Subash Abhinov Kasiviswanathan says:
====================
net: qualcomm: rmnet: Enable csum offloads
This series introduces the MAPv4 packet format for checksum
offload plus some other minor changes.
Patches 1-3 are cleanups.
Patch 4 renames the ingress format to data format so that all data
formats can be configured using this going forward.
Patch 5 uses the pacing helper to improve TCP transmit performance.
Patch 6-9 defines the the MAPv4 for checksum offload for RX and TX.
A new header and trailer format are used as part of MAPv4.
For RX checksum offload, only the 1's complement of the IP payload
portion is computed by hardware. The meta data from RX header is
used to verify the checksum field in the packet. Note that the
IP packet and its field itself is not modified by hardware.
This gives metadata to help with the RX checksum. For TX, the
required metadata is filled up so hardware can compute the
checksum.
Patch 10 enables GSO on rmnet devices
v1->v2: Fix sparse errors reported by kbuild test robot
v2->v3: Update the commit message for Patch 5 based on Eric's comments
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
Real devices may support scatter gather(SG), so enable SG on rmnet
devices to use GSO. GSO reduces CPU cycles by 20% for a rate of
146Mpbs for a single stream TCP connection.
Signed-off-by: Subash Abhinov Kasiviswanathan <subashab@codeaurora.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
TX checksum offload applies to TCP / UDP packets which are not
fragmented using the MAPv4 checksum trailer. The following needs to be
done to have checksum computed in hardware -
1. Set the checksum start offset and inset offset.
2. Set the csum_enabled bit
3. Compute and set 1's complement of partial checksum field in
transport header.
Signed-off-by: Subash Abhinov Kasiviswanathan <subashab@codeaurora.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
When using the MAPv4 packet format in conjunction with MAP commands,
a dummy DL checksum trailer will be appended to the packet. Before
this packet is sent out as an ACK, the DL checksum trailer needs to be
removed.
Signed-off-by: Subash Abhinov Kasiviswanathan <subashab@codeaurora.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
When using the MAPv4 packet format, receive checksum offload can be
enabled in hardware. The checksum computation over pseudo header is
not offloaded but the rest of the checksum computation over
the payload is offloaded. This applies only for TCP / UDP packets
which are not fragmented.
rmnet validates the TCP/UDP checksum for the packet using the checksum
from the checksum trailer added to the packet by hardware. The
validation performed is as following -
1. Perform 1's complement over the checksum value from the trailer
2. Compute 1's complement checksum over IPv4 / IPv6 header and
subtracts it from the value from step 1
3. Computes 1's complement checksum over IPv4 / IPv6 pseudo header and
adds it to the value from step 2
4. Subtracts the checksum value from the TCP / UDP header from the
value from step 3.
5. Compares the value from step 4 to the checksum value from the
TCP / UDP header.
6. If the comparison in step 5 succeeds, CHECKSUM_UNNECESSARY is set
and the packet is passed on to network stack. If there is a
failure, then the packet is passed on as such without modifying
the ip_summed field.
The checksum field is also checked for UDP checksum 0 as per RFC 768
and for unexpected TCP checksum of 0.
If checksum offload is disabled when using MAPv4 packet format in
receive path, the packet is queued as is to network stack without
the validations above.
Signed-off-by: Subash Abhinov Kasiviswanathan <subashab@codeaurora.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
The MAPv4 packet format adds support for RX / TX checksum offload.
For a bi-directional UDP stream at a rate of 570 / 146 Mbps, roughly
10% CPU cycles are saved.
For receive path, there is a checksum trailer appended to the end of
the MAP packet. The valid field indicates if hardware has computed
the checksum. csum_start_offset indicates the offset from the start
of the IP header from which hardware has computed checksum.
csum_length is the number of bytes over which the checksum was
computed and the resulting value is csum_value.
In the transmit path, a header is appended between the end of the MAP
header and the start of the IP packet. csum_start_offset is the offset
in bytes from which hardware will compute the checksum if the
csum_enabled bit is set. udp_ip4_ind indicates if the checksum
value of 0 is valid or not. csum_insert_offset is the offset from the
csum_start_offset where hardware will insert the computed checksum.
The use of this additional packet format for checksum offload is
explained in subsequent patches.
Signed-off-by: Subash Abhinov Kasiviswanathan <subashab@codeaurora.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
The real device over which the rmnet devices are installed also
aggregate multiple IP packets and sends them as a single large
aggregate frame to the hardware. This causes degraded throughput
for TCP TX due to bufferbloat.
