Current ConnectX HW is unable to perform VLAN pop in TX path and VLAN
push on RX path. As a workaround, untagged packets are tagged with
VID 0x000 allowing pop/push actions to be exchanged with VLAN rewrite
actions.
Use the ingress ACL table, preceding the FDB, to push VLAN 0x000 ID tag
for untagged packets and the egress ACL table, succeeding the FDB, to
pop VLAN 0x000 ID tag.
Signed-off-by: Eli Britstein <elibr@mellanox.com>
Reviewed-by: Oz Shlomo <ozsh@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
Hardware requires that all TIRs that steer traffic to the same RQ
should share identical tunneled_offload_en value.
For that, the tunneled_offload_en bit should be set/unset (according to
the HW capability) for all TIRs', not only the ones dedicated for
tunneled (inner) traffic.
Fixes: 1b223dd391 ("net/mlx5e: Fix checksum handling for non-stripped vlan packets")
Signed-off-by: Tariq Toukan <tariqt@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
Many TIR context settings are common to different TIR types,
take them into a common function.
Signed-off-by: Tariq Toukan <tariqt@mellanox.com>
Reviewed-by: Aya Levin <ayal@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
It is a followup after the fix in
commit 9c69a13205 ("route: Avoid crash from dereferencing NULL rt->from")
rt6_do_redirect():
1. NULL checking is needed on rt->from because a parallel
fib6_info delete could happen that sets rt->from to NULL.
(e.g. rt6_remove_exception() and fib6_drop_pcpu_from()).
2. fib6_info_hold() is not enough. Same reason as (1).
Meaning, holding dst->__refcnt cannot ensure
rt->from is not NULL or rt->from->fib6_ref is not 0.
Instead of using fib6_info_hold_safe() which ip6_rt_cache_alloc()
is already doing, this patch chooses to extend the rcu section
to keep "from" dereference-able after checking for NULL.
inet6_rtm_getroute():
1. NULL checking is also needed on rt->from for a similar reason.
Note that inet6_rtm_getroute() is using RTNL_FLAG_DOIT_UNLOCKED.
Fixes: a68886a691 ("net/ipv6: Make from in rt6_info rcu protected")
Signed-off-by: Martin KaFai Lau <kafai@fb.com>
Acked-by: Wei Wang <weiwan@google.com>
Reviewed-by: David Ahern <dsahern@gmail.com>
Reviewed-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
While the endiannes is being handled correctly as indicated by the comment
above the offending line - sparse was unhappy with the missing annotation
as be64_to_cpu() expects a __be64 argument. To mitigate this annotation
all involved variables are changed to a consistent __le64 and the
conversion to uint64_t delayed to the call to rds_cong_map_updated().
Signed-off-by: Nicholas Mc Guire <hofrat@osadl.org>
Acked-by: Santosh Shilimkar <santosh.shilimkar@oracle.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Maxime Chevallier says:
====================
net: mvpp2: cls: Add classification
This series is a rework of the previously standalone patch adding
classification support for mvpp2 :
https://lore.kernel.org/netdev/20190423075031.26074-1-maxime.chevallier@bootlin.com/
This patch has been reworked according to Saeed's review, to make sure
that the location of the rule is always respected and serves as a way to
prioritize rules between each other. This the 3rd iteration of this
submission, but since it's now a series, I reset the revision numbering.
This series implements that in a limited configuration for now, since we
limit the total number of rules per port to 4.
The main factors for this limitation are that :
- We share the classification tables between all ports (4 max, although
one is only used for internal loopback), hence we have to perform a
logical separation between rules, which is done today by dedicated
ranges for each port in each table
- The "Flow table", which dictates which lookups operations are
performed for an ingress packet, in subdivided into 22 "sub flows",
each corresponding to a traffic type based on the L3 proto, L4
proto, the presence or not of a VLAN tag and the L3 fragmentation.
This makes so that when adding a rule, it has to be added into each
of these subflows, introducing duplications of entries and limiting
our max number of entries.
These limitations can be overcomed in several ways, but for readability
sake, I'd rather submit basic classification offload support for now,
and improve it gradually.
This series also adds a small cosmetic cleanup patch (1), and also adds
support for the "Drop" action compared to the first submission of this
feature. It is simple enough to be added with this basic support.
Compared to the first submissions, the NETIF_F_NTUPLE flag was also
removed, following Saeed's comment.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
This commit introduces support for the "Drop" action in classification
offload. This corresponds to the "-1" action with ethtool -N.
