kbuild test robot reports for sparse warnings in
commits c2a4ffb70e ("ipvs: convert lblc scheduler to rcu")
and c5549571f9 ("ipvs: convert lblcr scheduler to rcu").
Fix it by removing extra __rcu annotation.
Signed-off-by: Julian Anastasov <[email protected]>
Signed-off-by: Simon Horman <[email protected]>
Use rcu_dereference_protected to resolve
sparse warning, found by kbuild test robot:
net/netfilter/ipvs/ip_vs_ctl.c:1464:35: warning: dereference of
noderef expression
Problem from commit 026ace060d
("ipvs: optimize dst usage for real server")
Signed-off-by: Julian Anastasov <[email protected]>
Signed-off-by: Simon Horman <[email protected]>
batadv_mesh_free() schedules some RCU callbacks which need the bat_priv struct
to do their jobs, while free_netdev(), which is called immediately after, is
destroying the private data.
Put an rcu_barrier() in the middle so that free_netdev() is invoked only after
all the callbacks returned.
This bug has been introduced by ab8f433dd39be94e8617cff2dfe9f7eca162eb15
("batman-adv: Move deinitialization of soft-interface to destructor")
Signed-off-by: Antonio Quartulli <[email protected]>
Signed-off-by: Marek Lindner <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
Conflicts:
drivers/net/ethernet/emulex/benet/be_main.c
drivers/net/ethernet/intel/igb/igb_main.c
drivers/net/wireless/brcm80211/brcmsmac/mac80211_if.c
include/net/scm.h
net/batman-adv/routing.c
net/ipv4/tcp_input.c
The e{uid,gid} --> {uid,gid} credentials fix conflicted with the
cleanup in net-next to now pass cred structs around.
The be2net driver had a bug fix in 'net' that overlapped with the VLAN
interface changes by Patrick McHardy in net-next.
An IGB conflict existed because in 'net' the build_skb() support was
reverted, and in 'net-next' there was a comment style fix within that
code.
Several batman-adv conflicts were resolved by making sure that all
calls to batadv_is_my_mac() are changed to have a new bat_priv first
argument.
Eric Dumazet's TS ECR fix in TCP in 'net' conflicted with the F-RTO
rewrite in 'net-next', mostly overlapping changes.
Thanks to Stephen Rothwell and Antonio Quartulli for help with several
of these merge resolutions.
Signed-off-by: David S. Miller <[email protected]>
This driver uses the crummy "| foo" style, putting the
logical operation at the beginning of lines. Then
when the VLAN tag flag bits got changed the operator
ended up at both the end and the beginning of some lines.
Fix the build error by having it uniformly use the operator
at the end of the line.
Reported-by: Stephen Rothwell <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
Several call sites were missed when the protocol argument was added to
__vlan_hwaccel_put_tag() in commit
86a9bad3ab ("net: vlan: add protocol
argument to packet tagging functions").
Reported-by: Stephen Rothwell <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
struct sctp_packet is currently embedded into sctp_transport or
sits on the stack as 'singleton' in sctp_outq_flush(). Therefore,
its member 'malloced' is always 0, thus a kfree() is never called.
Because of that, we can just remove this code.
Signed-off-by: Daniel Borkmann <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
Allow L2 redirection when VXLAN L3 switching is enabled
This patch restricts L3 switching to destination MAC addresses that are
marked as routers in order to allow virtual IP appliances that do L2
redirection to function with VXLAN L3 switching enabled.
We use L3 switching on VXLAN networks to avoid extra hops when the nominal
router for cross-subnet traffic for a VM is remote and the ultimate
destination may be local, or closer to the local node. Currently, the
destination IP address takes precedence over the MAC address in all cases.
Some network appliances receive packets for a virtualized IP address and
redirect by changing the destination MAC address (only) to be the final
destination for packet processing. VXLAN tunnel endpoints with L3 switching
enabled may then overwrite this destination MAC address based on the packet IP
address, resulting in potential loops and, at least, breaking L2 redirections
that travel through tunnel endpoints.
