Define the MAC_PUSH action which pushes an MPLS LSE before the mac
header (instead of between the mac and the network headers as the
plain PUSH action does).

The only special case is when the skb has an offloaded VLAN. In that
case, it has to be inlined before pushing the MPLS header.
Signed-off-by: Guillaume Nault <gnault@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Implement TCA_VLAN_ACT_POP_ETH and TCA_VLAN_ACT_PUSH_ETH, to
respectively pop and push a base Ethernet header at the beginning of a
frame.

POP_ETH is just a matter of pulling ETH_HLEN bytes. VLAN tags, if any,
must be stripped before calling POP_ETH.

PUSH_ETH is restricted to skbs with no mac_header, and only the MAC
addresses can be configured. The Ethertype is automatically set from
skb->protocol. These restrictions ensure that all skb's fields remain
consistent, so that this action can't confuse other part of the
networking stack (like GSO).

Since openvswitch already had these actions, consolidate the code in
skbuff.c (like for vlan and mpls push/pop).
Signed-off-by: Guillaume Nault <gnault@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Currently the comparisons of u16 integers value and sopass_val with
less than zero for error checking is always false because the values
are unsigned. Fix this by making these variables int.  This does not
affect the shift and mask operations performed on these variables

Addresses-Coverity: ("Unsigned compared against zero")
Fixes: 49fc2301 ("net: phy: dp83869: support Wake on LAN")
Signed-off-by: Colin Ian King <colin.king@canonical.com>
Acked-by: Dan Murphy <dmurphy@ti.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

The check for old hardware versions that did not have SMCDv2 support was
using suspicious pointer magic. Address the fields using an array.
Signed-off-by: Karsten Graul <kgraul@linux.ibm.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

When building a CLC proposal message then the list of ISM devices does
not need to contain multiple devices that have the same chid value,
all these devices use the same function at the end.
Improve smc_find_ism_v2_device_clnt() to collect only ISM devices that
have unique chid values.
Signed-off-by: Karsten Graul <kgraul@linux.ibm.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Vladimir Oltean says:

===================
Add Seville Ethernet switch to T1040RDB

Seville is a DSA switch that is embedded inside the T1040 SoC, and
supported by the mscc_seville DSA driver inside drivers/net/dsa/ocelot.

This series adds this switch to the SoC's dtsi files and to the T1040RDB
board file.

I would like to send this series through net-next. There is no conflict
with other patches submitted to T1040 device tree. Maybe this could at
least get an ACK from devicetree maintainers.
===================
Signed-off-by: David S. Miller <davem@davemloft.net>

Define the network interface names for the switch ports and hook them up
to the 2 QSGMII PHYs that are onboard.

A conscious decision was taken to go along with the numbers that are
written on the front panel of the board and not with the hardware
numbers of the switch chip ports.
Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Reviewed-by: Maxim Kochetkov <fido_max@inbox.ru>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: David S. Miller <davem@davemloft.net>

Add the description of the embedded L2 switch inside the SoC dtsi file
for NXP T1040.
Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Reviewed-by: Maxim Kochetkov <fido_max@inbox.ru>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: David S. Miller <davem@davemloft.net>

The retransmission refactoring patch
68698970 ("tcp: simplify tcp_mark_skb_lost")
does not properly update the total lost packet counter which may
break the policer mode in BBR. This patch fixes it.

Fixes: 68698970 ("tcp: simplify tcp_mark_skb_lost")
Reported-by: Neal Cardwell <ncardwell@google.com>
Signed-off-by: Yuchung Cheng <ycheng@google.com>
Signed-off-by: Neal Cardwell <ncardwell@google.com>
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Julian Wiedmann says:

====================
net/iucv: updates 2020-10-01

Just two (rare) patches, and both deal with smatch warnings.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>

smatch complains about
net/iucv/iucv.c:1119 __iucv_message_receive() warn: inconsistent indenting

While touching this line, also make the return logic consistent and thus
get rid of a goto label.
Signed-off-by: Julian Wiedmann <jwi@linux.ibm.com>
Reviewed-by: Karsten Graul <kgraul@linux.ibm.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

smatch complains about
net/iucv/af_iucv.c:624 iucv_sock_bind() error: memcpy() 'sa->siucv_user_id' too small (8 vs 9)

Which is absolutely correct - the memcpy() takes 9 bytes (sizeof(uid))
from an 8-byte field (sa->siucv_user_id).
Luckily the sockaddr_iucv struct contains more data after the
.siucv_user_id field, and we checked the size of the passed data earlier
on. So the memcpy() won't accidentally read from an invalid location.

