFLA_LINK should not be used in rmnet_changelink().
3. The third patch fixes suspicious RCU usage in rmnet_get_port().
rmnet_get_port() uses rcu_dereference_rtnl().
But, rmnet_get_port() is used by datapath.
So, rcu_dereference_bh() should be used instead of rcu_dereference_rtnl().
4. The fourth patch fixes suspicious RCU usage in
rmnet_force_unassociate_device().
RCU critical section should not be scheduled.
But, unregister_netdevice_queue() in the rmnet_force_unassociate_device()
would be scheduled.
So, the RCU warning occurs.
In this patch, the rcu_read_lock() in the rmnet_force_unassociate_device()
is removed because it's unnecessary.
5. The fifth patch fixes duplicate MUX ID case.
RMNET MUX ID is unique.
So, rmnet interface isn't allowed to be created, which have
a duplicate MUX ID.
But, only rmnet_newlink() checks this condition, rmnet_changelink()
doesn't check this.
So, duplicate MUX ID case would happen.
6. The sixth patch fixes upper/lower interface relationship problems.
When IFLA_LINK is used, the upper/lower infrastructure should be used.
Because it checks the maximum depth of upper/lower interfaces and it also
checks circular interface relationship, etc.
In this patch, netdev_upper_dev_link() is used.
7. The seventh patch fixes bridge related problems.
a) ->ndo_del_slave() doesn't work.
b) It couldn't detect circular upper/lower interface relationship.
c) It couldn't prevent stack overflow because of too deep depth
of upper/lower interface
d) It doesn't check the number of lower interfaces.
e) Panics because of several reasons.
These problems are actually the same problem.
So, this patch fixes these problems.
8. The eighth patch fixes packet forwarding issue in bridge mode
Packet forwarding is not working in rmnet bridge mode.
Because when a packet is forwarded, skb_push() for an ethernet header
is needed. But it doesn't call skb_push().
So, the ethernet header will be lost.
Change log:
- update commit logs.
- drop two patches in this patchset because of wrong target branch.
- ("net: rmnet: add missing module alias")
- ("net: rmnet: print error message when command fails")
- remove unneessary rcu_read_lock() in the third patch.
- use rcu_dereference_bh() instead of rcu_dereference in third patch.
- do not allow to add a bridge device if rmnet interface is already
bridge mode in the seventh patch.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
Taehee Yoo [Thu, 27 Feb 2020 12:26:15 +0000 (12:26 +0000)]
net: rmnet: fix packet forwarding in rmnet bridge mode
Packet forwarding is not working in rmnet bridge mode.
Because when a packet is forwarded, skb_push() for an ethernet header
is needed. But it doesn't call skb_push().
So, the ethernet header will be lost.
Test commands:
modprobe rmnet
ip netns add nst
ip netns add nst2
ip link add veth0 type veth peer name veth1
ip link add veth2 type veth peer name veth3
ip link set veth1 netns nst
ip link set veth3 netns nst2
ip link add rmnet0 link veth0 type rmnet mux_id 1
ip link set veth2 master rmnet0
ip link set veth0 up
ip link set veth2 up
ip link set rmnet0 up
ip a a 192.168.100.1/24 dev rmnet0
ip netns exec nst ip link set veth1 up
ip netns exec nst ip a a 192.168.100.2/24 dev veth1
ip netns exec nst2 ip link set veth3 up
ip netns exec nst2 ip a a 192.168.100.3/24 dev veth3
ip netns exec nst2 ping 192.168.100.2
Fixes:
60d58f971c10 ("net: qualcomm: rmnet: Implement bridge mode")
Signed-off-by: Taehee Yoo <ap420073@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Taehee Yoo [Thu, 27 Feb 2020 12:26:02 +0000 (12:26 +0000)]
net: rmnet: fix bridge mode bugs
In order to attach a bridge interface to the rmnet interface,
"master" operation is used.
(e.g. ip link set dummy1 master rmnet0)
But, in the rmnet_add_bridge(), which is a callback of ->ndo_add_slave()
doesn't register lower interface.