To overcome this problem, pacing shift value of 8 is set using the
sk_pacing_shift_update() helper. This value was determined based
on experiments with a single stream TCP TX using iperf for a
duration of 30s.
Pacing shift | Observed data rate (Mbps)
10 | 9
9 | 140
8 | 146 (Max link rate)
Signed-off-by: Subash Abhinov Kasiviswanathan <subashab@codeaurora.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
This is done so that we can use this field for both ingress and
egress flags.
Signed-off-by: Subash Abhinov Kasiviswanathan <subashab@codeaurora.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
rmnet_map_demultiplex() is only declared but not defined anywhere,
so remove it.
Signed-off-by: Subash Abhinov Kasiviswanathan <subashab@codeaurora.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
rmnet devices cannot have a mux id of 255. This is validated when
assigning the mux id to the rmnet devices. As a result, checking for
mux id 255 does not apply in egress path.
Signed-off-by: Subash Abhinov Kasiviswanathan <subashab@codeaurora.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
We already check the headroom once in rmnet_map_egress_handler(),
so this is not needed.
Signed-off-by: Subash Abhinov Kasiviswanathan <subashab@codeaurora.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
Calling acpi_wmi_init() at the subsys_initcall() level causes ordering
issues to appear on some systems and they are difficult to reproduce,
because there is no guaranteed ordering between subsys_initcall()
calls, so they may occur in different orders on different systems.
In particular, commit 86d9f48534 (mm/slab: fix kmemcg cache
creation delayed issue) exposed one of these issues where genl_init()
and acpi_wmi_init() are both called at the same initcall level, but
the former must run before the latter so as to avoid a NULL pointer
dereference.
For this reason, move the acpi_wmi_init() invocation to the
initcall_sync level which should still be early enough for things
to work correctly in the WMI land.
Link: https://marc.info/?t=151274596700002&r=1&w=2
Reported-by: Jonathan McDowell <noodles@earth.li>
Reported-by: Joonsoo Kim <iamjoonsoo.kim@lge.com>
Tested-by: Jonathan McDowell <noodles@earth.li>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Signed-off-by: Darren Hart (VMware) <dvhart@infradead.org>
Currently the less than zero error check on ret is incorrect
as it is checking a far earlier ret assignment rather than the
return from the call to wl1251_acx_arp_ip_filter. Fix this by
adding in the missing assginment.
Detected by CoverityScan, CID#1164835 ("Logically dead code")
Fixes: 204cc5c44f ("wl1251: implement hardware ARP filtering")
Signed-off-by: Colin Ian King <colin.king@canonical.com>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
Pausing queue without checking threshold is racy with txdone path.
Moreover we do not need pause queue on any error, but only if queue
is full - in case when we send RTS frame ( other cases of almost full
queue are already handled in rt2x00queue_write_tx_frame() ).
Patch fixes of theoretically possible problem of pausing empty
queue.
Signed-off-by: Stanislaw Gruszka <sgruszka@redhat.com>
Tested-by: Enrico Mioso <mrkiko.rs@gmail.com>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
Do not drop &queue->tx_lock and acquire it again to pause queue when
available entries are below the threshold.
Patch should mitigate problem of frequently printed errors:
"Error - Dropping frame due to full tx queue"
Signed-off-by: Stanislaw Gruszka <sgruszka@redhat.com>
Tested-by: Enrico Mioso@gmail.com
Tested-by: Enrico Mioso <mrkiko.rs@gmail.com>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
The last command used to shutdown firmware might be timeout,
and trigger firmware dump in asynchronous pcie/sdio work.
The remove/shutdown handler will continue free core data
structure private/adapter, which might be dereferenced in
pcie/sdio work, finally crash the kernel.
Sync and Cancel pcie/sdio work, could be a fix for above
cornel case. In this way, the last command timeout could
be handled properly.
Signed-off-by: Xinming Hu <huxm@marvell.com>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
This patch extend device_dump debugfs function to make it
works for usb interface.
For command timeouts, USB firmware will automatically emit
firmware dump events, so we don't implement device_dump().
For user-initiated dumps, we trigger it by issue firmware
dump event command to firmware, as there is no command
response, do not start 10s timer.
Signed-off-by: Xinming Hu <huxm@marvell.com>
Signed-off-by: Cathy Luo <cluo@marvell.com>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
Firmware dump on usb interface is different with current
sdio/pcie chipset, which is based on register operation.
When firmware hang on usb interface, context dump will be
upload to host using 0x73 firmware debug event.
This patch store dump data from debug event and send to
userspace using device coredump API.
