Daniel Borkmann says:
====================
pull-request: bpf 2018-01-09
The following pull-request contains BPF updates for your *net* tree.
The main changes are:
1) Prevent out-of-bounds speculation in BPF maps by masking the
index after bounds checks in order to fix spectre v1, and
add an option BPF_JIT_ALWAYS_ON into Kconfig that allows for
removing the BPF interpreter from the kernel in favor of
JIT-only mode to make spectre v2 harder, from Alexei.
2) Remove false sharing of map refcount with max_entries which
was used in spectre v1, from Daniel.
3) Add a missing NULL psock check in sockmap in order to fix
a race, from John.
4) Fix test_align BPF selftest case since a recent change in
verifier rejects the bit-wise arithmetic on pointers
earlier but test_align update was missing, from Alexei.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
Since commit f11a04464a ("i2c: gpio: Enable working over slow
can_sleep GPIOs"), probing the i2c RTC connected to an i2c-gpio bus on
r8a7740/armadillo fails with:
rtc-s35390a 0-0030: error resetting chip
rtc-s35390a: probe of 0-0030 failed with error -5
More debug code reveals:
i2c i2c-0: master_xfer[0] R, addr=0x30, len=1
i2c i2c-0: NAK from device addr 0x30 msg #0
s35390a_get_reg: ret = -6
Commit 02e479808b ("gpio: Alter semantics of *raw* operations to
actually be raw") moved open drain/source handling from
gpiod_set_raw_value_commit() to gpiod_set_value(), but forgot to take
into account that gpiod_set_value_cansleep() also needs this handling.
The i2c protocol mandates that i2c signals are open drain, hence i2c
communication fails.
Fix this by adding the missing handling to gpiod_set_value_cansleep(),
using a new common helper gpiod_set_value_nocheck().
Fixes: 02e479808b ("gpio: Alter semantics of *raw* operations to actually be raw")
Signed-off-by: Geert Uytterhoeven <geert+renesas@glider.be>
[removed underscore syntax, added kerneldoc]
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
Jakub Kicinski says:
====================
This series starts with a fix to Jesper's recent work, somehow I forgot
about control rings during review. Second patch is cleaning up a vNIC
header, in kdoc we should not use @ for #define constants. Aligning of
the top of the stack as well as bottom (last bytes will be unused) helps
the performance. We should check offload datapath's max MTU when program
is loaded and we can allow TC hw offload flag to be changed freely while
XDP offload is active.
Next group of patches adds more fully featured relocation support. Due
to limited amount of code space we only load the image to NIC's memory
when program is attached. Since we can't predict which programs are
loaded later, we should translate as if image was to be loaded at offset
zero and only apply relocations at load time. Many more advanced features
(eg. tail class, subprograms, dynamic allocation of program space and
sharing it between ports) will depend on this.
Nic adds support for signed comparison instructions.
Quentin makes use of the verifier log in our driver, the verifier print
function (verbose()) has to be renamed and exported.
v2:
- replace #define by function aliasing for verbose() in patch 13
====================
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Now that `bpf_verifier_log_write()` is exported from the verifier and
makes it possible to reuse the verifier log to print messages to the
standard output, use this instead of the kernel logs in the nfp driver
for printing error messages occurring at verification time.
Signed-off-by: Quentin Monnet <quentin.monnet@netronome.com>
Reviewed-by: Jakub Kicinski <jakub.kicinski@netronome.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Rename the BPF verifier `verbose()` to `bpf_verifier_log_write()` and
export it, so that other components (in particular, drivers for BPF
offload) can reuse the user buffer log to dump error messages at
verification time.
Renaming `verbose()` was necessary in order to avoid a name so generic
to be exported to the global namespace. However to prevent too much pain
for backports, the calls to `verbose()` in the kernel BPF verifier were
not changed. Instead, use function aliasing to make `verbose` point to
`bpf_verifier_log_write`. Another solution could consist in making a
wrapper around `verbose()`, but since it is a variadic function, I don't
see a clean way without creating two identical wrappers, one for the
verifier and one to export.
