Some firmwares in Huawei E3372H devices have been observed to switch back
to NTB 32-bit format after altsetting switch.
This patch implements a driver flag to check for the device settings and
set NTB format to 16-bit again if needed.
The flag has been activated for devices controlled by the huawei_cdc_ncm.c
driver.
V1->V2:
- fixed broken error checks
- some corrections to the commit message
V2->V3:
- variable name changes, to clarify what's happening
- check (and possibly set) the NTB format later in the common bind code path
Signed-off-by: Enrico Mioso <mrkiko.rs@gmail.com>
Reported-and-tested-by: Christian Panton <christian@panton.org>
Reviewed-by: Bjørn Mork <bjorn@mork.no>
CC: Bjørn Mork <bjorn@mork.no>
CC: Christian Panton <christian@panton.org>
CC: linux-usb@vger.kernel.org
CC: netdev@vger.kernel.org
CC: Oliver Neukum <oliver@neukum.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
The CDC-NCM driver can require large amounts of memory to create
skb's and this can be a problem when the memory becomes fragmented.
This especially affects embedded systems that have constrained
resources but wish to maximise the throughput of CDC-NCM with 16KiB
NTB's.
The issue is after running for a while the kernel memory can become
fragmented and it needs compacting.
If the NTB allocation is needed before the memory has been compacted
the atomic allocation can fail which can cause increased latency,
large re-transmissions or disconnections depending upon the data
being transmitted at the time.
This situation occurs for less than a second until the kernel has
compacted the memory but the failed devices can take a lot longer to
recover from the failed TX packets.
To ease this temporary situation I modified the CDC-NCM TX path to
temporarily switch into a reduced memory mode which allocates an NTB
that will fit into a USB_CDC_NCM_NTB_MIN_OUT_SIZE (default 2048 Bytes)
sized memory block and only transmit NTB's with a single network frame
until the memory situation is resolved.
Each time this issue occurs we wait for an increasing number of
reduced size allocations before requesting a full size one to not
put additional pressure on a low memory system.
Once the memory is compacted the CDC-NCM data can resume transmitting
at the normal tx_max rate once again.
Signed-off-by: Jim Baxter <jim_baxter@mentor.com>
Reviewed-by: Bjørn Mork <bjorn@mork.no>
Signed-off-by: David S. Miller <davem@davemloft.net>
Joe and Bjørn suggested that it'd be nicer to not have the
cast in the fairly common case of doing
*(u8 *)skb_put(skb, 1) = c;
Add skb_put_u8() for this case, and use it across the code,
using the following spatch:
@@
expression SKB, C, S;
typedef u8;
identifier fn = {skb_put};
fresh identifier fn2 = fn ## "_u8";
@@
- *(u8 *)fn(SKB, S) = C;
+ fn2(SKB, C);
Note that due to the "S", the spatch isn't perfect, it should
have checked that S is 1, but there's also places that use a
sizeof expression like sizeof(var) or sizeof(u8) etc. Turns
out that nobody ever did something like
*(u8 *)skb_put(skb, 2) = c;
which would be wrong anyway since the second byte wouldn't be
initialized.
Suggested-by: Joe Perches <joe@perches.com>
Suggested-by: Bjørn Mork <bjorn@mork.no>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
It seems like a historic accident that these return unsigned char *,
and in many places that means casts are required, more often than not.
Make these functions (skb_put, __skb_put and pskb_put) return void *
and remove all the casts across the tree, adding a (u8 *) cast only
where the unsigned char pointer was used directly, all done with the
following spatch:
@@
expression SKB, LEN;
typedef u8;
identifier fn = { skb_put, __skb_put };
@@
- *(fn(SKB, LEN))
+ *(u8 *)fn(SKB, LEN)
@@
expression E, SKB, LEN;
identifier fn = { skb_put, __skb_put };
type T;
@@
- E = ((T *)(fn(SKB, LEN)))
+ E = fn(SKB, LEN)
which actually doesn't cover pskb_put since there are only three
users overall.
A handful of stragglers were converted manually, notably a macro in
drivers/isdn/i4l/isdn_bsdcomp.c and, oddly enough, one of the many
instances in net/bluetooth/hci_sock.c. In the former file, I also
had to fix one whitespace problem spatch introduced.