Fix the warning by reducing the size of the uid variable to what's
actually needed, and thus reducing the amount of copied data.
Signed-off-by: Julian Wiedmann <jwi@linux.ibm.com>
Reviewed-by: Karsten Graul <kgraul@linux.ibm.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

NETDEV_HW_ADDR_T_SLAVE is not used anymore, remove it.
Signed-off-by: Taehee Yoo <ap420073@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Johannes Berg says:

====================
genetlink per-op policy export

Here's a respin, now including Jakub's patch last so that it will
do the right thing from the start.

The first patch remains the same, of course; the others have mostly
some rebasing going on, except for the actual export patch (patch 4)
which is adjusted per Jakub's review comments about exporting the
policy only if it's actually used for do/dump.

To see that, the dump for "nlctrl" (i.e. the generic netlink control)
is instructive, because the ops are this:

        {
                .cmd            = CTRL_CMD_GETFAMILY,
                .validate = GENL_DONT_VALIDATE_STRICT | GENL_DONT_VALIDATE_DUMP,
                .policy         = ctrl_policy_family,
                .maxattr        = ARRAY_SIZE(ctrl_policy_family) - 1,
                .doit           = ctrl_getfamily,
                .dumpit         = ctrl_dumpfamily,
        },
        {
                .cmd            = CTRL_CMD_GETPOLICY,
                .policy         = ctrl_policy_policy,
                .maxattr        = ARRAY_SIZE(ctrl_policy_policy) - 1,
                .start          = ctrl_dumppolicy_start,
                .dumpit         = ctrl_dumppolicy,
                .done           = ctrl_dumppolicy_done,
        },

So we exercise both "don't have doit" and "GENL_DONT_VALIDATE_DUMP"
parts, and get (with the current genl patch):

$ genl ctrl policy name nlctrl
	ID: 0x10  op 3 policies: do=0
	ID: 0x10  op 10 policies: dump=1
	ID: 0x10  policy[0]:attr[1]: type=U16 range:[0,65535]
	ID: 0x10  policy[0]:attr[2]: type=NUL_STRING max len:15
	ID: 0x10  policy[1]:attr[1]: type=U16 range:[0,65535]
	ID: 0x10  policy[1]:attr[2]: type=NUL_STRING max len:15
	ID: 0x10  policy[1]:attr[10]: type=U32 range:[0,4294967295]
====================
Reviewed-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>

Right now CTRL_CMD_GETPOLICY can only dump the family-wide
policy. Support dumping policy of a specific op.

v3:
 - rebase after per-op policy export and handle that
v2:
 - make cmd U32, just in case.
v1:
 - don't echo op in the output in a naive way, this should
   make it cleaner to extend the output format for dumping
   policies for all the commands at once in the future.
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Link: https://lore.kernel.org/r/20201001225933.1373426-11-kuba@kernel.orgSigned-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Add support for per-op policy dumping. The data is pretty much
as before, except that now the assumption that the policy with
index 0 is "the" policy no longer holds - you now need to look
at the new CTRL_ATTR_OP_POLICY attribute which is a nested attr
(indexed by op) containing attributes for do and dump policies.

When a single op is requested, the CTRL_ATTR_OP_POLICY will be
added in the same way, since do and dump policies may differ.

v2:
 - conditionally advertise per-command policies only if there
   actually is a policy being used for the do/dump and it's
   present at all
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

We'll need this later for the per-op policy index dump.
Reviewed-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Rework the policy dump code a bit to support adding multiple
policies to a single dump, in order to e.g. support per-op
policies in generic netlink.

v2:
 - move kernel-doc to implementation [Jakub]
 - squash the first patch to not flip-flop on the prototype
   [Jakub]
 - merge netlink_policy_dump_get_policy_idx() with the old
   get_policy_idx() we already had
 - rebase without Jakub's patch to have per-op dump
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

The maxtype is really an integral part of the policy, and while we
haven't gotten into a situation yet where this happens, it seems
that some developer might eventually have two places pointing to
identical policies, with different maxattr to exclude some attrs
in one of the places.

Even if not, it's really the right thing to compare both since the
two data items fundamentally belong together.

v2:
 - also do the proper comparison in get_policy_idx()
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Jakub Kicinski says:

====================
genetlink: support per-command policy dump

The objective of this series is to dump ethtool policies
to be able to tell which flags are supported by the kernel.
Current release adds ETHTOOL_FLAG_STATS for dumping extra
stats, but because of strict checking we need to make sure
that the flag is actually supported before setting it in
a request.

Ethtool policies are per command, and so far only dumping
family policies was supported.