This patch limits L3 switching to the intended case where the original
destination MAC address is a next-hop router and relies on the destination
MAC address for all other cases, thus allowing L2 redirection and L3 switching
to coexist peacefully.
Signed-Off-By: David L Stevens <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
The return value from list_netdevice() is not used and no need, so remove it.
Signed-off-by: Ding Tianhong <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
dl_next member in struct request_sock doesn't need to be first.
We expect to insert a "struct common_sock" or a subset of it,
so this claim had to be verified.
Signed-off-by: Eric Dumazet <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
qeth_hdr_chk_and_bounce() can possibly shift the skb->data
pointer. However, the existing code didn't update the hdr pointer,
which should point to skb->data, accordingly.
Symptoms of this issue are sporadic recoveries.
Signed-off-by: Stefan Raspl <[email protected]>
Signed-off-by: Frank Blaschka <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
Initialization of bss_descriptor is unnecessary as the entire
structure will be overwritten by a memcpy followed by.
Initialize disable_11ac flag properly by setting it to true in
mwifiex_fill_new_bss_desc().
Reported-by: Doug Anderson <[email protected]>
Reviewed-by: Doug Anderson <[email protected]>
Reviewed-by: Paul Stewart <[email protected]>
Signed-off-by: Bing Zhao <[email protected]>
Signed-off-by: Yogesh Ashok Powar <[email protected]>
Signed-off-by: John W. Linville <[email protected]>
beacon_ie buffer is allocated in mwifiex_fill_new_bss_desc()
and the buffer pointer is saved in bss_desc->beacon_buf.
beacon_ie is freed before the function returns. However,
bss_desc->beacon_buf is still being accessed afterwards.
Fix it by freeing beacon_ie (bss_desc->beacon_buf) in
caller's scope.
Reviewed-by: Doug Anderson <[email protected]>
Reviewed-by: Paul Stewart <[email protected]>
Signed-off-by: Bing Zhao <[email protected]>
Signed-off-by: John W. Linville <[email protected]>
Some frontend drivers are sending packets > 64 KiB in length. This length
overflows the length field in the first slot making the following slots have
an invalid length.
Turn this error back into a non-fatal error by dropping the packet. To avoid
having the following slots having fatal errors, consume all slots in the
packet.
This does not reopen the security hole in XSA-39 as if the packet as an
invalid number of slots it will still hit fatal error case.
Signed-off-by: David Vrabel <[email protected]>
Signed-off-by: Wei Liu <[email protected]>
Acked-by: Ian Campbell <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
This patch tries to coalesce tx requests when constructing grant copy
structures. It enables netback to deal with situation when frontend's
MAX_SKB_FRAGS is larger than backend's MAX_SKB_FRAGS.
With the help of coalescing, this patch tries to address two regressions
avoid reopening the security hole in XSA-39.
Regression 1. The reduction of the number of supported ring entries (slots)
per packet (from 18 to 17). This regression has been around for some time but
remains unnoticed until XSA-39 security fix. This is fixed by coalescing
slots.
Regression 2. The XSA-39 security fix turning "too many frags" errors from
just dropping the packet to a fatal error and disabling the VIF. This is fixed
by coalescing slots (handling 18 slots when backend's MAX_SKB_FRAGS is 17)
which rules out false positive (using 18 slots is legit) and dropping packets
using 19 to `max_skb_slots` slots.
To avoid reopening security hole in XSA-39, frontend sending packet using more
than max_skb_slots is considered malicious.
The behavior of netback for packet is thus:
1-18 slots: valid
19-max_skb_slots slots: drop and respond with an error
max_skb_slots+ slots: fatal error
max_skb_slots is configurable by admin, default value is 20.
Also change variable name from "frags" to "slots" in netbk_count_requests.