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
A common pattern with skb_put() is to just want to memcpy()
some data into the new space, introduce skb_put_data() for
this.
An spatch similar to the one for skb_put_zero() converts many
of the places using it:
@@
identifier p, p2;
expression len, skb, data;
type t, t2;
@@
(
-p = skb_put(skb, len);
+p = skb_put_data(skb, data, len);
|
-p = (t)skb_put(skb, len);
+p = skb_put_data(skb, data, len);
)
(
p2 = (t2)p;
-memcpy(p2, data, len);
|
-memcpy(p, data, len);
)
@@
type t, t2;
identifier p, p2;
expression skb, data;
@@
t *p;
...
(
-p = skb_put(skb, sizeof(t));
+p = skb_put_data(skb, data, sizeof(t));
|
-p = (t *)skb_put(skb, sizeof(t));
+p = skb_put_data(skb, data, sizeof(t));
)
(
p2 = (t2)p;
-memcpy(p2, data, sizeof(*p));
|
-memcpy(p, data, sizeof(*p));
)
@@
expression skb, len, data;
@@
-memcpy(skb_put(skb, len), data, len);
+skb_put_data(skb, data, len);
(again, manually post-processed to retain some comments)
Reviewed-by: Stephen Hemminger <stephen@networkplumber.org>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
There were many places that my previous spatch didn't find,
as pointed out by yuan linyu in various patches.
The following spatch found many more and also removes the
now unnecessary casts:
@@
identifier p, p2;
expression len;
expression skb;
type t, t2;
@@
(
-p = skb_put(skb, len);
+p = skb_put_zero(skb, len);
|
-p = (t)skb_put(skb, len);
+p = skb_put_zero(skb, len);
)
... when != p
(
p2 = (t2)p;
-memset(p2, 0, len);
|
-memset(p, 0, len);
)
@@
type t, t2;
identifier p, p2;
expression skb;
@@
t *p;
...
(
-p = skb_put(skb, sizeof(t));
+p = skb_put_zero(skb, sizeof(t));
|
-p = (t *)skb_put(skb, sizeof(t));
+p = skb_put_zero(skb, sizeof(t));
)
... when != p
(
p2 = (t2)p;
-memset(p2, 0, sizeof(*p));
|
-memset(p, 0, sizeof(*p));
)
@@
expression skb, len;
@@
-memset(skb_put(skb, len), 0, len);
+skb_put_zero(skb, len);
Apply it to the tree (with one manual fixup to keep the
comment in vxlan.c, which spatch removed.)
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
The zero padding that is added to NTB's does
not zero the memory correctly.
This is because the skb_put modifies the value
of skb_out->len which results in the memset
command not setting any memory to zero as
(ctx->tx_max - skb_out->len) == 0.
I have resolved this by storing the size of
the memory to be zeroed before the skb_put
and using this in the memset call.
Signed-off-by: Jim Baxter <jim_baxter@mentor.com>
Reviewed-by: Bjørn Mork <bjorn@mork.no>
Signed-off-by: David S. Miller <davem@davemloft.net>
Add support for the net stats64 counters to the usbnet core. With that
in place put the hooks into every usbnet driver to use it.
This is a strait forward addition of 64bit counters for RX and TX packet
and byte counts. It is done in the same style as for the other net drivers
that support stats64. Note that the other stats fields remain as 32bit
sized values (error counts, etc).
The motivation to add this is that it is not particularly difficult to
get the RX and TX byte counts to wrap on 32bit platforms.
Signed-off-by: Greg Ungerer <gerg@linux-m68k.org>
Acked-by: Bjørn Mork <bjorn@mork.no>
Signed-off-by: David S. Miller <davem@davemloft.net>
The ethtool api {get|set}_settings is deprecated.
We move this driver to new api {get|set}_link_ksettings.
As I don't have the hardware, I'd be very pleased if
someone may test this patch.
Signed-off-by: Philippe Reynes <tremyfr@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
ktime_set(S,N) was required for the timespec storage type and is still
useful for situations where a Seconds and Nanoseconds part of a time value
needs to be converted. For anything where the Seconds argument is 0, this
is pointless and can be replaced with a simple assignment.
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Cc: Peter Zijlstra <peterz@infradead.org>
Telit LE922A MBIM based composition does not work properly
with altsetting toggle done in cdc_ncm_bind_common.