The series adds new set of "light" ops to genl families which
don't have all the callbacks, and won't have the policy.
Most of families are then moved to these ops. This gives
us 4096B in savings on an allyesconfig build (not counting
the growth that would have happened when policy is added):

     text       data       bss        dec       hex
244415581  227958581  78372980  550747142  20d3bc06
244415581  227962677  78372980  550751238  20d3cc06

Next 5 patches deal the dumping per-op policy.

v3:

The actually patch to dump per-op policy was taken out and
will come in a series from Johannes, to make sure uAPI is
consistent from the start.

For dump-specific policies I think it should be fine to add
a new pair of members to the "full" ops, and not overthink it.

v2:
 - remove the stale comment in taskstats
 - split patch 8 -> 8, 9
 - now the getfamily policy is also in the op
 - make cmd u32
v1:
 - replace remaining uses of "light" with "small"
 - fix dump (ops can't be on the stack there)
 - coding changes in patch 4
 - new patch 7
 - don't echo op in responses - to make dump all easier

Dave - this series will cause a very trivial conflict with
the patch I sent to net. Both sides add some kdoc to struct
genl_ops so we'll need to keep it all.  I'm sending this
already because I also need to restructure ethool policies
in time for 5.10 if we want to use it for the stats flag.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>

In preparation for adding a new attribute to CTRL_CMD_GETPOLICY
split the policies for getpolicy and getfamily apart.

This will cause a slight user-visible change in that dumping
the policies will switch from per family to per op, but
supposedly sniffer-type applications (which are the main use
case for policy dumping thus far) should support both, anyway.
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Reviewed-by: Johannes Berg <johannes@sipsolutions.net>
Signed-off-by: David S. Miller <davem@davemloft.net>

Attributes are already parsed based on the policy specified
in the family and ready-to-use in info->attrs. No need to
call genlmsg_parse() again.
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Reviewed-by: Johannes Berg <johannes@sipsolutions.net>
Signed-off-by: David S. Miller <davem@davemloft.net>

commit 3b0f31f2 ("genetlink: make policy common to family")
had to work around removal of policy from ops by parsing in
the pre_doit callback. Now that policy is back in full ops
we can switch again. Set maxattr to actual size of the policies
- both commands set GENL_DONT_VALIDATE_STRICT so out of range
attributes will be silently ignored, anyway.

v2:
 - remove stale comment
Suggested-by: Johannes Berg <johannes@sipsolutions.net>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Reviewed-by: Johannes Berg <johannes@sipsolutions.net>
Signed-off-by: David S. Miller <davem@davemloft.net>

Add policy to the struct genl_ops structure, this time
with maxattr, so it can be used properly.

Propagate .policy and .maxattr from the family
in genl_get_cmd() if needed, this way the rest of the
code does not have to worry if the policy is per op
or global.
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Reviewed-by: Johannes Berg <johannes@sipsolutions.net>
Signed-off-by: David S. Miller <davem@davemloft.net>

The structure of ctrl_dumppolicy() is clearly split into
init and dumping. Move the init to a .start callback
for clarity, it's a more idiomatic netlink dump code structure.
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Reviewed-by: Johannes Berg <johannes@sipsolutions.net>
Signed-off-by: David S. Miller <davem@davemloft.net>

Whenever netlink dump uses more than 2 cb->args[] entries
code gets hard to read. We're about to add more state to
ctrl_dumppolicy() so create a structure.

Since the structure is typed and clearly named we can remove
the local fam_id variable and use ctx->fam_id directly.

v3:
 - rebase onto explicit free fix
v1:
 - s/nl_policy_dump/netlink_policy_dump_state/
 - forward declare struct netlink_policy_dump_state,
   and move from passing unsigned long to actual pointer type
 - add build bug on
 - u16 fam_id
 - s/args/ctx/
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Reviewed-by: Johannes Berg <johannes@sipsolutions.net>
Signed-off-by: David S. Miller <davem@davemloft.net>

Bulk of the genetlink users can use smaller ops, move them.
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Reviewed-by: Johannes Berg <johannes@sipsolutions.net>
Signed-off-by: David S. Miller <davem@davemloft.net>

We want to add maxattr and policy back to genl_ops, to enable
dumping per command policy to user space. This, however, would
cause bloat for all the families with global policies. Introduce
smaller version of ops (half the size of genl_ops). Translate
these smaller ops into a full blown struct before use in the
core.

v1:
 - use struct assignment
 - put a full copy of the op in struct genl_dumpit_info
 - s/light/small/
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Reviewed-by: Johannes Berg <johannes@sipsolutions.net>
Signed-off-by: David S. Miller <davem@davemloft.net>

There are holes and oversized members in struct genl_family.

Before: /* size: 104, cachelines: 2, members: 16 */
After:  /* size:  88, cachelines: 2, members: 16 */

The command field in struct genlmsghdr is a u8, so no point
in the operation count being 32 bit. Also operation 0 is
usually undefined, so we only need 255 entries.

netnsok and parallel_ops are only ever initialized to true.