So, ->ndo_del_slave() doesn't work.
There are other problems too.
1. It couldn't detect circular upper/lower interface relationship.
2. It couldn't prevent stack overflow because of too deep depth
of upper/lower interface
3. It doesn't check the number of lower interfaces.
4. Panics because of several reasons.
The root problem of these issues is actually the same.
So, in this patch, these all problems will be fixed.
Test commands:
modprobe rmnet
ip link add dummy0 type dummy
ip link add rmnet0 link dummy0 type rmnet mux_id 1
ip link add dummy1 master rmnet0 type dummy
ip link add dummy2 master rmnet0 type dummy
ip link del rmnet0
ip link del dummy2
ip link del dummy1
Splat looks like:
[ 41.867595][ T1164] general protection fault, probably for non-canonical address 0xdffffc0000000101I
[ 41.869993][ T1164] KASAN: null-ptr-deref in range [0x0000000000000808-0x000000000000080f]
[ 41.872950][ T1164] CPU: 0 PID: 1164 Comm: ip Not tainted 5.6.0-rc1+ #447
[ 41.873915][ T1164] Hardware name: innotek GmbH VirtualBox/VirtualBox, BIOS VirtualBox 12/01/2006
[ 41.875161][ T1164] RIP: 0010:rmnet_unregister_bridge.isra.6+0x71/0xf0 [rmnet]
[ 41.876178][ T1164] Code: 48 89 ef 48 89 c6 5b 5d e9 fc fe ff ff e8 f7 f3 ff ff 48 8d b8 08 08 00 00 48 ba 00 7
[ 41.878925][ T1164] RSP: 0018:
ffff8880c4d0f188 EFLAGS:
00010202
[ 41.879774][ T1164] RAX:
0000000000000000 RBX:
0000000000000000 RCX:
0000000000000101
[ 41.887689][ T1164] RDX:
dffffc0000000000 RSI:
ffffffffb8cf64f0 RDI:
0000000000000808
[ 41.888727][ T1164] RBP:
ffff8880c40e4000 R08:
ffffed101b3c0e3c R09:
0000000000000001
[ 41.889749][ T1164] R10:
0000000000000001 R11:
ffffed101b3c0e3b R12:
1ffff110189a1e3c
[ 41.890783][ T1164] R13:
ffff8880c4d0f200 R14:
ffffffffb8d56160 R15:
ffff8880ccc2c000
[ 41.891794][ T1164] FS:
00007f4300edc0c0(0000) GS:
ffff8880d9c00000(0000) knlGS:
0000000000000000
[ 41.892953][ T1164] CS: 0010 DS: 0000 ES: 0000 CR0:
0000000080050033
[ 41.893800][ T1164] CR2:
00007f43003bc8c0 CR3:
00000000ca53e001 CR4:
00000000000606f0
[ 41.894824][ T1164] Call Trace:
[ 41.895274][ T1164] ? rcu_is_watching+0x2c/0x80
[ 41.895895][ T1164] rmnet_config_notify_cb+0x1f7/0x590 [rmnet]
[ 41.896687][ T1164] ? rmnet_unregister_bridge.isra.6+0xf0/0xf0 [rmnet]
[ 41.897611][ T1164] ? rmnet_unregister_bridge.isra.6+0xf0/0xf0 [rmnet]
[ 41.898508][ T1164] ? __module_text_address+0x13/0x140
[ 41.899162][ T1164] notifier_call_chain+0x90/0x160
[ 41.899814][ T1164] rollback_registered_many+0x660/0xcf0
[ 41.900544][ T1164] ? netif_set_real_num_tx_queues+0x780/0x780
[ 41.901316][ T1164] ? __lock_acquire+0xdfe/0x3de0
[ 41.901958][ T1164] ? memset+0x1f/0x40
[ 41.902468][ T1164] ? __nla_validate_parse+0x98/0x1ab0
[ 41.903166][ T1164] unregister_netdevice_many.part.133+0x13/0x1b0
[ 41.903988][ T1164] rtnl_delete_link+0xbc/0x100
[ ... ]
Fixes:
60d58f971c10 ("net: qualcomm: rmnet: Implement bridge mode")
Signed-off-by: Taehee Yoo <ap420073@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Taehee Yoo [Thu, 27 Feb 2020 12:25:43 +0000 (12:25 +0000)]
net: rmnet: use upper/lower device infrastructure
netdev_upper_dev_link() is useful to manage lower/upper interfaces.