This is achieved using the color marking actions available in the C2
engine, which associate a color to a packet. These colors can be either
Green, Yellow or Red, Red meaning that the packet should be dropped.
Green and Yellow colors are interpreted by the Policer, which isn't
supported yet.
This method of dropping using the Classifier is different than the
already existing early-drop features, such as VLAN filtering and MAC
UC/MC filtering, which are performed during the Parsing step, and
therefore take precedence over classification actions.
Signed-off-by: Maxime Chevallier <maxime.chevallier@bootlin.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
This commit introduces basic classification offloading support for the
PPv2 controller.
The PPv2 classifier has many classification engines, for now we only use
the C2 TCAM match engine.
This engine allows to perform ternary lookups on 64 bits keys (called
Header Extracted Key), that are built by extracting fields from the packet
header and concatenating them. At most 4 fields can be extracted for a
single lookup.
This basic implementation allows to build the HEK from the following
fields :
- L4 source and destination ports (for UDP and TCP)
More fields are to be added in the future.
Classification flows are added through the ethtool interface, using the
newly introduced flow_rule infrastructure as an internal rule
representation, allowing to more easily implement tc flower rules if
need be.
The internal design for now allocates one range of 4 rules per port
due to the internal design of the flow table, which uses 22 sub-flows.
When inserting a classification rule, the rule is created in every
relevant sub-flow.
This low rule-count is a very simple design which reaches quickly the
limitations of the flow table ordering, but guarantees that the rule
ordering will always be respected.
This commit only introduces support for the "steer to rxq" action.
Signed-off-by: Maxime Chevallier <maxime.chevallier@bootlin.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
As of today, the classification code is used only for RSS. We split the
incoming traffic into multiple flows, that correspond to the ethtool
flow_type parameter.
We don't want to use the ethtool flow definitions such as TCP_V4_FLOW,
for several reason :
- We want to decorrelate the driver code from ethtool as much as
possible, so that we can easily use other interfaces such as tc flower,
- We want the flow_type to be a bitfield, so that we can match flows
embedded into each other, such as TCP4 which is a subset of IP4.
This commit does the conversion to the newer type.
Signed-off-by: Maxime Chevallier <maxime.chevallier@bootlin.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Cosmetic patch removing extra whitespaces when writing the flow_table
entries
Signed-off-by: Maxime Chevallier <maxime.chevallier@bootlin.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
This merge commit includes some misc shared code updates from mlx5-next branch needed
for net-next.
1) From Aya: Enable general events on all physical link types and
restrict general event handling of subtype DELAY_DROP_TIMEOUT in mlx5 rdma
driver to ethernet links only as it was intended.
2) From Eli: Introduce low level bits for prio tag mode
3) From Maor: Low level steering updates to support RDMA RX flow
steering and enables RoCE loopback traffic when switchdev is enabled.
4) From Vu and Parav: Two small mlx5 core cleanups
5) From Yevgeny add HW definitions of geneve offloads
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
- regression in memset if line size !64
- avoid panic if PAE and IOC
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1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=Qm+2
-----END PGP SIGNATURE-----
Merge tag 'arc-5.1-final' of git://git.kernel.org/pub/scm/linux/kernel/git/vgupta/arc
Pull ARC fixes from Vineet Gupta:
"A few minor fixes for ARC.
- regression in memset if line size !64
- avoid panic if PAE and IOC"
* tag 'arc-5.1-final' of git://git.kernel.org/pub/scm/linux/kernel/git/vgupta/arc:
ARC: memset: fix build with L1_CACHE_SHIFT != 6
ARC: [hsdk] Make it easier to add PAE40 region to DTB
ARC: PAE40: don't panic and instead turn off hw ioc
The cited commit broke the offsets of hca cap struct, fix it.
While at it, cleanup a white space introduced by the same commit.
Fixes: b169e64a24 ("net/mlx5: Geneve, Add flow table capabilities for Geneve decap with TLV options")
Reported-by: Qian Cai <cai@lca.pw>
Cc: Yevgeny Kliteynik <kliteyn@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
The Interrupt Message Number in the PCIe Capabilities register (PCIe r4.0,
sec 7.5.3.2) indicates which MSI/MSI-X vector is shared by interrupts
related to the PCIe Capability, including Link Bandwidth Management and
Link Autonomous Bandwidth Interrupts (Link Control, 7.5.3.7), Command
Completed and Hot-Plug Interrupts (Slot Control, 7.5.3.10), and the PME
Interrupt (Root Control, 7.5.3.12).
pcie_message_numbers() checked whether we want to enable PME or Hot-Plug
interrupts but neglected to check for Link Bandwidth Management, so if we
only wanted the Bandwidth Management interrupts, it decided we didn't need
any vectors at all. Then pcie_port_enable_irq_vec() tried to reallocate
zero vectors, which failed, resulting in fallback to INTx.