Signed-off-by: Quentin Monnet <quentin.monnet@netronome.com>
Reviewed-by: Jakub Kicinski <jakub.kicinski@netronome.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
This patch adds signed jump instructions (jsgt, jsge, jslt, jsle)
to the nfp jit. As well as adding the additional required raw
assembler branch mask to nfp_asm.h
Signed-off-by: Nic Viljoen <nick.viljoen@netronome.com>
Reviewed-by: Jiong Wang <jiong.wang@netronome.com>
Reviewed-by: Jakub Kicinski <jakub.kicinski@netronome.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Instead of having an app callback per message type hand off
all offload-related handling to apps with one "rest of ndo_bpf"
callback.
Signed-off-by: Jakub Kicinski <jakub.kicinski@netronome.com>
Reviewed-by: Quentin Monnet <quentin.monnet@netronome.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
To make absolute relocated branches (branches which will be completely
rewritten with br_set_offset()) distinguishable in user space dumps
from normal jumps add a large offset to them.
Signed-off-by: Jakub Kicinski <jakub.kicinski@netronome.com>
Reviewed-by: Jiong Wang <jiong.wang@netronome.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
The translator pre-allocates a buffer of maximal program size.
Due to HW/FW limitations the program buffer can't currently be
longer than 128Kb, so we used to kmalloc() it, and then map for
DMA directly.
Now that the late branch resolution is copying the program image
anyway, we can just kvmalloc() the buffer. While at it, after
translation reallocate the buffer to save space.
Signed-off-by: Jakub Kicinski <jakub.kicinski@netronome.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Don't translate the program assuming it will be loaded at a given
address. This will be required for sharing programs between ports
of the same NIC, tail calls and subprograms. It will also make the
jump targets easier to understand when dumping the program to user
space.
Translate the program as if it was going to be loaded at address
zero. When load happens add the load offset in and set addresses
of special branches.
Signed-off-by: Jakub Kicinski <jakub.kicinski@netronome.com>
Reviewed-by: Jiong Wang <jiong.wang@netronome.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
In preparation for better handling of relocations move existing
helper for setting branch offset to nfp_asm.c and add two more.
Signed-off-by: Jakub Kicinski <jakub.kicinski@netronome.com>
Reviewed-by: Jiong Wang <jiong.wang@netronome.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Jump target resolution should be in jit.c not offload.c.
No functional changes.
Signed-off-by: Jakub Kicinski <jakub.kicinski@netronome.com>
Reviewed-by: Jiong Wang <jiong.wang@netronome.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
TC BPF offload was added first, so we used to assume that
the ethtool TC HW offload flag cannot be touched whenever
any BPF program is loaded on the NIC. This unncessarily
limits changes to the TC flag when offloaded program is XDP.
Signed-off-by: Jakub Kicinski <jakub.kicinski@netronome.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
When BPF offload is active we need may need to restrict the MTU
changes more than just to the limitation of the kernel XDP datapath.
Allow the BPF code to veto a MTU change.
Signed-off-by: Jakub Kicinski <jakub.kicinski@netronome.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Kernel enforces the alignment of the bottom of the stack, NFP
deals with positive offsets better so we should align the top
of the stack. Round the stack size to NFP word size (4B).
Signed-off-by: Jakub Kicinski <jakub.kicinski@netronome.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
We should use % instead of @ for documenting preprocessor defines.
Add missing documentation of __NFP_REPR_TYPE_MAX. This gets rid
of all remaining kdoc warnings in the driver.
Signed-off-by: Jakub Kicinski <jakub.kicinski@netronome.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Some RX rings are used for control messages, those will not have
a netdev pointer in dp. Skip XDP rxq handling on those rings.
Fixes: 7f1c684a89 ("nfp: setup xdp_rxq_info")
Signed-off-by: Jakub Kicinski <jakub.kicinski@netronome.com>
Reviewed-by: Dirk van der Merwe <dirk.vandermerwe@netronome.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Jakub Kicinski says:
====================
Two more trivial fixes to the recent XDP RXQ series.