This patch adds CDC_MBIM_FLAG_AVOID_ALTSETTING_TOGGLE quirk
to avoid this procedure that, instead, is mandatory for
other modems.
Signed-off-by: Daniele Palmas <dnlplm@gmail.com>
Reviewed-by: Bjørn Mork <bjorn@mork.no>
Signed-off-by: David S. Miller <davem@davemloft.net>
usbnet:
- Remove stale new_mtu <= 0 check in usbnet.c
- Set min_mtu = 0, max_mtu = 65535 (sub-drivers must set their own
max_mtu and/or min_mtu as needed)
r8152:
- Set appropriate max_mtu for different variants (1500 or 9194)
lan78xx:
- Set max_mtu = 9000
asix_driver:
- max_mtu = 16384 for ax88178 variant
ax88179:
- max_mtu = 4088
cdc_ncm:
- max_mtu from hardware
cdc-phonet:
- min_mtu = 6, max_mtu = 65541
sierra_net:
- max_mtu = 1500, call usbnet_change_mtu directly
- sierra_net_change_mtu checked for MTU > 1500, then called
usbnet_change_mtu, but if we set max_mtu to let the network core handle
the range check, then we can simply call usbnet_change_mtu directly
smsc75xx:
- max_mtu = 9000
CC: netdev@vger.kernel.org
CC: Woojung Huh <woojung.huh@microchip.com>
CC: Microchip Linux Driver Support <UNGLinuxDriver@microchip.com>
CC: Hayes Wang <hayeswang@realtek.com>
CC: Oliver Neukum <oneukum@suse.com>
CC: Steve Glendinning <steve.glendinning@shawell.net>
Signed-off-by: Jarod Wilson <jarod@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Several Lenovo users have reported problems with their Sierra
Wireless EM7455 modem. The driver has loaded successfully and
the MBIM management channel has appeared to work, including
establishing a connection to the mobile network. But no frames
have been received over the data interface.
The problem affects all EM7455 and MC7455, and is assumed to
affect other modems based on the same Qualcomm chipset and
baseband firmware.
Testing narrowed the problem down to what seems to be a
firmware timing bug during initialization. Adding a short sleep
while probing is sufficient to make the problem disappear.
Experiments have shown that 1-2 ms is too little to have any
effect, while 10-20 ms is enough to reliably succeed.
Reported-by: Stefan Armbruster <ml001@armbruster-it.de>
Reported-by: Ralph Plawetzki <ralph@purejava.org>
Reported-by: Andreas Fett <andreas.fett@secunet.com>
Reported-by: Rasmus Lerdorf <rasmus@lerdorf.com>
Reported-by: Samo Ratnik <samo.ratnik@gmail.com>
Reported-and-tested-by: Aleksander Morgado <aleksander@aleksander.es>
Signed-off-by: Bjørn Mork <bjorn@mork.no>
Signed-off-by: David S. Miller <davem@davemloft.net>
Current implementation updates the mtu size and notify cdc_ncm
device using USB_CDC_SET_MAX_DATAGRAM_SIZE request about datagram
size change instead of changing rx_urb_size.
Whenever mtu is being changed, datagram size should also be
updated. Also updating maxmtu formula so it takes max_datagram_size with
use of cdc_ncm_max_dgram_size() and not ctx.
Signed-off-by: Robert Dobrowolski <robert.dobrowolski@linux.intel.com>
Signed-off-by: Rafal Redzimski <rafal.f.redzimski@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Telit LE910 V2 is a mobile broadband card with no ARP capabilities:
the patch makes this device to use wwan_noarp_info struct
Signed-off-by: Daniele Palmas <dnlplm@gmail.com>
Reviewed-by: Bjørn Mork <bjorn@mork.no>
Signed-off-by: David S. Miller <davem@davemloft.net>
usbnet_link_change will call schedule_work and should be
avoided if bind is failing. Otherwise we will end up with
scheduled work referring to a netdev which has gone away.
Instead of making the call conditional, we can just defer
it to usbnet_probe, using the driver_info flag made for
this purpose.
Fixes: 8a34b0ae87 ("usbnet: cdc_ncm: apply usbnet_link_change")
Reported-by: Andrey Konovalov <andreyknvl@gmail.com>
Suggested-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Bjørn Mork <bjorn@mork.no>
Signed-off-by: David S. Miller <davem@davemloft.net>
Some devices will silently fail setup unless they are reset first.