On some systems, e.g., an X79-based workstation, that INTx seems broken or
not handled correctly, so we got spurious IRQ16 interrupts for Bandwidth
Management events.
Change pcie_message_numbers() so that if we want Link Bandwidth Management
interrupts, we use the shared MSI/MSI-X vector from the PCIe Capabilities
register.
Fixes: e8303bb7a7 ("PCI/LINK: Report degraded links via link bandwidth notification")
Link: https://lore.kernel.org/lkml/155597243666.19387.1205950870601742062.stgit@gimli.home
Signed-off-by: Alex Williamson <alex.williamson@redhat.com>
[bhelgaas: changelog]
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Revert a recent ACPICA change that caused initialization to fail on
systems with Thunderbolt docking stations connected at the init time.
-----BEGIN PGP SIGNATURE-----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=2XoK
-----END PGP SIGNATURE-----
Merge tag 'acpi-5.1-rc8' of git://git.kernel.org/pub/scm/linux/kernel/git/rafael/linux-pm
Pull ACPI fix from Rafael Wysocki:
"Revert a recent ACPICA change that caused initialization to fail on
systems with Thunderbolt docking stations connected at the init time"
* tag 'acpi-5.1-rc8' of git://git.kernel.org/pub/scm/linux/kernel/git/rafael/linux-pm:
Revert "ACPICA: Clear status of GPEs before enabling them"
The 'extent_type' variable does seem to be reliably initialized, but
it's _very_ non-obvious, since there's a "goto next" case that jumps
over the normal initialization. That will then always trigger the
"start >= extent_end" test, which will end up never falling through to
the use of that variable.
But the code is certainly not obvious, and the compiler warning looks
reasonable. Make 'extent_type' an int, and initialize it to an invalid
negative value, which seems to be the common pattern in other places.
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Esben Haabendal says:
====================
net: ll_temac: x86_64 support
This patch series adds support for use of ll_temac driver with
platform_data configuration and fixes endianess and 64-bit problems so
that it can be used on x86_64 platform.
A few bugfixes are also included.
Changes since v2:
- Fixed lp->indirect_mutex initialization regression for OF
platforms introduced in v2
Changes since v1:
- Make indirect_mutex specification mandatory when using platform_data
- Move header to include/linux/platform_data
- Enable COMPILE_TEST for XILINX_LL_TEMAC
- Rebased to v5.1-rc7
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
As soon as TAILDESCR_PTR is written, DMA transfers might start.
Let's ensure we are ready to receive DMA IRQ's before doing that.
Signed-off-by: Esben Haabendal <esben@geanix.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
This allows custom setup of IRQ coalescing for platforms using legacy
platform_device. The irq timeout and count parameters can be used for
tuning cpu load vs. latency.
I have maintained the 0x00000400 bit in TX_CHNL_CTRL. It is specified as
unused in the documentation I have available. It does not make any
difference in the hardware I have available, so it is left in to not risk
breaking other platforms where it might be used.
Signed-off-by: Esben Haabendal <esben@geanix.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Use usleep_range() to avoid problems with msleep() actually sleeping
much longer than expected.
Signed-off-by: Esben Haabendal <esben@geanix.com>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: David S. Miller <davem@davemloft.net>
As we are actually using a BD for both the skb and each frag contained in
it, the oldest TX BD would be overwritten when there was exactly one BD
less than needed.
Signed-off-by: Esben Haabendal <esben@geanix.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Unmap the actual buffer length, not the amount of data received, avoiding
resource exhaustion of swiotlb (seen on x86_64 platform).
Signed-off-by: Esben Haabendal <esben@geanix.com>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: David S. Miller <davem@davemloft.net>
Indirect register access goes through a DCR bus bridge, which
allows only one outstanding transaction. And to make matters
worse, each TEMAC IP block contains two Ethernet interfaces, and
although they seem to have separate registers for indirect access,
they actually share the registers. Or to be more specific, MSW, LSW
and CTL registers are physically shared between Ethernet interfaces
in same TEMAC IP, with RDY register being (almost) specificic to
the Ethernet interface. The 0x10000 bit in RDY reflects combined
bus ready state though.
So we need to take care to synchronize not only within a single
device, but also between devices in same TEMAC IP.
This commit allows to do that with legacy platform devices.