And this function internally validates looping, maximum depth.
All or most virtual interfaces that could have a real interface
(e.g. macsec, macvlan, ipvlan etc.) use lower/upper infrastructure.
Test commands:
modprobe rmnet
ip link add dummy0 type dummy
ip link add rmnet1 link dummy0 type rmnet mux_id 1
for i in {2..100}
do
let A=$i-1
ip link add rmnet$i link rmnet$A type rmnet mux_id $i
done
ip link del dummy0
The purpose of the test commands is to make stack overflow.
Splat looks like:
[ 52.411438][ T1395] BUG: KASAN: slab-out-of-bounds in find_busiest_group+0x27e/0x2c00
[ 52.413218][ T1395] Write of size 64 at addr
ffff8880c774bde0 by task ip/1395
[ 52.414841][ T1395]
[ 52.430720][ T1395] CPU: 1 PID: 1395 Comm: ip Not tainted 5.6.0-rc1+ #447
[ 52.496511][ T1395] Hardware name: innotek GmbH VirtualBox/VirtualBox, BIOS VirtualBox 12/01/2006
[ 52.513597][ T1395] Call Trace:
[ 52.546516][ T1395]
[ 52.558773][ T1395] Allocated by task
3171537984:
[ 52.588290][ T1395] BUG: unable to handle page fault for address:
ffffffffb999e260
[ 52.589311][ T1395] #PF: supervisor read access in kernel mode
[ 52.590529][ T1395] #PF: error_code(0x0000) - not-present page
[ 52.591374][ T1395] PGD
d6818067 P4D
d6818067 PUD
d6819063 PMD 0
[ 52.592288][ T1395] Thread overran stack, or stack corrupted
[ 52.604980][ T1395] Oops: 0000 [#1] SMP DEBUG_PAGEALLOC KASAN PTI
[ 52.605856][ T1395] CPU: 1 PID: 1395 Comm: ip Not tainted 5.6.0-rc1+ #447
[ 52.611764][ T1395] Hardware name: innotek GmbH VirtualBox/VirtualBox, BIOS VirtualBox 12/01/2006
[ 52.621520][ T1395] RIP: 0010:stack_depot_fetch+0x10/0x30
[ 52.622296][ T1395] Code: ff e9 f9 fe ff ff 48 89 df e8 9c 1d 91 ff e9 ca fe ff ff cc cc cc cc cc cc cc 89 f8 0
[ 52.627887][ T1395] RSP: 0018:
ffff8880c774bb60 EFLAGS:
00010006
[ 52.628735][ T1395] RAX:
00000000001f8880 RBX:
ffff8880c774d140 RCX:
0000000000000000
[ 52.631773][ T1395] RDX:
000000000000001d RSI:
ffff8880c774bb68 RDI:
0000000000003ff0
[ 52.649584][ T1395] RBP:
ffffea00031dd200 R08:
ffffed101b43e403 R09:
ffffed101b43e403
[ 52.674857][ T1395] R10:
0000000000000001 R11:
ffffed101b43e402 R12:
ffff8880d900e5c0
[ 52.678257][ T1395] R13:
ffff8880c774c000 R14:
0000000000000000 R15:
dffffc0000000000
[ 52.694541][ T1395] FS:
00007fe867f6e0c0(0000) GS:
ffff8880da000000(0000) knlGS:
0000000000000000
[ 52.764039][ T1395] CS: 0010 DS: 0000 ES: 0000 CR0:
0000000080050033
[ 52.815008][ T1395] CR2:
ffffffffb999e260 CR3:
00000000c26aa005 CR4:
00000000000606e0
[ 52.862312][ T1395] Call Trace:
[ 52.887133][ T1395] Modules linked in: dummy rmnet veth openvswitch nsh nf_conncount nf_nat nf_conntrack nf_dex
[ 52.936749][ T1395] CR2:
ffffffffb999e260
[ 52.965695][ T1395] ---[ end trace
7e32ca99482dbb31 ]---
[ 52.966556][ T1395] RIP: 0010:stack_depot_fetch+0x10/0x30
[ 52.971083][ T1395] Code: ff e9 f9 fe ff ff 48 89 df e8 9c 1d 91 ff e9 ca fe ff ff cc cc cc cc cc cc cc 89 f8 0