====================
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Looks like commit e817f85652 ("xdp: generic XDP handling of
xdp_rxq_info") replaced kvfree(dev->_rx) in free_netdev() with
a call to netif_free_rx_queues() which doesn't actually free
the rings?
While at it remove the unnecessary temporary variable.
Fixes: e817f85652 ("xdp: generic XDP handling of xdp_rxq_info")
Signed-off-by: Jakub Kicinski <jakub.kicinski@netronome.com>
Reviewed-by: Simon Horman <simon.horman@netronome.com>
Acked-by: Jesper Dangaard Brouer <brouer@redhat.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
kvzalloc'ed memory should be kvfree'd.
Fixes: e817f85652 ("xdp: generic XDP handling of xdp_rxq_info")
Signed-off-by: Jakub Kicinski <jakub.kicinski@netronome.com>
Reviewed-by: Simon Horman <simon.horman@netronome.com>
Acked-by: Jesper Dangaard Brouer <brouer@redhat.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
This contains what I hope are the last RISC-V changes to go into 4.15.
I know it's a bit last minute, but I think they're all fairly small
changes:
* SR_* constants have been renamed to match the latest ISA
specification.
* Some CONFIG_MMU #ifdef cruft has been removed. We've never supported
!CONFIG_MMU.
* __NR_riscv_flush_icache is now visible to userspace. We were hoping
to avoid making this public in order to force userspace to call the
vDSO entry, but it looks like QEMU's user-mode emulation doesn't want
to emulate a vDSO. In order to allow glibc to fall back to a system
call when the vDSO entry doesn't exist we're just
* Our defconfig is no long empty. This is another one that just slipped
through the cracks. The defconfig isn't perfect, but it's at least
close to what users will want for the first RISC-V development board.
Getting closer is kind of splitting hairs here: none of the RISC-V
specific drivers are in yet, so it's not like things will boot out of
the box.
The only one that's strictly necessary is the __NR_riscv_flush_icache
change, as I want that to be part of the public API starting from our
first kernel so nobody has to worry about it. The others are nice to
haves, but they seem sane for 4.15 to me.
-----BEGIN PGP SIGNATURE-----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=PhUK
-----END PGP SIGNATURE-----
Merge tag 'riscv-for-linus-4.15-rc8_cleanups' of git://git.kernel.org/pub/scm/linux/kernel/git/palmer/linux
Pull RISC-V updates from Palmer Dabbelt:
"This contains what I hope are the last RISC-V changes to go into 4.15.
I know it's a bit last minute, but I think they're all fairly small
changes:
- SR_* constants have been renamed to match the latest ISA
specification.
- Some CONFIG_MMU #ifdef cruft has been removed. We've never
supported !CONFIG_MMU.
- __NR_riscv_flush_icache is now visible to userspace. We were hoping
to avoid making this public in order to force userspace to call the
vDSO entry, but it looks like QEMU's user-mode emulation doesn't
want to emulate a vDSO. In order to allow glibc to fall back to a
system call when the vDSO entry doesn't exist we're just
- Our defconfig is no long empty. This is another one that just
slipped through the cracks. The defconfig isn't perfect, but it's
at least close to what users will want for the first RISC-V
development board. Getting closer is kind of splitting hairs here:
none of the RISC-V specific drivers are in yet, so it's not like
things will boot out of the box.
The only one that's strictly necessary is the __NR_riscv_flush_icache
change, as I want that to be part of the public API starting from our
first kernel so nobody has to worry about it. The others are nice to
haves, but they seem sane for 4.15 to me"
* tag 'riscv-for-linus-4.15-rc8_cleanups' of git://git.kernel.org/pub/scm/linux/kernel/git/palmer/linux:
riscv: rename SR_* constants to match the spec
riscv: remove CONFIG_MMU ifdefs
RISC-V: Make __NR_riscv_flush_icache visible to userspace
RISC-V: Add a basic defconfig
Pull MIPS fixes from Ralf Baechle:
"Another round of MIPS fixes for 4.15.