For OF devices, the xlnx,compound parent of the temac node should be
used to find siblings, and setup a shared indirect_mutex between them.
I will leave this work to somebody else, as I don't have hardware to
test that. No regression is introduced by that, as before this commit
using two Ethernet interfaces in same TEMAC block is simply broken.
Signed-off-by: Esben Haabendal <esben@geanix.com>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: David S. Miller <davem@davemloft.net>
With little-endian and 64-bit support in place, the ll_temac driver can
now be used on x86 and x86_64 platforms.
And while at it, enable COMPILE_TEST also.
Signed-off-by: Esben Haabendal <esben@geanix.com>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: David S. Miller <davem@davemloft.net>
Both TEMAC and SDMA is big-endian, so make sure that all values in SDMA
buffer descriptors (cmdac_bd) are handled as big-endian, independent of the
host endianness. With all currently supported platforms being big-endian,
this change does not make a change for any of them.
Note, when using app3 and app4 for piggybacking skb pointers there is no
need to care about endianness, as neither TEMAC nor SDMA access app3 and
app4 in TX buffer descriptors.
Signed-off-by: Esben Haabendal <esben@geanix.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Replace the powerpc specific MMIO register access functions with the
generic big-endian mmio access functions, and add support for
little-endian access depending on configuration.
Big-endian access is maintained as the default, but little-endian can
be configured in device-tree binding or in platform data.
The temac_ior()/temac_iow() functions are replaced with macro wrappers
to avoid modifying existing code more than necessary.
Signed-off-by: Esben Haabendal <esben@geanix.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
The use of buffer descriptor APP4 field (32-bit) for storing skb pointer
obviously does not work on 64-bit platforms.
As APP3 is also unused, we can use that to store the other half of 64-bit
pointer values.
Contrary to what is hinted at in commit message of commit 15bfe05c8d
("net: ethernet: xilinx: Mark XILINX_LL_TEMAC broken on 64-bit")
there are no other pointers stored in cdmac_bd.
Signed-off-by: Esben Haabendal <esben@geanix.com>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: David S. Miller <davem@davemloft.net>
Support initialization with platdata, so the driver can be used on
non-device-tree platforms.
For currently supported device-tree platforms, the driver should behave
as before.
Signed-off-by: Esben Haabendal <esben@geanix.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
As a side effect, a few error cases are fixed.
If of_iomap() of sdma_regs failed, no error code was returned. Fixed to
return -ENOMEM similar to of_iomap() fail of regs.
If sysfs_create_group() or register_netdev() failed, lp->phy_node was not
released.
Finally, the order in remove function is corrected to be reverse order
of what is done in probe, i.e. calling temac_mdio_teardown() last, so we
unregister the netdev that most likely is using the mdio_bus first.
Signed-off-by: Esben Haabendal <esben@geanix.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Fixes: 65b2b4939a ("selftests: net: initial fib rule tests")
Signed-off-by: Hangbin Liu <liuhangbin@gmail.com>
Reviewed-by: David Ahern <dsahern@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Fix inconsistent IS_ERR and PTR_ERR in cpsw_probe,
The proper pointer to use is clk instead of mode.
This issue was detected with the help of Coccinelle.
Fixes: 83a8471ba2 ("net: ethernet: ti: cpsw: refactor probe to group common hw initialization")
Signed-off-by: YueHaibing <yuehaibing@huawei.com>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: David S. Miller <davem@davemloft.net>
The pvlock_page and hvclock_page variables are (as the name implies)
addresses to pages, created by the linker script.
But we declared them as just "extern u8" variables, which _works_, but
now that gcc does some more bounds checking, it causes warnings like
warning: array subscript 1 is outside array bounds of ‘u8[1]’
when we then access more than one byte from those variables.
Fix this by simply making the declaration of the variables match
reality, which makes the compiler happy too.
Signed-off-by: Linus Torvalds <torvalds@-linux-foundation.org>
I'm not sure what made gcc warn about this code now. The 'ret' variable
does end up initialized in all cases, but it's definitely not obvious,
so the compiler is quite reasonable to warn about this.
So just add initialization to make it all much more obvious both to
compilers and to humans.
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
We already did this for clang, but now gcc has that warning too. Yes,
yes, the address may be unaligned. And that's kind of the point.
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Previously, during fragmentation after forwarding, skb->skb_iif isn't
preserved, i.e. 'ip_copy_metadata' does not copy skb_iif from given
'from' skb.
As a result, ip_do_fragment's creates fragments with zero skb_iif,
leading to inconsistent behavior.