[ 53.003650][ T1395] RSP: 0018:
ffff8880c774bb60 EFLAGS:
00010006
[ 53.043183][ T1395] RAX:
00000000001f8880 RBX:
ffff8880c774d140 RCX:
0000000000000000
[ 53.076480][ T1395] RDX:
000000000000001d RSI:
ffff8880c774bb68 RDI:
0000000000003ff0
[ 53.093858][ T1395] RBP:
ffffea00031dd200 R08:
ffffed101b43e403 R09:
ffffed101b43e403
[ 53.112795][ T1395] R10:
0000000000000001 R11:
ffffed101b43e402 R12:
ffff8880d900e5c0
[ 53.139837][ T1395] R13:
ffff8880c774c000 R14:
0000000000000000 R15:
dffffc0000000000
[ 53.141500][ T1395] FS:
00007fe867f6e0c0(0000) GS:
ffff8880da000000(0000) knlGS:
0000000000000000
[ 53.143343][ T1395] CS: 0010 DS: 0000 ES: 0000 CR0:
0000000080050033
[ 53.152007][ T1395] CR2:
ffffffffb999e260 CR3:
00000000c26aa005 CR4:
00000000000606e0
[ 53.156459][ T1395] Kernel panic - not syncing: Fatal exception
[ 54.213570][ T1395] Shutting down cpus with NMI
[ 54.354112][ T1395] Kernel Offset: 0x33000000 from 0xffffffff81000000 (relocation range: 0xffffffff80000000-0x)
[ 54.355687][ T1395] Rebooting in 5 seconds..
Fixes:
b37f78f234bf ("net: qualcomm: rmnet: Fix crash on real dev unregistration")
Signed-off-by: Taehee Yoo <ap420073@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Taehee Yoo [Thu, 27 Feb 2020 12:25:19 +0000 (12:25 +0000)]
net: rmnet: do not allow to change mux id if mux id is duplicated
Basically, duplicate mux id isn't be allowed.
So, the creation of rmnet will be failed if there is duplicate mux id
is existing.
But, changelink routine doesn't check duplicate mux id.
Test commands:
modprobe rmnet
ip link add dummy0 type dummy
ip link add rmnet0 link dummy0 type rmnet mux_id 1
ip link add rmnet1 link dummy0 type rmnet mux_id 2
ip link set rmnet1 type rmnet mux_id 1
Fixes:
23790ef12082 ("net: qualcomm: rmnet: Allow to configure flags for existing devices")
Signed-off-by: Taehee Yoo <ap420073@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Taehee Yoo [Thu, 27 Feb 2020 12:25:05 +0000 (12:25 +0000)]
net: rmnet: remove rcu_read_lock in rmnet_force_unassociate_device()
The notifier_call() of the slave interface removes rmnet interface with
unregister_netdevice_queue().
But, before calling unregister_netdevice_queue(), it acquires
rcu readlock.
In the RCU critical section, sleeping isn't be allowed.
But, unregister_netdevice_queue() internally calls synchronize_net(),
which would sleep.
So, suspicious RCU usage warning occurs.
Test commands:
modprobe rmnet
ip link add dummy0 type dummy
ip link add dummy1 type dummy
ip link add rmnet0 link dummy0 type rmnet mux_id 1
ip link set dummy1 master rmnet0
ip link del dummy0
Splat looks like:
[ 79.639245][ T1195] =============================
[ 79.640134][ T1195] WARNING: suspicious RCU usage
[ 79.640852][ T1195] 5.6.0-rc1+ #447 Not tainted
[ 79.641657][ T1195] -----------------------------
[ 79.642472][ T1195] ./include/linux/rcupdate.h:273 Illegal context switch in RCU read-side critical section!