Signed-off-by: Xinming Hu <huxm@marvell.com>
Signed-off-by: Cathy Luo <cluo@marvell.com>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
This patch refactor current device dump code to make it generic
for subsequent implementation on usb interface.
Signed-off-by: Xinming Hu <huxm@marvell.com>
Signed-off-by: Cathy Luo <cluo@marvell.com>
Signed-off-by: Ganapathi Bhat <gbhat@marvell.com>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
Currently if the allocation of skb fails and returns NULL then the
call to skb_put will cause a null pointer dereference. Fix this by
checking for a null skb and returning NULL. Note that calls to
function mt76x2_mcu_msg_alloc don't directly check the null return
but instead pass the NULL pointer to mt76x2_mcu_msg_send which
checks for the NULL and returns ENOMEM in this case.
Detected by CoverityScan, CID#1462624 ("Dereference null return value")
Fixes: 7bc04215a6 ("mt76: add driver code for MT76x2e")
Signed-off-by: Colin Ian King <colin.king@canonical.com>
Acked-by: Felix Fietkau <nbd@nbd.name>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
mt76x2_mac_start writes dev->rxfilter to the hardware. It also happens
during init, before dev->rxfilter is filled with the initval register
value, leading to issues like promisc mode being enabled
unconditionally.
Fix this by reading the default value into dev->rxfilter earlier
Signed-off-by: Lorenzo Bianconi <lorenzo.bianconi83@gmail.com>
Signed-off-by: Felix Fietkau <nbd@nbd.name>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
Due to an unrelated issue, the MT_RX_FILTR_CFG_PROMISC flag is currently
unset, which means that monitor mode is unconditionally enabled.
Toggle this flag based on the mac80211 monitor mode setting instead
Signed-off-by: Felix Fietkau <nbd@nbd.name>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
It is no longer necessary for client mode operation, vif index entries
8-16 are no longer used
Signed-off-by: Lorenzo Bianconi <lorenzo.bianconi83@gmail.com>
Signed-off-by: Felix Fietkau <nbd@nbd.name>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
Using both chains adds max. 3 dBm. A similar worst-case calculation is
being used in ath9k as well to ensure that the hardware stays within
regulatory limits
Signed-off-by: Felix Fietkau <nbd@nbd.name>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
This allows user space to query the real hardware limits directly
Signed-off-by: Felix Fietkau <nbd@nbd.name>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
Discard affected packets instead. Should reduce the frequency of bogus
bug reports
Signed-off-by: Felix Fietkau <nbd@nbd.name>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
drivers/net/wireless/mediatek/mt76/mt76x2_main.c:86:5-8: Unneeded variable: "ret". Return "0" on line 112
Remove unneeded variable used to store return value.
Generated by: scripts/coccinelle/misc/returnvar.cocci
Fixes: a5f6039c8f9c ("mt76: add driver code for MT76x2e")
CC: Felix Fietkau <nbd@nbd.name>
Signed-off-by: Fengguang Wu <fengguang.wu@intel.com>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
drivers/net/wireless/mediatek/mt76/debugfs.c:36:0-23: WARNING: fops_regval should be defined with DEFINE_DEBUGFS_ATTRIBUTE
Use DEFINE_DEBUGFS_ATTRIBUTE rather than DEFINE_SIMPLE_ATTRIBUTE
for debugfs files.
Semantic patch information:
Rationale: DEFINE_SIMPLE_ATTRIBUTE + debugfs_create_file()
imposes some significant overhead as compared to
DEFINE_DEBUGFS_ATTRIBUTE + debugfs_create_file_unsafe().
Generated by: scripts/coccinelle/api/debugfs/debugfs_simple_attr.cocci
Fixes: a5f6039c8f9c ("mt76: add driver code for MT76x2e")
CC: Felix Fietkau <nbd@nbd.name>
Signed-off-by: Fengguang Wu <fengguang.wu@intel.com>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
This patch will add 43455 into the save-restore(SR) capable chip list, so
the SR engine will be enabled with 43455 FW which built-in the -sr
function.