Assume for example an eBPF program attached at tc egress (post
forwarding) that examines __sk_buff->ingress_ifindex:
- the correct iif is observed if forwarding path does not involve
fragmentation/refragmentation
- a bogus iif is observed if forwarding path involves
fragmentation/refragmentatiom
Fix, by preserving skb_iif during 'ip_copy_metadata'.
Signed-off-by: Shmulik Ladkani <shmulik.ladkani@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
* share more code across drivers
* new driver for MT7615 chipsets
* rework DMA API
* tx/rx performance optimizations
* use NAPI for tx cleanup on mt76x02
* AP mode support for USB devices
* USB stability fixes
* tx power handling fixes for 76x2
* endian fixes
-----BEGIN PGP SIGNATURE-----
Version: GnuPG/MacGPG2 v2
Comment: GPGTools - http://gpgtools.org
iEYEABECAAYFAlzJfSQACgkQ130UHQKnbvUqiACfeNJFtx6+0CbEvDXDGKJ7a7Mx
w28An1uKNBgAKSbNoWny3ZFbOoHuApq9
=i9U/
-----END PGP SIGNATURE-----
Merge tag 'mt76-for-kvalo-2019-05-01' of https://github.com/nbd168/wireless
mt76 patches for 5.2
* share more code across drivers
* new driver for MT7615 chipsets
* rework DMA API
* tx/rx performance optimizations
* use NAPI for tx cleanup on mt76x02
* AP mode support for USB devices
* USB stability fixes
* tx power handling fixes for 76x2
* endian fixes
In io_sqe_buffer_register() we allocate a number of arrays based on the
iov_len from the user-provided iov. While we limit iov_len to SZ_1G,
we can still attempt to allocate arrays exceeding MAX_ORDER.
On a 64-bit system with 4KiB pages, for an iov where iov_base = 0x10 and
iov_len = SZ_1G, we'll calculate that nr_pages = 262145. When we try to
allocate a corresponding array of (16-byte) bio_vecs, requiring 4194320
bytes, which is greater than 4MiB. This results in SLUB warning that
we're trying to allocate greater than MAX_ORDER, and failing the
allocation.
Avoid this by using kvmalloc() for allocations dependent on the
user-provided iov_len. At the same time, fix a leak of imu->bvec when
registration fails.
Full splat from before this patch:
WARNING: CPU: 1 PID: 2314 at mm/page_alloc.c:4595 __alloc_pages_nodemask+0x7ac/0x2938 mm/page_alloc.c:4595
Kernel panic - not syncing: panic_on_warn set ...
CPU: 1 PID: 2314 Comm: syz-executor326 Not tainted 5.1.0-rc7-dirty #4
Hardware name: linux,dummy-virt (DT)
Call trace:
dump_backtrace+0x0/0x2f0 include/linux/compiler.h:193
show_stack+0x20/0x30 arch/arm64/kernel/traps.c:158
__dump_stack lib/dump_stack.c:77 [inline]
dump_stack+0x110/0x190 lib/dump_stack.c:113
panic+0x384/0x68c kernel/panic.c:214
__warn+0x2bc/0x2c0 kernel/panic.c:571
report_bug+0x228/0x2d8 lib/bug.c:186
bug_handler+0xa0/0x1a0 arch/arm64/kernel/traps.c:956
call_break_hook arch/arm64/kernel/debug-monitors.c:301 [inline]
brk_handler+0x1d4/0x388 arch/arm64/kernel/debug-monitors.c:316
do_debug_exception+0x1a0/0x468 arch/arm64/mm/fault.c:831
el1_dbg+0x18/0x8c
__alloc_pages_nodemask+0x7ac/0x2938 mm/page_alloc.c:4595
alloc_pages_current+0x164/0x278 mm/mempolicy.c:2132
alloc_pages include/linux/gfp.h:509 [inline]
kmalloc_order+0x20/0x50 mm/slab_common.c:1231
kmalloc_order_trace+0x30/0x2b0 mm/slab_common.c:1243
kmalloc_large include/linux/slab.h:480 [inline]
__kmalloc+0x3dc/0x4f0 mm/slub.c:3791
kmalloc_array include/linux/slab.h:670 [inline]
io_sqe_buffer_register fs/io_uring.c:2472 [inline]
__io_uring_register fs/io_uring.c:2962 [inline]
__do_sys_io_uring_register fs/io_uring.c:3008 [inline]
__se_sys_io_uring_register fs/io_uring.c:2990 [inline]
__arm64_sys_io_uring_register+0x9e0/0x1bc8 fs/io_uring.c:2990
__invoke_syscall arch/arm64/kernel/syscall.c:35 [inline]
invoke_syscall arch/arm64/kernel/syscall.c:47 [inline]
el0_svc_common.constprop.0+0x148/0x2e0 arch/arm64/kernel/syscall.c:83
el0_svc_handler+0xdc/0x100 arch/arm64/kernel/syscall.c:129
el0_svc+0x8/0xc arch/arm64/kernel/entry.S:948
SMP: stopping secondary CPUs
Dumping ftrace buffer:
(ftrace buffer empty)
Kernel Offset: disabled
CPU features: 0x002,23000438
Memory Limit: none
Rebooting in 1 seconds..