Please note that RX path still has dependency on MAX_SKB_FRAGS. This will be
fixed with separate patch.
Signed-off-by: Wei Liu <[email protected]>
Acked-by: Ian Campbell <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
The maximum packet including header that can be handled by netfront / netback
wire format is 65535. Reduce gso_max_size accordingly.
Drop skb and print warning when skb->len > 65535. This can 1) save the effort
to send malformed packet to netback, 2) help spotting misconfiguration of
netfront in the future.
Signed-off-by: Wei Liu <[email protected]>
Acked-by: Ian Campbell <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
This patch sets the coherent DMA mask to 64-bit after the be2net driver has
been acknowledged that the system is 64-bit DMA capable. The coherent DMA
mask is examined by the Intel IOMMU driver to determine whether to allow
pass through context mapping for all devices. With this patch, the be2net
driver combined with be2net compatible hardware provides comparable
performance to the case where vt-d is disabled. The main use-case for this
change is to decrease the time necessary to copy virtual machine memory
during KVM live migration instantiations.
This patch was tested on a system that enables the IOMMU in non-coherent
mode. Two DMA remapper issues were encountered in the previous version and
both patches have been committed.
commit ea2447f700
commit 2e12bc29fc
Signed-off-by: Craig Hada <[email protected]>
Signed-off-by: Sathya Perla <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
Use GET_PROFILE_CONFIG_V1 cmd for BE3-R, to query the maximum number of
TX rings available per function. On SH-R the same is queried via the
GET_FUNCTION_CONFIG cmd.
Signed-off-by: Vasundhara Volam <[email protected]>
Signed-off-by: Sathya Perla <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
Avoid flashing BE3 UFI on BE3-R chip by verifying asic_revision
number of the chip.
Signed-off-by: Vasundhara Volam <[email protected]>
Signed-off-by: Sathya Perla <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
Don't log "Out of MCCQ wrbs" msg. When the driver doesn't receive any
response from the FW it already logs a "FW not responding" message.
The driver runs out of MCCQ wrbs much later. Also, this message can
swamp the kernel log in HW/FW error scenarios.
Signed-off-by: Vasundhara Volam <[email protected]>
Signed-off-by: Sathya Perla <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
Skyhawk-R and BE3-R (SuperNIC profile) require V2 version
of TXQ_CREATE cmd to be used.
Signed-off-by: Vasundhara Volam <[email protected]>
Signed-off-by: Sathya Perla <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
We have blocked association attempts on interfaces configured in
AP and AD-HOC modes. P2P mode should be blocked too.
Furthermore, an error code must be returned if we are unable to
associate.
Signed-off-by: Bing Zhao <[email protected]>
Signed-off-by: John W. Linville <[email protected]>
priv->bss_mode uses NL80211_IFTYPE_* definitions.
HostCmd_BSS_MODE_IBSS is used in ad-hoc start/join command between
driver and firmware.
Coincidentally both HostCmd_BSS_MODE_IBSS and NL80211_IFTYPE_STATION
are defined as 2. That explains why nobody complained.
Signed-off-by: Bing Zhao <[email protected]>
Signed-off-by: John W. Linville <[email protected]>
There is a typo in mwifiex_cleanup_rxq_ring() which uses
PCI_DMA_TODEVICE while unmapping PCI memory.
We should actually use PCI_DMA_FROMDEVICE.
Signed-off-by: Avinash Patil <[email protected]>
Signed-off-by: Yogesh Ashok Powar <[email protected]>
Signed-off-by: Bing Zhao <[email protected]>
Signed-off-by: John W. Linville <[email protected]>
Send P2P_MODE_CFG cmd to firmware when p2p interface is created.
Without proper p2p configuration firmware may behave incorrectly
while handling commands sent through this interface.