- Maciej Rozycki found another series of FP issues which requires a
seven part series to restructure and fix.
- James fixes a warning about .set mt which gas doesn't like when
building for R1 processors"
* 'upstream' of git://git.linux-mips.org/pub/scm/ralf/upstream-linus:
MIPS: Validate PR_SET_FP_MODE prctl(2) requests against the ABI of the task
MIPS: Disallow outsized PTRACE_SETREGSET NT_PRFPREG regset accesses
MIPS: Also verify sizeof `elf_fpreg_t' with PTRACE_SETREGSET
MIPS: Fix an FCSR access API regression with NT_PRFPREG and MSA
MIPS: Consistently handle buffer counter with PTRACE_SETREGSET
MIPS: Guard against any partial write attempt with PTRACE_SETREGSET
MIPS: Factor out NT_PRFPREG regset access helpers
MIPS: CPS: Fix r1 .set mt assembler warning
The BPF interpreter has been used as part of the spectre 2 attack CVE-2017-5715.
A quote from goolge project zero blog:
"At this point, it would normally be necessary to locate gadgets in
the host kernel code that can be used to actually leak data by reading
from an attacker-controlled location, shifting and masking the result
appropriately and then using the result of that as offset to an
attacker-controlled address for a load. But piecing gadgets together
and figuring out which ones work in a speculation context seems annoying.
So instead, we decided to use the eBPF interpreter, which is built into
the host kernel - while there is no legitimate way to invoke it from inside
a VM, the presence of the code in the host kernel's text section is sufficient
to make it usable for the attack, just like with ordinary ROP gadgets."
To make attacker job harder introduce BPF_JIT_ALWAYS_ON config
option that removes interpreter from the kernel in favor of JIT-only mode.
So far eBPF JIT is supported by:
x64, arm64, arm32, sparc64, s390, powerpc64, mips64
The start of JITed program is randomized and code page is marked as read-only.
In addition "constant blinding" can be turned on with net.core.bpf_jit_harden
v2->v3:
- move __bpf_prog_ret0 under ifdef (Daniel)
v1->v2:
- fix init order, test_bpf and cBPF (Daniel's feedback)
- fix offloaded bpf (Jakub's feedback)
- add 'return 0' dummy in case something can invoke prog->bpf_func
- retarget bpf tree. For bpf-next the patch would need one extra hunk.
It will be sent when the trees are merged back to net-next
Considered doing:
int bpf_jit_enable __read_mostly = BPF_EBPF_JIT_DEFAULT;
but it seems better to land the patch as-is and in bpf-next remove
bpf_jit_enable global variable from all JITs, consolidate in one place
and remove this jit_init() function.
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Pull block fixes from Jens Axboe:
"A set of fixes that should go into this release. This contains:
- An NVMe pull request from Christoph, with a few critical fixes for
NVMe.
- A block drain queue fix from Ming.
- The concurrent lo_open/release fix for loop"
* 'for-linus' of git://git.kernel.dk/linux-block:
loop: fix concurrent lo_open/lo_release
block: drain queue before waiting for q_usage_counter becoming zero
nvme-fcloop: avoid possible uninitialized variable warning
nvme-mpath: fix last path removal during traffic
nvme-rdma: fix concurrent reset and reconnect
nvme: fix sector units when going between formats
nvme-pci: move use_sgl initialization to nvme_init_iod()
In addition to commit b2157399cc ("bpf: prevent out-of-bounds
speculation") also change the layout of struct bpf_map such that
false sharing of fast-path members like max_entries is avoided
when the maps reference counter is altered. Therefore enforce
them to be placed into separate cachelines.