Fixes: edafccee56 ("io_uring: add support for pre-mapped user IO buffers")
Signed-off-by: Mark Rutland <mark.rutland@arm.com>
Cc: Alexander Viro <viro@zeniv.linux.org.uk>
Cc: Jens Axboe <axboe@kernel.dk>
Cc: linux-fsdevel@vger.kernel.org
Cc: linux-block@vger.kernel.org
Cc: linux-kernel@vger.kernel.org
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Vinicius Costa Gomes says:
====================
net/sched: taprio change schedules
Changes from RFC:
- Removed the patches for taprio offloading, because of the lack of
in-tree users;
- Updated the links to point to the PATCH version of this series;
Original cover letter:
Overview
--------
This RFC has two objectives, it adds support for changing the running
schedules during "runtime", explained in more detail later, and
proposes an interface between taprio and the drivers for hardware
offloading.
These two different features are presented together so it's clear what
the "final state" would look like. But after the RFC stage, they can
be proposed (and reviewed) separately.
Changing the schedules without disrupting traffic is important for
handling dynamic use cases, for example, when streams are
added/removed and when the network configuration changes.
Hardware offloading support allows schedules to be more precise and
have lower resource usage.
Changing schedules
------------------
The same as the other interfaces we proposed, we try to use the same
concepts as the IEEE 802.1Q-2018 specification. So, for changing
schedules, there are an "oper" (operational) and an "admin" schedule.
The "admin" schedule is mutable and not in use, the "oper" schedule is
immutable and is in use.
That is, when the user first adds an schedule it is in the "admin"
state, and it becomes "oper" when its base-time (basically when it
starts) is reached.
What this means is that now it's possible to create taprio with a schedule:
$ tc qdisc add dev IFACE parent root handle 100 taprio \
num_tc 3 \
map 2 2 1 0 2 2 2 2 2 2 2 2 2 2 2 2 \
queues 1@0 1@1 2@2 \
base-time 10000000 \
sched-entry S 03 300000 \
sched-entry S 02 300000 \
sched-entry S 06 400000 \
clockid CLOCK_TAI
And then, later, after the previous schedule is "promoted" to "oper",
add a new ("admin") schedule to be used some time later:
$ tc qdisc change dev IFACE parent root handle 100 taprio \
base-time 1553121866000000000 \
sched-entry S 02 500000 \
sched-entry S 0f 400000 \
clockid CLOCK_TAI
When enabling the ability to change schedules, it makes sense to add
two more defined knobs to schedules: "cycle-time" allows to truncate a
cycle to some value, so it repeats after a well-defined value;
"cycle-time-extension" controls how much an entry can be extended if
it's the last one before the change of schedules, the reason is to
avoid a very small cycle when transitioning from a schedule to
another.
With these, taprio in the software mode should provide a fairly
complete implementation of what's defined in the Enhancements for
Scheduled Traffic parts of the specification.
Hardware offload
----------------
Some workloads require better guarantees from their schedules than
what's provided by the software implementation. This series proposes
an interface for configuring schedules into compatible network
controllers.
This part is proposed together with the support for changing
schedules, because it raises questions like, should the "qdisc" side
be responsible of providing visibility into the schedules or should it
be the driver?
In this proposal, the driver is called passing the new schedule as
soon as it is validated, and the "core" qdisc takes care of displaying
(".dump()") the correct schedules at all times. It means that some
logic would need to be duplicated in the driver, if the hardware
doesn't have support for multiple schedules. But as taprio doesn't
have enough information about the underlying controller to know how
much in advance a schedule needs to be informed to the hardware, it
feels like a fair compromise.
The hardware offloading part of this proposal also tries to define an
interface for frame-preemption and how it interacts with the
scheduling of traffic, see Section 8.6.8.4 of IEEE 802.1Q-2018 for
more information.