Signed-off-by: Bing Zhao <[email protected]>
Signed-off-by: Stone Piao <[email protected]>
Signed-off-by: John W. Linville <[email protected]>
p2p devices should identify themselves as such to userspace at
startup, so the connection manager can decide which interface
to start wpa_supplicant instances on.
Signed-off-by: Paul Stewart <[email protected]>
Reviewed-by: Bing Zhao <[email protected]>
Signed-off-by: John W. Linville <[email protected]>
The function is used for BBP register initialization,
fix the typo in the function name to reflect that.
The patch contains no functional changes.
Signed-off-by: Gabor Juhos <[email protected]>
Acked-by: Gertjan van Wingerde <[email protected]>
Signed-off-by: John W. Linville <[email protected]>
The dereference to 'ifp' in debug code should be moved below the NULL test.
Signed-off-by: Wei Yongjun <[email protected]>
Acked-by: Arend van Spriel <[email protected]>
Signed-off-by: John W. Linville <[email protected]>
Helpers to fetch various fields from the Rx descriptor
Signed-off-by: Vladimir Kondratiev <[email protected]>
Signed-off-by: John W. Linville <[email protected]>
Original Tx descriptor stored is in non-cached area for DMA;
copy it to the cached memory to speed-up access
Signed-off-by: Vladimir Kondratiev <[email protected]>
Signed-off-by: John W. Linville <[email protected]>
vring size is known from the beginning, fill it immediately
in the struct initializer
This is minor optimization that reduces code size.
Signed-off-by: Vladimir Kondratiev <[email protected]>
Signed-off-by: John W. Linville <[email protected]>
Rx descriptors stored in non-cacheable memory area for DMA.
Non-cacheable memory causes long access time from CPU.
Copy rx descriptor to the skb->cb, and use this copy.
It provides faster memory access, and will be usefull to keep
Rx information for later processing (BACK reorder)
Signed-off-by: Vladimir Kondratiev <[email protected]>
Signed-off-by: John W. Linville <[email protected]>
In the dequeue worker the function brcmf_commit_skb() is called.
However, instead of increment the credit count upon success it
should break the for loop upon failure. Otherwise, it will result
in an endless loop.
Reviewed-by: Hante Meuleman <[email protected]>
Reviewed-by: Pieter-Paul Giesberts <[email protected]>
Signed-off-by: Arend van Spriel <[email protected]>
Signed-off-by: John W. Linville <[email protected]>
The function brcmf_rollback_toq() is already called in error path
and its result should not override the initial error value. As the
function releases the sk_buff there is no need to return anything
so change return type to void.
Reviewed-by: Hante Meuleman <[email protected]>
Reviewed-by: Pieter-Paul Giesberts <[email protected]>
Signed-off-by: Arend van Spriel <[email protected]>
Signed-off-by: John W. Linville <[email protected]>
All rollback failures should result in freeing of the sk_buff
by calling brcmf_txfinalize().
Reviewed-by: Hante Meuleman <[email protected]>
Reviewed-by: Pieter-Paul Giesberts <[email protected]>
Signed-off-by: Arend van Spriel <[email protected]>
Signed-off-by: John W. Linville <[email protected]>
When deleting an interface in firmware-signalling module it will
clear any destination descriptors. To avoid concurrency issues it
should take the lock using brcmf_fws_lock().
Reviewed-by: Hante Meuleman <[email protected]>
Reviewed-by: Pieter-Paul Giesberts <[email protected]>
Signed-off-by: Arend van Spriel <[email protected]>
Signed-off-by: John W. Linville <[email protected]>
Several firmware signals should be considered as opportunity to
send packets to the firmware. This patch adds conditional scheduling
of the dequeue worker thread while handling those signals.
Reviewed-by: Hante Meuleman <[email protected]>
Reviewed-by: Pieter-Paul Giesberts <[email protected]>
Signed-off-by: Arend van Spriel <[email protected]>
Signed-off-by: John W. Linville <[email protected]>