This is necessary even if the data interface is already in
altsetting 0, which it will be when the device is probed for the
first time. Briefly toggling the altsetting forces a function
reset regardless of the initial state.
This fixes a setup problem observed on a number of Huawei devices,
appearing to operate in NTB-32 mode even if we explicitly set them
to NTB-16 mode.
Signed-off-by: Bjørn Mork <bjorn@mork.no>
Signed-off-by: David S. Miller <davem@davemloft.net>
NCM buffer sizes are negotiated with the device independently of
the network device MTU. The RX buffers are allocated by the
usbnet framework based on the rx_urb_size value set by cdc_ncm. A
single RX buffer can hold a number of MTU sized packets.
The default usbnet change_mtu ndo only modifies rx_urb_size if it
is equal to hard_mtu. And the cdc_ncm driver will set rx_urb_size
and hard_mtu independently of each other, based on dwNtbInMaxSize
and dwNtbOutMaxSize respectively. It was therefore assumed that
usbnet_change_mtu() would never touch rx_urb_size. This failed to
consider the case where dwNtbInMaxSize and dwNtbOutMaxSize happens
to be equal.
Fix by implementing an NCM specific change_mtu ndo, modifying the
netdev MTU without touching the buffer size settings.
Signed-off-by: Bjørn Mork <bjorn@mork.no>
Signed-off-by: David S. Miller <davem@davemloft.net>
Unlike DW5550, Dell DW5813 is a mobile broadband card with no ARP
capabilities: the patch makes this device to use wwan_noarp_info struct
Signed-off-by: Daniele Palmas <dnlplm@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Unlike DW5550, Dell DW5812 is a mobile broadband card with no ARP
capabilities: the patch makes this device to use wwan_noarp_info struct
Signed-off-by: Daniele Palmas <dnlplm@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Conflicts:
drivers/net/geneve.c
Here we had an overlapping change, where in 'net' the extraneous stats
bump was being removed whilst in 'net-next' the final argument to
udp_tunnel6_xmit_skb() was being changed.
Signed-off-by: David S. Miller <davem@davemloft.net>
Adding a writable sysfs attribute for the "NDP to end"
quirk flag.
This makes it easier for end users to test new devices for
this firmware bug. We've been lucky so far, but we should
not depend on reporters capable of rebuilding the driver.
Cc: Enrico Mioso <mrkiko.rs@gmail.com>
Signed-off-by: Bjørn Mork <bjorn@mork.no>
Signed-off-by: David S. Miller <davem@davemloft.net>
The Huawei E3372 (12d1:157d) needs this quirk in MBIM mode
as well. Allow this by forcing the NTB to contain only a
single NDP, and add a device specific entry for this ID.
Due to the way Huawei use device IDs, this might be applied
to other modems as well. It is assumed that those modems
will be based on the same firmware and will need this quirk
too. If not, it will still not harm normal usage, although
multiplexing performance could be impacted.
Cc: Enrico Mioso <mrkiko.rs@gmail.com>
Reported-by: Sami Farin <hvtaifwkbgefbaei@gmail.com>
Signed-off-by: Bjørn Mork <bjorn@mork.no>
Acked-By: Enrico Mioso <mrkiko.rs@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Commit 77b0a09967 ("cdc-ncm: use common parser") added a dangerous
new trust in the CDC functional descriptors presented by the device,
unconditionally assuming that any device handled by the driver has
a CDC Union descriptor.
This descriptor is required by the NCM and MBIM specs, but crashing
on non-compliant devices is still unacceptable. Not only will that
allow malicious devices to crash the kernel, but in this case it is
also well known that there are non-compliant real devices on the
market - as shown by the comment accompanying the IAD workaround
in the same function.
The Sierra Wireless EM7305 is an example of such device, having
a CDC header and a CDC MBIM descriptor but no CDC Union:
Interface Descriptor:
bLength 9
bDescriptorType 4
bInterfaceNumber 12
bAlternateSetting 0
bNumEndpoints 1
bInterfaceClass 2 Communications
bInterfaceSubClass 14
bInterfaceProtocol 0
iInterface 0
CDC Header:
bcdCDC 1.10
CDC MBIM:
bcdMBIMVersion 1.00
wMaxControlMessage 4096
bNumberFilters 16
bMaxFilterSize 128
wMaxSegmentSize 4064
bmNetworkCapabilities 0x20
8-byte ntb input size
Endpoint Descriptor:
..