One important difference between the qdisc interface and the
qdisc-driver interface, is that the "gate mask" on the qdisc side
references traffic classes, that is bit 0 of the gate mask means
Traffic Class 0, and in the driver interface, it specifies the queues,
that is bit 0 means queue 0. That is to say that taprio converts the
references to traffic classes to references to queues before sending
the offloading request to the driver.
Request for help
----------------
I would like that interested driver maintainers could take a look at
the proposed interface and see if it's going to be too awkward for any
particular device. Also, pointers to available documentation would be
appreciated. The idea here is to start a discussion so we can have an
interface that would work for multiple vendors.
Links
-----
kernel patches:
https://github.com/vcgomes/net-next/tree/taprio-add-support-for-change-v3
iproute2 patches:
https://github.com/vcgomes/iproute2/tree/taprio-add-support-for-change-v3
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
IEEE 802.1Q-2018 defines the concept of a cycle-time-extension, so the
last entry of a schedule before the start of a new schedule can be
extended, so "too-short" entries can be avoided.
Signed-off-by: Vinicius Costa Gomes <vinicius.gomes@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
IEEE 802.1Q-2018 defines that a the cycle-time of a schedule may be
overridden, so the schedule is truncated to a determined "width".
Signed-off-by: Vinicius Costa Gomes <vinicius.gomes@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
The IEEE 802.1Q-2018 defines two "types" of schedules, the "Oper" (from
operational?) and "Admin" ones. Up until now, 'taprio' only had
support for the "Oper" one, added when the qdisc is created. This adds
support for the "Admin" one, which allows the .change() operation to
be supported.
Just for clarification, some quick (and dirty) definitions, the "Oper"
schedule is the currently (as in this instant) running one, and it's
read-only. The "Admin" one is the one that the system configurator has
installed, it can be changed, and it will be "promoted" to "Oper" when
it's 'base-time' is reached.
The idea behing this patch is that calling something like the below,
(after taprio is already configured with an initial schedule):
$ tc qdisc change taprio dev IFACE parent root \
base-time X \
sched-entry <CMD> <GATES> <INTERVAL> \
...
Will cause a new admin schedule to be created and programmed to be
"promoted" to "Oper" at instant X. If an "Admin" schedule already
exists, it will be overwritten with the new parameters.
Up until now, there was some code that was added to ease the support
of changing a single entry of a schedule, but was ultimately unused.
Now, that we have support for "change" with more well thought
semantics, updating a single entry seems to be less useful.
So we remove what is in practice dead code, and return a "not
supported" error if the user tries to use it. If changing a single
entry would make the user's life easier we may ressurrect this idea,
but at this point, removing it simplifies the code.
For now, only the schedule specific bits are allowed to be added for a
new schedule, that means that 'clockid', 'num_tc', 'map' and 'queues'
cannot be modified.
Example:
$ tc qdisc change dev IFACE parent root handle 100 taprio \
base-time $BASE_TIME \
sched-entry S 00 500000 \
sched-entry S 0f 500000 \
clockid CLOCK_TAI
The only change in the netlink API introduced by this change is the
introduction of an "admin" type in the response to a dump request,
that type allows userspace to separate the "oper" schedule from the
"admin" schedule. If userspace doesn't support the "admin" type, it
will only display the "oper" schedule.
Signed-off-by: Vinicius Costa Gomes <vinicius.gomes@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Right now, this isn't a problem, but the next commit allows schedules
to be added during runtime. When a new schedule transitions from the
inactive to the active state ("admin" -> "oper") the previous one can
be freed, if it's freed just after the RCU read lock is released, we
may access an invalid entry.
So, we should take care to protect the dequeue() flow, so all the
places that access the entries are protected by the RCU read lock.
Signed-off-by: Vinicius Costa Gomes <vinicius.gomes@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Yuchung Cheng says:
====================
undo congestion window on spurious SYN or SYNACK timeout
Linux TCP currently uses the initial congestion window of 1 packet
if multiple SYN or SYNACK timeouts per RFC6298. However such
timeouts are often spurious on wireless or cellular networks that
experience high delay variances (e.g. ramping up dormant radios or
local link retransmission). Another case is when the underlying
path is longer than the default SYN timeout (e.g. 1 second). In
these cases starting the transfer with a minimal congestion window
is detrimental to the performance for short flows.
One naive approach is to simply ignore SYN or SYNACK timeouts and
always use a larger or default initial window. This approach however
risks pouring gas to the fire when the network is already highly
congested. This is particularly true in data center where application
could start thousands to millions of connections over a single or
multiple hosts resulting in high SYN drops (e.g. incast).