pahole dump after change:
struct bpf_map {
const struct bpf_map_ops * ops; /* 0 8 */
struct bpf_map * inner_map_meta; /* 8 8 */
void * security; /* 16 8 */
enum bpf_map_type map_type; /* 24 4 */
u32 key_size; /* 28 4 */
u32 value_size; /* 32 4 */
u32 max_entries; /* 36 4 */
u32 map_flags; /* 40 4 */
u32 pages; /* 44 4 */
u32 id; /* 48 4 */
int numa_node; /* 52 4 */
bool unpriv_array; /* 56 1 */
/* XXX 7 bytes hole, try to pack */
/* --- cacheline 1 boundary (64 bytes) --- */
struct user_struct * user; /* 64 8 */
atomic_t refcnt; /* 72 4 */
atomic_t usercnt; /* 76 4 */
struct work_struct work; /* 80 32 */
char name[16]; /* 112 16 */
/* --- cacheline 2 boundary (128 bytes) --- */
/* size: 128, cachelines: 2, members: 17 */
/* sum members: 121, holes: 1, sum holes: 7 */
};
Now all entries in the first cacheline are read only throughout
the life time of the map, set up once during map creation. Overall
struct size and number of cachelines doesn't change from the
reordering. struct bpf_map is usually first member and embedded
in map structs in specific map implementations, so also avoid those
members to sit at the end where it could potentially share the
cacheline with first map values e.g. in the array since remote
CPUs could trigger map updates just as well for those (easily
dirtying members like max_entries intentionally as well) while
having subsequent values in cache.
Quoting from Google's Project Zero blog [1]:
Additionally, at least on the Intel machine on which this was
tested, bouncing modified cache lines between cores is slow,
apparently because the MESI protocol is used for cache coherence
[8]. Changing the reference counter of an eBPF array on one
physical CPU core causes the cache line containing the reference
counter to be bounced over to that CPU core, making reads of the
reference counter on all other CPU cores slow until the changed
reference counter has been written back to memory. Because the
length and the reference counter of an eBPF array are stored in
the same cache line, this also means that changing the reference
counter on one physical CPU core causes reads of the eBPF array's
length to be slow on other physical CPU cores (intentional false
sharing).
While this doesn't 'control' the out-of-bounds speculation through
masking the index as in commit b2157399cc, triggering a manipulation
of the map's reference counter is really trivial, so lets not allow
to easily affect max_entries from it.
Splitting to separate cachelines also generally makes sense from
a performance perspective anyway in that fast-path won't have a
cache miss if the map gets pinned, reused in other progs, etc out
of control path, thus also avoids unintentional false sharing.
[1] https://googleprojectzero.blogspot.ch/2018/01/reading-privileged-memory-with-side.html
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Heiner Kallweit says:
====================
r8169: improve runtime pm
On my system with two network ports I found that runtime PM didn't
suspend the unused port. Therefore I checked runtime pm in this driver
in somewhat more detail and this series improves runtime pm in general
and solves the mentioned issue.
Tested on a system with RTL8168evl (MAC version 34).
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
So far rpm doesn't cover cases like unused ports which are never
brought up. If they are active at probe time they remain in this state.
Included in this patch:
- Let the idle notification check whether we can suspend and let it
schedule the suspend. This way we don't need to have calls to
pm_schedule_suspend in different places.
- At the end of rtl_open and rtl_init_one send an idle notification
to allow suspending if the link is down. If a cable is plugged in
aneg is finished before the suspend timer expires and the suspend
request is cancelled.
- Change rtl8169_runtime_suspend to power down the chip if the
interface is down.
Successfully tested on a RTL8168evl (mac version 34).
Signed-off-by: Heiner Kallweit <hkallweit1@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
This patch partially reverts commit e4fbce740f "r8169: Fix runtime
power management" from 2010. At that time the suspend delay was 100ms
and therefore suspending happened during initial aneg. Currently
suspend delay is 5s, so suspend starts after aneg and the issue
doesn't exist any longer. On my system aneg takes almost 3s, to be on
the safe side let's increase the suspend delay to 10s.
Signed-off-by: Heiner Kallweit <hkallweit1@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
This patch reverts commit 2a15cd2ff4 "r8169: runtime resume before
shutdown" from 2012. Few months after this change the underlying issue
was solved in the PCI core with commit 3ff2de9ba1 "PCI/PM: Resume
device before shutdown".
Signed-off-by: Heiner Kallweit <hkallweit1@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Jon Maloy says:
====================
tipc: improvements to group messaging
We make a number of simplifications and improvements to the group
messaging service. They aim at readability/maintainability of the code
as well as scalability.