[ 79.644043][ T1195]
[ 79.644043][ T1195] other info that might help us debug this:
[ 79.644043][ T1195]
[ 79.645682][ T1195]
[ 79.645682][ T1195] rcu_scheduler_active = 2, debug_locks = 1
[ 79.646980][ T1195] 2 locks held by ip/1195:
[ 79.647629][ T1195] #0:
ffffffffa3cf64f0 (rtnl_mutex){+.+.}, at: rtnetlink_rcv_msg+0x457/0x890
[ 79.649312][ T1195] #1:
ffffffffa39256c0 (rcu_read_lock){....}, at: rmnet_config_notify_cb+0xf0/0x590 [rmnet]
[ 79.651717][ T1195]
[ 79.651717][ T1195] stack backtrace:
[ 79.652650][ T1195] CPU: 3 PID: 1195 Comm: ip Not tainted 5.6.0-rc1+ #447
[ 79.653702][ T1195] Hardware name: innotek GmbH VirtualBox/VirtualBox, BIOS VirtualBox 12/01/2006
[ 79.655037][ T1195] Call Trace:
[ 79.655560][ T1195] dump_stack+0x96/0xdb
[ 79.656252][ T1195] ___might_sleep+0x345/0x440
[ 79.656994][ T1195] synchronize_net+0x18/0x30
[ 79.661132][ T1195] netdev_rx_handler_unregister+0x40/0xb0
[ 79.666266][ T1195] rmnet_unregister_real_device+0x42/0xb0 [rmnet]
[ 79.667211][ T1195] rmnet_config_notify_cb+0x1f7/0x590 [rmnet]
[ 79.668121][ T1195] ? rmnet_unregister_bridge.isra.6+0xf0/0xf0 [rmnet]
[ 79.669166][ T1195] ? rmnet_unregister_bridge.isra.6+0xf0/0xf0 [rmnet]
[ 79.670286][ T1195] ? __module_text_address+0x13/0x140
[ 79.671139][ T1195] notifier_call_chain+0x90/0x160
[ 79.671973][ T1195] rollback_registered_many+0x660/0xcf0
[ 79.672893][ T1195] ? netif_set_real_num_tx_queues+0x780/0x780
[ 79.675091][ T1195] ? __lock_acquire+0xdfe/0x3de0
[ 79.675825][ T1195] ? memset+0x1f/0x40
[ 79.676367][ T1195] ? __nla_validate_parse+0x98/0x1ab0
[ 79.677290][ T1195] unregister_netdevice_many.part.133+0x13/0x1b0
[ 79.678163][ T1195] rtnl_delete_link+0xbc/0x100
[ ... ]
Fixes:
ceed73a2cf4a ("drivers: net: ethernet: qualcomm: rmnet: Initial implementation")
Signed-off-by: Taehee Yoo <ap420073@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Taehee Yoo [Thu, 27 Feb 2020 12:24:45 +0000 (12:24 +0000)]
net: rmnet: fix suspicious RCU usage
rmnet_get_port() internally calls rcu_dereference_rtnl(),
which checks RTNL.
But rmnet_get_port() could be called by packet path.
The packet path is not protected by RTNL.
So, the suspicious RCU usage problem occurs.
Test commands:
modprobe rmnet
ip netns add nst
ip link add veth0 type veth peer name veth1
ip link set veth1 netns nst
ip link add rmnet0 link veth0 type rmnet mux_id 1
ip netns exec nst ip link add rmnet1 link veth1 type rmnet mux_id 1
ip netns exec nst ip link set veth1 up
ip netns exec nst ip link set rmnet1 up
ip netns exec nst ip a a 192.168.100.2/24 dev rmnet1
ip link set veth0 up
ip link set rmnet0 up
ip a a 192.168.100.1/24 dev rmnet0
ping 192.168.100.2
Splat looks like:
[ 146.630958][ T1174] WARNING: suspicious RCU usage
[ 146.631735][ T1174] 5.6.0-rc1+ #447 Not tainted
[ 146.632387][ T1174] -----------------------------
[ 146.633151][ T1174] drivers/net/ethernet/qualcomm/rmnet/rmnet_config.c:386 suspicious rcu_dereference_check() !