The conversion to a common parser also left the local cdc_union
variable untouched. This caused the IAD workaround code to be applied
to all devices with an IAD descriptor, which was never intended. Finish
the conversion by testing for hdr.usb_cdc_union_desc instead.
Cc: Oliver Neukum <oneukum@suse.com>
Fixes: 77b0a09967 ("cdc-ncm: use common parser")
Signed-off-by: Bjørn Mork <bjorn@mork.no>
Signed-off-by: David S. Miller <davem@davemloft.net>
This moves cdc-ncm to the common parser for CDC user
to reduce code duplication.
Signed-off-by: Oliver Neukum <oneukum@suse.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Update referenced specs link to reflect actual file version and location.
Signed-off-by: Enrico Mioso <mrkiko.rs@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
NCM specs are not actually mandating a specific position in the frame for
the NDP (Network Datagram Pointer). However, some Huawei devices will
ignore our aggregates if it is not placed after the datagrams it points
to. Add support for doing just this, in a per-device configurable way.
While at it, update NCM subdrivers, disabling this functionality in all of
them, except in huawei_cdc_ncm where it is enabled instead.
We aren't making any distinction between different Huawei NCM devices,
based on what the vendor driver does. Standard NCM devices are left
unaffected: if they are compliant, they should be always usable, still
stay on the safe side.
This change has been tested and working with a Huawei E3131 device (which
works regardless of NDP position), a Huawei E3531 (also working both
ways) and an E3372 (which mandates NDP to be after indexed datagrams).
V1->V2:
- corrected wrong NDP acronym definition
- fixed possible NULL pointer dereference
- patch cleanup
V2->V3:
- Properly account for the NDP size when writing new packets to SKB
Signed-off-by: Enrico Mioso <mrkiko.rs@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
The tx_curr_frame_payload field is u32. When we try to calculate a
small negative delta based on it, we end up with a positive integer
close to 2^32 instead. So the tx_bytes pointer increases by about
2^32 for every transmitted frame.
Fix by calculating the delta as a signed long.
Cc: Ben Hutchings <ben.hutchings@codethink.co.uk>
Reported-by: Florian Bruhin <me@the-compiler.org>
Fixes: 7a1e890e21 ("usbnet: Fix tx_bytes statistic running backward in cdc_ncm")
Signed-off-by: Bjørn Mork <bjorn@mork.no>
Signed-off-by: David S. Miller <davem@davemloft.net>
cdc_ncm disagrees with usbnet about how much framing overhead should
be counted in the tx_bytes statistics, and tries 'fix' this by
decrementing tx_bytes on the transmit path. But statistics must never
be decremented except due to roll-over; this will thoroughly confuse
user-space. Also, tx_bytes is only incremented by usbnet in the
completion path.
Fix this by requiring drivers that set FLAG_MULTI_FRAME to set a
tx_bytes delta along with the tx_packets count.
Fixes: beeecd42c3 ("net: cdc_ncm/cdc_mbim: adding NCM protocol statistics")
Signed-off-by: Ben Hutchings <ben.hutchings@codethink.co.uk>
Signed-off-by: Bjørn Mork <bjorn@mork.no>
Currently the usbnet core does not update the tx_packets statistic for
drivers with FLAG_MULTI_PACKET and there is no hook in the TX
completion path where they could do this.
cdc_ncm and dependent drivers are bumping tx_packets stat on the
transmit path while asix and sr9800 aren't updating it at all.
Add a packet count in struct skb_data so these drivers can fill it
in, initialise it to 1 for other drivers, and add the packet count
to the tx_packets statistic on completion.
Signed-off-by: Ben Hutchings <ben.hutchings@codethink.co.uk>
Tested-by: Bjørn Mork <bjorn@mork.no>
Signed-off-by: David S. Miller <davem@davemloft.net>
The min_tx_pkt variable decides the cutoff point where the driver
will stop padding out NTBs to maximum size. The padding is a tradeoff
where we use some USB bus bandwidth to allow the device to receive
fixed size buffers. Different devices will have different optimal
settings, spanning from no padding at all to padding every NTB.