We can grow the fields as needed, compiler should warn us
if someone tries to assign larger constants.
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Reviewed-by: Johannes Berg <johannes@sipsolutions.net>
Signed-off-by: David S. Miller <davem@davemloft.net>

Ioana Ciornei says:

====================
dpaa2-eth: add devlink parser error drop trap support

This patch set adds support in the dpaa2-eth driver for a new group of
devlink drop traps - PARSER_ERROR_DROPS.

The first patch adds a new generic trap group and associated traps,
their definitions in devlink and their corresponding entries in the
Documentation.

Because there might be more devices (besides DPAA2) which do not support
changing the action independently on each trap, a nre devlink callback
is introduced - .trap_group_action_set(). If this callback is populated,
it will take precedence over .trap_action_set() when the user requests
changing the action on all the traps in a group.

The next patches add basic linkage with devlink for the dpaa2-eth driver
and support for the newly added PARSER_ERROR_DROPS. Nothing special
here, just setting up the Rx error queue, interpreting the parse result,
and then reporting any frame received on that queue to devlink.

Changes in v2:
 - fix build error in 3/4

Changes in v3:
 - removed a commented line in 4/4
 - added an extack in 4/4
 - fixed up a warning on 32bit in 4/4
 - reworded the trap_group_action_set() description in 2/4
====================
Signed-off-by: David S. Miller <davem@davemloft.net>

Add support for the new group of devlink traps - PARSER_ERROR_DROPS.
This consists of registering the array of parser error drops supported,
controlling their action through the .trap_group_action_set() callback
and reporting an erroneous skb received on the error queue
appropriately.
DPAA2 devices do not support controlling the action of independent
parser error traps, thus the .trap_action_set() callback just returns an
EOPNOTSUPP while .trap_group_action_set() actually notifies the hardware
what it should do with a frame marked as having a header error.
Signed-off-by: Ioana Ciornei <ioana.ciornei@nxp.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Add basic support in dpaa2-eth for devlink. For the moment, just
register the device with devlink, add the corresponding devlink port and
implement the .info_get() callback.
Signed-off-by: Ioana Ciornei <ioana.ciornei@nxp.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Add a new devlink callback, .trap_group_action_set(), which can be used
by device drivers which do not support controlling the action (drop,
trap) on each trap but rather on the entire group trap.
If this new callback is populated, it will take precedence over the
.trap_action_set() callback when the user requests a change of all the
traps in a group.
Signed-off-by: Ioana Ciornei <ioana.ciornei@nxp.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Add parser error drop packet traps, so that capable device driver could
register them with devlink. The new packet trap group holds any drops of
packets which were marked by the device as erroneous during header
parsing. Add documentation for every added packet trap and packet trap
group.
Signed-off-by: Ioana Ciornei <ioana.ciornei@nxp.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Shannon Nelson says:

====================
ionic error recovery

This set of patches comes mostly from error recovery path testing,
as well as a couple of upstream review comments.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>

If the new firmware image downladed for update is corrupted
or is a bad format, the download process will report a status
code specifically for that.
Signed-off-by: Shannon Nelson <snelson@pensando.io>
Signed-off-by: David S. Miller <davem@davemloft.net>

Use the lif's ident information for the uc and mc filter
counts rather than the ionic's version, to be sure
we're getting the info that is specific to this lif.

While we're thinking about it, add some missing error
checking where we get the lif's identity information.
Signed-off-by: Shannon Nelson <snelson@pensando.io>
Signed-off-by: David S. Miller <davem@davemloft.net>

After we do a fw upgrade and refill the ionic->ident.dev, we
also need to update the other identity info.  Since the lif
identity needs to be updated each time the ionic identity is
refreshed, we can pull it into ionic_identify().

The debugfs entry is moved so that it doesn't cause an
error message when the data is refreshed after the fw upgrade.
Signed-off-by: Shannon Nelson <snelson@pensando.io>
Signed-off-by: David S. Miller <davem@davemloft.net>

Some time ago we short-circuited the queue disables on a timeout
error in order to not have to wait on every queue when we already
know it will time out.  However, this meant that we're not
properly stopping all the interrupts and napi contexts.  This
changes queue disable to always call ionic_qcq_disable() and to
give it an argument to know when to not do the adminq request.
Signed-off-by: Shannon Nelson <snelson@pensando.io>
Signed-off-by: David S. Miller <davem@davemloft.net>

There are a couple of error recovery paths that can come through
ionic_qcq_disable() without having set up the qcq, so we need
to make sure we have a valid qcq pointer before using it.
Signed-off-by: Shannon Nelson <snelson@pensando.io>
Signed-off-by: David S. Miller <davem@davemloft.net>