The series is based on commit f9c935db80 ("tipc: fix problems with
multipoint-to-point flow control) which has been applied to 'net' but
not yet to 'net-next'.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
The current criteria for returning POLLOUT from a group member socket is
too simplistic. It basically returns POLLOUT as soon as the group has
external destinations, something obviously leading to a lot of spinning
during destination congestion situations. At the same time, the internal
congestion handling is unnecessarily complex.
We now change this as follows.
- We introduce an 'open' flag in struct tipc_group. This flag is used
only to help poll() get the setting of POLLOUT right, and *not* for
congeston handling as such. This means that a user can choose to
ignore an EAGAIN for a destination and go on sending messages to
other destinations in the group if he wants to.
- The flag is set to false every time we return EAGAIN on a send call.
- The flag is set to true every time any member, i.e., not necessarily
the member that caused EAGAIN, is removed from the small_win list.
- We remove the group member 'usr_pending' flag. The size of the send
window and presence in the 'small_win' list is sufficient criteria
for recognizing congestion.
This solution seems to be a reasonable compromise between 'anycast',
which is normally not waiting for POLLOUT for a specific destination,
and the other three send modes, which are.
Acked-by: Ying Xue <ying.xue@windriver.com>
Signed-off-by: Jon Maloy <jon.maloy@ericsson.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
When a member joins a group, it also indicates a binding scope. This
makes it possible to create both node local groups, invisible to other
nodes, as well as cluster global groups, visible everywhere.
In order to avoid that different members end up having permanently
differing views of group size and memberhip, we must inhibit locally
and globally bound members from joining the same group.
We do this by using the binding scope as an additional separator between
groups. I.e., a member must ignore all membership events from sockets
using a different scope than itself, and all lookups for message
destinations must require an exact match between the message's lookup
scope and the potential target's binding scope.
Apart from making it possible to create local groups using the same
identity on different nodes, a side effect of this is that it now also
becomes possible to create a cluster global group with the same identity
across the same nodes, without interfering with the local groups.
Acked-by: Ying Xue <ying.xue@windriver.com>
Signed-off-by: Jon Maloy <jon.maloy@ericsson.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Currently, when a user is subscribing for binding table publications,
he will receive a PUBLISH event for all already existing matching items
in the binding table.
However, a group socket making a subscriptions doesn't need this initial
status update from the binding table, because it has already scanned it
during the join operation. Worse, the multiplicatory effect of issuing
mutual events for dozens or hundreds group members within a short time
frame put a heavy load on the topology server, with the end result that
scale out operations on a big group tend to take much longer than needed.
We now add a new filter option, TIPC_SUB_NO_STATUS, for topology server
subscriptions, so that this initial avalanche of events is suppressed.
This change, along with the previous commit, significantly improves the
range and speed of group scale out operations.
We keep the new option internal for the tipc driver, at least for now.
Acked-by: Ying Xue <ying.xue@windriver.com>
Signed-off-by: Jon Maloy <jon.maloy@ericsson.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
When a socket is joining a group, we look up in the binding table to
find if there are already other members of the group present. This is
used for being able to return EAGAIN instead of EHOSTUNREACH if the
user proceeds directly to a send attempt.
However, the information in the binding table can be used to directly
set the created member in state MBR_PUBLISHED and send a JOIN message
to the peer, instead of waiting for a topology PUBLISH event to do this.
When there are many members in a group, the propagation time for such
events can be significant, and we can save time during the join
operation if we use the initial lookup result fully.
In this commit, we eliminate the member state MBR_DISCOVERED which has
been the result of the initial lookup, and do instead go directly to
MBR_PUBLISHED, which initiates the setup.