There is no way to automatically figure out which setting is best
for a specific device.
The default value is a reasonable tradeoff, calculated based on the
USB packet size and out NTB max size. This may have to be changed
along with any tx_max changes.
Signed-off-by: Bjørn Mork <bjorn@mork.no>
Signed-off-by: David S. Miller <davem@davemloft.net>
The mandatory GetNtbParameters control request is an important part of
the host <-> device protocol negotiation in CDC NCM (and CDC MBIM). It
gives device limits which the host must obey when configuring the
protocol aggregation variables. The driver will enforce this by
rejecting attempts to set any of the tunable variables to a value
which is not supported by the device. Exporting the parameter block
helps userspace decide which values are allowed without resorting
to trial and error.
Signed-off-by: Bjørn Mork <bjorn@mork.no>
Signed-off-by: David S. Miller <davem@davemloft.net>
The ethtool coalesce API is not applicable for this driver. Forcing
it to fit the NCM aggregation redefined the API in a driver specific
way, which is much worse than defining a clean new API. These ethtool
coalesce functions have therefore been replaced by a new sysfs API.
Signed-off-by: Bjørn Mork <bjorn@mork.no>
Signed-off-by: David S. Miller <davem@davemloft.net>
Attach a driver specific sysfs group to the netdev, and use it
for the rx/tx aggregation variables.
The datagram aggregation defined by the CDC NCM specification is
specific to this device class (including CDC MBIM). Using the
ethtool interrupt coalesce API as an interface to the aggregation
parameters redefined that API in a driver specific and confusing
way. A sysfs group
- makes it clear that this is a driver specific userspace API, and
- allows us to export the real values instead of some translated
version, and
- lets us include more aggregation variables which were impossible
to force into the ethtool API.
Additionally, using sysfs allows tuning the driver on space
constrained hosts where userspace tools like ethtool are undesired.
Suggested-by: Peter Stuge <peter@stuge.se>
Signed-off-by: Bjørn Mork <bjorn@mork.no>
Signed-off-by: David S. Miller <davem@davemloft.net>
It doesn't matter whether the buffer size goes up or down. We have to
keep usbnet and device syncronized to be able to split transfers at the
correct boundaries. The spec allow skipping short packets when using
max sized transfers. If we don't tell usbnet about our new expected rx
buffer size, then it will merge and/or split NTBs. The driver does not
support this, and the result will be lots of framing errors.
Fix by always reallocating usbnet rx buffers when the rx_max value
changes.
Fixes: 68864abf08 ("net: cdc_ncm: support rx_max/tx_max updates when running")
Signed-off-by: Bjørn Mork <bjorn@mork.no>
Signed-off-by: David S. Miller <davem@davemloft.net>
We are calling usbnet_start_xmit() to flush any remaining data,
depending on the side effect that tx_curr_skb is set to NULL,
ensuring a new allocation using the updated tx_max. But this
side effect will only happen if there were any cached data ready
to transmit. If not, then an empty tx_curr_skb is still allocated
using the old tx_max size. Free it to avoid a buffer overrun.
Fixes: 68864abf08 ("net: cdc_ncm: support rx_max/tx_max updates when running")
Signed-off-by: Bjørn Mork <bjorn@mork.no>
Signed-off-by: David S. Miller <davem@davemloft.net>
Cloning the big skbs we use for USB buffering chokes up TCP and
SCTP because the socket memory limits are hitting earlier than
they should. It is better to unconditionally copy the unwrapped
packets to freshly allocated skbs.
Reported-by: Jim Baxter <jim_baxter@mentor.com>
Acked-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: Bjørn Mork <bjorn@mork.no>
Signed-off-by: David S. Miller <davem@davemloft.net>
There is a typo here where we test for USB_CDC_NCM_NTH32_SIGN instead
of USB_CDC_NCM_NTB32_SUPPORTED. The test probably still works as
written because 0x686D636E has (1 << 1) set and doesn't have (1 << 0)
set.
Fixes: f8afb73da3 ('net: cdc_ncm: factor out one-time device initialization')
Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
The upper timer_interval limit is arbitrary and much higher
than anything usable in the real world. Reducing it from 15s
to ~4s to make the timer_interval fit in an u32 does not make
much difference. The limit is still outside the practical
bounds.