[ 146.634742][ T1174]
[ 146.634742][ T1174] other info that might help us debug this:
[ 146.634742][ T1174]
[ 146.645992][ T1174]
[ 146.645992][ T1174] rcu_scheduler_active = 2, debug_locks = 1
[ 146.646937][ T1174] 5 locks held by ping/1174:
[ 146.647609][ T1174] #0:
ffff8880c31dea70 (sk_lock-AF_INET){+.+.}, at: raw_sendmsg+0xab8/0x2980
[ 146.662463][ T1174] #1:
ffffffff93925660 (rcu_read_lock_bh){....}, at: ip_finish_output2+0x243/0x2150
[ 146.671696][ T1174] #2:
ffffffff93925660 (rcu_read_lock_bh){....}, at: __dev_queue_xmit+0x213/0x2940
[ 146.673064][ T1174] #3:
ffff8880c19ecd58 (&dev->qdisc_running_key#7){+...}, at: ip_finish_output2+0x714/0x2150
[ 146.690358][ T1174] #4:
ffff8880c5796898 (&dev->qdisc_xmit_lock_key#3){+.-.}, at: sch_direct_xmit+0x1e2/0x1020
[ 146.699875][ T1174]
[ 146.699875][ T1174] stack backtrace:
[ 146.701091][ T1174] CPU: 0 PID: 1174 Comm: ping Not tainted 5.6.0-rc1+ #447
[ 146.705215][ T1174] Hardware name: innotek GmbH VirtualBox/VirtualBox, BIOS VirtualBox 12/01/2006
[ 146.706565][ T1174] Call Trace:
[ 146.707102][ T1174] dump_stack+0x96/0xdb
[ 146.708007][ T1174] rmnet_get_port.part.9+0x76/0x80 [rmnet]
[ 146.709233][ T1174] rmnet_egress_handler+0x107/0x420 [rmnet]
[ 146.710492][ T1174] ? sch_direct_xmit+0x1e2/0x1020
[ 146.716193][ T1174] rmnet_vnd_start_xmit+0x3d/0xa0 [rmnet]
[ 146.717012][ T1174] dev_hard_start_xmit+0x160/0x740
[ 146.717854][ T1174] sch_direct_xmit+0x265/0x1020
[ 146.718577][ T1174] ? register_lock_class+0x14d0/0x14d0
[ 146.719429][ T1174] ? dev_watchdog+0xac0/0xac0
[ 146.723738][ T1174] ? __dev_queue_xmit+0x15fd/0x2940
[ 146.724469][ T1174] ? lock_acquire+0x164/0x3b0
[ 146.725172][ T1174] __dev_queue_xmit+0x20c7/0x2940
[ ... ]
Fixes:
ceed73a2cf4a ("drivers: net: ethernet: qualcomm: rmnet: Initial implementation")
Signed-off-by: Taehee Yoo <ap420073@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Taehee Yoo [Thu, 27 Feb 2020 12:24:26 +0000 (12:24 +0000)]
net: rmnet: fix NULL pointer dereference in rmnet_changelink()
In the rmnet_changelink(), it uses IFLA_LINK without checking
NULL pointer.
tb[IFLA_LINK] could be NULL pointer.
So, NULL-ptr-deref could occur.
rmnet already has a lower interface (real_dev).
So, after this patch, rmnet_changelink() does not use IFLA_LINK anymore.