This patch-set detects spurious SYN and SYNACK timeouts upon
completing the handshake via the widely-supported TCP timestamp
options. Upon such events the sender reverts to the default
initial window to start the data transfer so it gets best of both
worlds. This patch-set supports this feature for both active and
passive as well as Fast Open or regular connections.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
Relocate the congestion window initialization from tcp_init_metrics()
to tcp_init_transfer() to improve code readability.
Signed-off-by: Yuchung Cheng <ycheng@google.com>
Signed-off-by: Neal Cardwell <ncardwell@google.com>
Signed-off-by: Soheil Hassas Yeganeh <soheil@google.com>
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Use a helper to consolidate two identical code block for passive TFO.
Signed-off-by: Yuchung Cheng <ycheng@google.com>
Signed-off-by: Neal Cardwell <ncardwell@google.com>
Signed-off-by: Soheil Hassas Yeganeh <soheil@google.com>
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
This patch makes passive Fast Open reverts the cwnd to default
initial cwnd (10 packets) if the SYNACK timeout is spurious.
Passive Fast Open uses a full socket during handshake so it can
use the existing undo logic to detect spurious retransmission
by recording the first SYNACK timeout in key state variable
retrans_stamp. Upon receiving the ACK of the SYNACK, if the socket
has sent some data before the timeout, the spurious timeout
is detected by tcp_try_undo_recovery() in tcp_process_loss()
in tcp_ack().
But if the socket has not send any data yet, tcp_ack() does not
execute the undo code since no data is acknowledged. The fix is to
check such case explicitly after tcp_ack() during the ACK processing
in SYN_RECV state. In addition this is checked in FIN_WAIT_1 state
in case the server closes the socket before handshake completes.
Signed-off-by: Yuchung Cheng <ycheng@google.com>
Signed-off-by: Neal Cardwell <ncardwell@google.com>
Signed-off-by: Soheil Hassas Yeganeh <soheil@google.com>
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
TCP sender would use congestion window of 1 packet on the second SYN
and SYNACK timeout except passive TCP Fast Open. This makes passive
TFO too aggressive and unfair during congestion at handshake. This
patch fixes this issue so TCP (fast open or not, passive or active)
always conforms to the RFC6298.
Note that tcp_enter_loss() is called only once during recurring
timeouts. This is because during handshake, high_seq and snd_una
are the same so tcp_enter_loss() would incorrect set the undo state
variables multiple times.
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>
Linux implements RFC6298 and use an initial congestion window
of 1 upon establishing the connection if the SYNACK packet is
retransmitted 2 or more times. In cellular networks SYNACK timeouts
are often spurious if the wireless radio was dormant or idle. Also
some network path is longer than the default SYNACK timeout. In
both cases falsely starting with a minimal cwnd are detrimental
to performance.
This patch avoids doing so when the final ACK's TCP timestamp
indicates the original SYNACK was delivered. It remembers the
original SYNACK timestamp when SYNACK timeout has occurred and
re-uses the function to detect spurious SYN timeout conveniently.
Note that a server may receives multiple SYNs from and immediately
retransmits SYNACKs without any SYNACK timeout. This often happens
on when the client SYNs have timed out due to wireless delay
above. In this case since the server will still use the default
initial congestion (e.g. 10) because tp->undo_marker is reset in
tcp_init_metrics(). This is an intentional design because packets
are not lost but delayed.
This patch only covers regular TCP passive open. Fast Open is
supported in the next patch.
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>
Detecting spurious SYNACK timeout using timestamp option requires
recording the exact SYNACK skb timestamp. Previously the SYNACK
sent timestamp was stamped slightly earlier before the skb
was transmitted. This patch uses the SYNACK skb transmission
timestamp directly.
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>
Linux implements RFC6298 and use an initial congestion window of 1
upon establishing the connection if the SYN packet is retransmitted 2
or more times. In cellular networks SYN timeouts are often spurious
if the wireless radio was dormant or idle. Also some network path
is longer than the default SYN timeout. Having a minimal cwnd on
both cases are detrimental to TCP startup performance.
This patch extends TCP undo feature (RFC3522 aka TCP Eifel) to detect
spurious SYN timeout via TCP timestamps. Since tp->retrans_stamp
records the initial SYN timestamp instead of first retransmission, we
have to implement a different undo code additionally. The detection
also must happen before tcp_ack() as retrans_stamp is reset when
SYN is acknowledged.
Note this patch covers both active regular and fast open.
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>