This eliminates the need for a 64bit timer_interval, fixing a
build error related to 64bit division:
drivers/built-in.o: In function `cdc_ncm_get_coalesce':
ak8975.c:(.text+0x1ac994): undefined reference to `__aeabi_uldivmod'
Reported-by: Stephen Rothwell <sfr@canb.auug.org.au>
Signed-off-by: Bjørn Mork <bjorn@mork.no>
Signed-off-by: David S. Miller <davem@davemloft.net>
We can end up with a freshly allocated tx_curr_skb with no frames
in it. In this case it does not make any sense to start the timer.
This avoids the timer periodically trying to start tx when there
is nothing in the queue.
Signed-off-by: Bjørn Mork <bjorn@mork.no>
Signed-off-by: David S. Miller <davem@davemloft.net>
Calling netif_carrier_{on,off} is sufficient. There is no need
to duplicate the carrier state in a driver specific flag.
Acked-by: Enrico Mioso <mrkiko.rs@gmail.com>
Signed-off-by: Bjørn Mork <bjorn@mork.no>
Signed-off-by: David S. Miller <davem@davemloft.net>
Lots of devices request much larger buffers than reasonable. This
cause real problems for users of hosts with limited resources.
Reducing the default buffer size to 16kB for such devices is
a reasonable trade-off between allowing them to aggregate traffic
and avoiding memory exhaustion on resource restrained hosts.
Signed-off-by: Bjørn Mork <bjorn@mork.no>
Signed-off-by: David S. Miller <davem@davemloft.net>
To have an idea of the effects of the protocol coalescing
it's useful to have some counters showing the different
aspects.
Due to the asymmetrical usbnet interface the netdev
rx_bytes counter has been counting real received payload,
while the tx_bytes counter has included the NCM/MBIM
framing overhead. This overhead can be many times the
payload because of the aggressive padding strategy of
this driver, and will vary a lot depending on device
and traffic.
With very few exceptions, users are only interested in
the payload size. Having an somewhat accurate payload
byte counter is particularly important for mobile
broadband devices, which many NCM devices and of course
all MBIM devices are. Users and userspace applications
will use this counter to monitor account quotas.
Having protocol specific counters for the overhead, we are
now able to correct the tx_bytes netdev counter so that
it shows the real payload
Signed-off-by: Bjørn Mork <bjorn@mork.no>
Signed-off-by: David S. Miller <davem@davemloft.net>
We pad frames larger than X to maximum size for devices which
don't need a ZLP after maximum sized frames. This allows the
device to optimize its transfers for one fixed buffer size.
X was arbitrarily set at 512 bytes regardless of real buffer
maximum, causing extreme overheads due to excessive padding of
larger tx buffers. Limit the padding to at most 3 full USB
packets, still allowing the overhead to payload ratio of 3/1.
Signed-off-by: Bjørn Mork <bjorn@mork.no>
Signed-off-by: David S. Miller <davem@davemloft.net>
Many newer NCM and MBIM devices will request a maximum tx
datagram count which is much smaller than our hard-coded
absolute max. We can reduce the overhead without sacrificing
any of the simplicity for these devices, by simply using the
true negotiated count in when calculated the maximum NTH and
NDP header sizes.
Signed-off-by: Bjørn Mork <bjorn@mork.no>
Signed-off-by: David S. Miller <davem@davemloft.net>
Datagram coalescing is an integral part of the NCM and MBIM
protocols, intended to reduce the interrupt load primarily
on the device end of the USB link. As with all coalescing
solutions, there is a trade-off between buffering and
interrupts.
The current defaults are based on the assumption that device
side buffers should be the limiting factor. However, many
modern high speed LTE modems suffers from buffer-bloat,
making this assumption fail. This results in sub-optimal
performance due to excessive coalescing. And in cases where
such modems are connected to cheap embedded hosts there is
often severe buffer allocation issues, giving very noticeable
performance degradation .
A start on improving this is going from build time hard
coded limits to per device user configurable limits. The
ethtool coalescing API was selected as user interface
because, although the tuned values are buffer sizes, these
settings directly control datagram coalescing.
Signed-off-by: Bjørn Mork <bjorn@mork.no>
Signed-off-by: David S. Miller <davem@davemloft.net>