Test commands:
modprobe rmnet
ip link add dummy0 type dummy
ip link add rmnet0 link dummy0 type rmnet mux_id 1
ip link set rmnet0 type rmnet mux_id 2
Splat looks like:
[ 90.578726][ T1131] general protection fault, probably for non-canonical address 0xdffffc0000000000I
[ 90.581121][ T1131] KASAN: null-ptr-deref in range [0x0000000000000000-0x0000000000000007]
[ 90.582380][ T1131] CPU: 2 PID: 1131 Comm: ip Not tainted 5.6.0-rc1+ #447
[ 90.584285][ T1131] Hardware name: innotek GmbH VirtualBox/VirtualBox, BIOS VirtualBox 12/01/2006
[ 90.587506][ T1131] RIP: 0010:rmnet_changelink+0x5a/0x8a0 [rmnet]
[ 90.588546][ T1131] Code: 83 ec 20 48 c1 ea 03 80 3c 02 00 0f 85 6f 07 00 00 48 8b 5e 28 48 b8 00 00 00 00 00 0
[ 90.591447][ T1131] RSP: 0018:
ffff8880ce78f1b8 EFLAGS:
00010247
[ 90.592329][ T1131] RAX:
dffffc0000000000 RBX:
0000000000000000 RCX:
ffff8880ce78f8b0
[ 90.593253][ T1131] RDX:
0000000000000000 RSI:
ffff8880ce78f4a0 RDI:
0000000000000004
[ 90.594058][ T1131] RBP:
ffff8880cf543e00 R08:
0000000000000002 R09:
0000000000000002
[ 90.594859][ T1131] R10:
ffffffffc0586a40 R11:
0000000000000000 R12:
ffff8880ca47c000
[ 90.595690][ T1131] R13:
ffff8880ca47c000 R14:
ffff8880cf545000 R15:
0000000000000000
[ 90.596553][ T1131] FS:
00007f21f6c7e0c0(0000) GS:
ffff8880da400000(0000) knlGS:
0000000000000000
[ 90.597504][ T1131] CS: 0010 DS: 0000 ES: 0000 CR0:
0000000080050033
[ 90.599418][ T1131] CR2:
0000556e413db458 CR3:
00000000c917a002 CR4:
00000000000606e0
[ 90.600289][ T1131] Call Trace:
[ 90.600631][ T1131] __rtnl_newlink+0x922/0x1270
[ 90.601194][ T1131] ? lock_downgrade+0x6e0/0x6e0
[ 90.601724][ T1131] ? rtnl_link_unregister+0x220/0x220
[ 90.602309][ T1131] ? lock_acquire+0x164/0x3b0
[ 90.602784][ T1131] ? is_bpf_image_address+0xff/0x1d0
[ 90.603331][ T1131] ? rtnl_newlink+0x4c/0x90
[ 90.603810][ T1131] ? kernel_text_address+0x111/0x140
[ 90.604419][ T1131] ? __kernel_text_address+0xe/0x30
[ 90.604981][ T1131] ? unwind_get_return_address+0x5f/0xa0
[ 90.605616][ T1131] ? create_prof_cpu_mask+0x20/0x20
[ 90.606304][ T1131] ? arch_stack_walk+0x83/0xb0
[ 90.606985][ T1131] ? stack_trace_save+0x82/0xb0
[ 90.607656][ T1131] ? stack_trace_consume_entry+0x160/0x160
[ 90.608503][ T1131] ? deactivate_slab.isra.78+0x2c5/0x800
[ 90.609336][ T1131] ? kasan_unpoison_shadow+0x30/0x40
[ 90.610096][ T1131] ? kmem_cache_alloc_trace+0x135/0x350
[ 90.610889][ T1131] ? rtnl_newlink+0x4c/0x90
[ 90.611512][ T1131] rtnl_newlink+0x65/0x90
[ ... ]
Fixes:
23790ef12082 ("net: qualcomm: rmnet: Allow to configure flags for existing devices")
Signed-off-by: Taehee Yoo <ap420073@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Taehee Yoo [Thu, 27 Feb 2020 12:23:52 +0000 (12:23 +0000)]
net: rmnet: fix NULL pointer dereference in rmnet_newlink()
rmnet registers IFLA_LINK interface as a lower interface.
But, IFLA_LINK could be NULL.