Signed-off-by: Double Lo <double.lo@cypress.com>
Signed-off-by: Wright Feng <wright.feng@cypress.com>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
syzbot reported the following panic in the verifier triggered
by kmalloc error injection:
kasan: GPF could be caused by NULL-ptr deref or user memory access
RIP: 0010:copy_func_state kernel/bpf/verifier.c:403 [inline]
RIP: 0010:copy_verifier_state+0x364/0x590 kernel/bpf/verifier.c:431
Call Trace:
pop_stack+0x8c/0x270 kernel/bpf/verifier.c:449
push_stack kernel/bpf/verifier.c:491 [inline]
check_cond_jmp_op kernel/bpf/verifier.c:3598 [inline]
do_check+0x4b60/0xa050 kernel/bpf/verifier.c:4731
bpf_check+0x3296/0x58c0 kernel/bpf/verifier.c:5489
bpf_prog_load+0xa2a/0x1b00 kernel/bpf/syscall.c:1198
SYSC_bpf kernel/bpf/syscall.c:1807 [inline]
SyS_bpf+0x1044/0x4420 kernel/bpf/syscall.c:1769
when copy_verifier_state() aborts in the middle due to kmalloc failure
some of the frames could have been partially copied while
current free_verifier_state() loop
for (i = 0; i <= state->curframe; i++)
assumed that all frames are non-null.
Simply fix it by adding 'if (!state)' to free_func_state().
Also avoid stressing copy frame logic more if kzalloc fails
in push_stack() free env->cur_state right away.
Fixes: f4d7e40a5b ("bpf: introduce function calls (verification)")
Reported-by: syzbot+32ac5a3e473f2e01cfc7@syzkaller.appspotmail.com
Reported-by: syzbot+fa99e24f3c29d269a7d5@syzkaller.appspotmail.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
In preparation for removing the function array, remove all code that
refers to function by index and replace with pointers to the function
itself.
Signed-off-by: Ian Molton <ian@mnementh.co.uk>
Reviewed-by: Arend van Spriel <arend.vanspriel@broadcom.com>
[arend: replace BUG() with WARN() macro]
Signed-off-by: Arend van Spriel <arend.vanspriel@broadcom.com>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
Its more efficient to test the register we're interested in first,
potentially avoiding two more comparisons, and therefore always avoiding
one comparison per call on all other chips.
Signed-off-by: Ian Molton <ian@mnementh.co.uk>
[arend: fix some checkpatch warnings]
Signed-off-by: Arend van Spriel <arend.vanspriel@broadcom.com>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
func0 is not provided by the mmc stack as a function when probing.
Instead providing specific access functions to read/write it.
This prepares for a patch to remove the actual array entry itself.
Signed-off-by: Ian Molton <ian@mnementh.co.uk>
Acked-by: Arend van Spriel <arend.vanspriel@broadcom.com>
[arend: rephrased the commit message]
[arend: removed unrelated comment for which separate patch is warranted]
Signed-off-by: Arend van Spriel <arend.vanspriel@broadcom.com>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
The IO functions operate within the Chipcommon IO window. Explicitly
set this, rather than relying on the last initialisation IO access to
leave it set to the right value by chance.
Signed-off-by: Ian Molton <ian@mnementh.co.uk>
Acked-by: Arend van Spriel <arend.vanspriel@broadcom.com>
Signed-off-by: Arend van Spriel <arend.vanspriel@broadcom.com>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
The matching of the counters was not taken into account, fixed.
Signed-off-by: Jozsef Kadlecsik <kadlec@blackhole.kfki.hu>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
Make use of the swap macro and remove unnecessary variables tmp.
This makes the code easier to read and maintain.
This code was detected with the help of Coccinelle.
Signed-off-by: Gustavo A. R. Silva <garsilva@embeddedor.com>
Signed-off-by: Jozsef Kadlecsik <kadlec@blackhole.kfki.hu>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
Add new instruction for the nf_tables VM that allows us to specify what
flows are offloaded into a given flow table via name. This new
instruction creates the flow entry and adds it to the flow table.
Only established flows, ie. we have seen traffic in both directions, are
added to the flow table. You can still decide to offload entries at a
later stage via packet counting or checking the ct status in case you
want to offload assured conntracks.
This new extension depends on the conntrack subsystem.
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
This patch adds the IPv6 flow table type, that implements the datapath
flow table to forward IPv6 traffic.
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
This patch adds the IPv6 flow table type, that implements the datapath
flow table to forward IPv6 traffic.
This patch exports ip6_dst_mtu_forward() that is required to check for
mtu to pass up packets that need PMTUD handling to the classic
forwarding path.
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
This patch adds the IPv4 flow table type, that implements the datapath
flow table to forward IPv4 traffic. Rationale is:
1) Look up for the packet in the flow table, from the ingress hook.
2) If there's a hit, decrement ttl and pass it on to the neighbour layer
for transmission.
3) If there's a miss, packet is passed up to the classic forwarding
path.
This patch also supports layer 3 source and destination NAT.
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