In the current code, rmnet doesn't check IFLA_LINK.
So, panic would occur.
Test commands:
modprobe rmnet
ip link add rmnet0 type rmnet mux_id 1
Splat looks like:
[ 36.826109][ T1115] general protection fault, probably for non-canonical address 0xdffffc0000000000I
[ 36.838817][ T1115] KASAN: null-ptr-deref in range [0x0000000000000000-0x0000000000000007]
[ 36.839908][ T1115] CPU: 1 PID: 1115 Comm: ip Not tainted 5.6.0-rc1+ #447
[ 36.840569][ T1115] Hardware name: innotek GmbH VirtualBox/VirtualBox, BIOS VirtualBox 12/01/2006
[ 36.841408][ T1115] RIP: 0010:rmnet_newlink+0x54/0x510 [rmnet]
[ 36.841986][ T1115] Code: 83 ec 18 48 c1 e9 03 80 3c 01 00 0f 85 d4 03 00 00 48 8b 6a 28 48 b8 00 00 00 00 00 c
[ 36.843923][ T1115] RSP: 0018:
ffff8880b7e0f1c0 EFLAGS:
00010247
[ 36.844756][ T1115] RAX:
dffffc0000000000 RBX:
ffff8880d14cca00 RCX:
1ffff11016fc1e99
[ 36.845859][ T1115] RDX:
0000000000000000 RSI:
ffff8880c3d04000 RDI:
0000000000000004
[ 36.846961][ T1115] RBP:
0000000000000000 R08:
ffff8880b7e0f8b0 R09:
ffff8880b6ac2d90
[ 36.848020][ T1115] R10:
ffffffffc0589a40 R11:
ffffed1016d585b7 R12:
ffffffff88ceaf80
[ 36.848788][ T1115] R13:
ffff8880c3d04000 R14:
ffff8880b7e0f8b0 R15:
ffff8880c3d04000
[ 36.849546][ T1115] FS:
00007f50ab3360c0(0000) GS:
ffff8880da000000(0000) knlGS:
0000000000000000
[ 36.851784][ T1115] CS: 0010 DS: 0000 ES: 0000 CR0:
0000000080050033
[ 36.852422][ T1115] CR2:
000055871afe5ab0 CR3:
00000000ae246001 CR4:
00000000000606e0
[ 36.853181][ T1115] Call Trace:
[ 36.853514][ T1115] __rtnl_newlink+0xbdb/0x1270
[ 36.853967][ T1115] ? lock_downgrade+0x6e0/0x6e0
[ 36.854420][ T1115] ? rtnl_link_unregister+0x220/0x220
[ 36.854936][ T1115] ? lock_acquire+0x164/0x3b0
[ 36.855376][ T1115] ? is_bpf_image_address+0xff/0x1d0
[ 36.855884][ T1115] ? rtnl_newlink+0x4c/0x90
[ 36.856304][ T1115] ? kernel_text_address+0x111/0x140
[ 36.856857][ T1115] ? __kernel_text_address+0xe/0x30
[ 36.857440][ T1115] ? unwind_get_return_address+0x5f/0xa0
[ 36.858063][ T1115] ? create_prof_cpu_mask+0x20/0x20
[ 36.858644][ T1115] ? arch_stack_walk+0x83/0xb0
[ 36.859171][ T1115] ? stack_trace_save+0x82/0xb0
[ 36.859710][ T1115] ? stack_trace_consume_entry+0x160/0x160
[ 36.860357][ T1115] ? deactivate_slab.isra.78+0x2c5/0x800
[ 36.860928][ T1115] ? kasan_unpoison_shadow+0x30/0x40
[ 36.861520][ T1115] ? kmem_cache_alloc_trace+0x135/0x350
[ 36.862125][ T1115] ? rtnl_newlink+0x4c/0x90
[ 36.864073][ T1115] rtnl_newlink+0x65/0x90
[ ... ]
Fixes:
ceed73a2cf4a ("drivers: net: ethernet: qualcomm: rmnet: Initial implementation")
Signed-off-by: Taehee Yoo <ap420073@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>