After this change, the tipc_member FSM looks as follows:
+-----------+
---->| PUBLISHED |-----------------------------------------------+
PUB- +-----------+ LEAVE/WITHRAW |
LISH |JOIN |
| +-------------------------------------------+ |
| | LEAVE/WITHDRAW | |
| | +------------+ | |
| | +----------->| PENDING |---------+ | |
| | |msg/maxactv +-+---+------+ LEAVE/ | | |
| | | | | WITHDRAW | | |
| | | +----------+ | | | |
| | | |revert/maxactv| | | |
| | | V V V V V
| +----------+ msg +------------+ +-----------+
+-->| JOINED |------>| ACTIVE |------>| LEAVING |--->
| +----------+ +--- -+------+ LEAVE/+-----------+DOWN
| A A | WITHDRAW A A A EVT
| | | |RECLAIM | | |
| | |REMIT V | | |
| | |== adv +------------+ | | |
| | +---------| RECLAIMING |--------+ | |
| | +-----+------+ LEAVE/ | |
| | |REMIT WITHDRAW | |
| | |< adv | |
| |msg/ V LEAVE/ | |
| |adv==ADV_IDLE+------------+ WITHDRAW | |
| +-------------| REMITTED |------------+ |
| +------------+ |
|PUBLISH |
JOIN +-----------+ LEAVE/WITHDRAW |
---->| JOINING |-----------------------------------------------+
+-----------+
Acked-by: Ying Xue <ying.xue@windriver.com>
Signed-off-by: Jon Maloy <jon.maloy@ericsson.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
After the changes in the previous commit the group LEAVE sequence
can be simplified.
We now let the arrival of a LEAVE message unconditionally issue a group
DOWN event to the user. When a topology WITHDRAW event is received, the
member, if it still there, is set to state LEAVING, but we only issue a
group DOWN event when the link to the peer node is gone, so that no
LEAVE message is to be expected.
Acked-by: Ying Xue <ying.xue@windriver.com>
Signed-off-by: Jon Maloy <jon.maloy@ericsson.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
In the current implementation, a group socket receiving topology
events about other members just converts the topology event message
into a group event message and stores it until it reaches the right
state to issue it to the user. This complicates the code unnecessarily,
and becomes impractical when we in the coming commits will need to
create and issue membership events independently.
In this commit, we change this so that we just notice the type and
origin of the incoming topology event, and then drop the buffer. Only
when it is time to actually send a group event to the user do we
explicitly create a new message and send it upwards.
Acked-by: Ying Xue <ying.xue@windriver.com>
Signed-off-by: Jon Maloy <jon.maloy@ericsson.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Analysis reveals that the member state MBR_QURANTINED in reality is
unnecessary, and can be replaced by the state MBR_JOINING at all
occurrencs.
Acked-by: Ying Xue <ying.xue@windriver.com>
Signed-off-by: Jon Maloy <jon.maloy@ericsson.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
We handle a corner case in the function tipc_group_update_rcv_win().
During extreme pessure it might happen that a message receiver has all
its active senders in RECLAIMING or REMITTED mode, meaning that there
is nobody to reclaim advertisements from if an additional sender tries
to go active.
Currently we just set the new sender to ACTIVE anyway, hence at least
theoretically opening up for a receiver queue overflow by exceeding the
MAX_ACTIVE limit. The correct solution to this is to instead add the
member to the pending queue, while letting the oldest member in that
queue revert to JOINED state.
In this commit we refactor the code for handling message arrival from
a JOINED member, both to make it more comprehensible and to cover the
case described above.
Acked-by: Ying Xue <ying.xue@windriver.com>
Signed-off-by: Jon Maloy <jon.maloy@ericsson.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
- We remove the 'reclaiming' member list in struct tipc_group, since
it doesn't serve any purpose.
- We simplify the GRP_REMIT_MSG branch of tipc_group_protocol_rcv().
Acked-by: Ying Xue <ying.xue@windriver.com>
Signed-off-by: Jon Maloy <jon.maloy@ericsson.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
In the current code, when creating a new fib6 table, tb6_root.leaf gets
initialized to net->ipv6.ip6_null_entry.
If a default route is being added with rt->rt6i_metric = 0xffffffff,
fib6_add() will add this route after net->ipv6.ip6_null_entry. As
null_entry is shared, it could cause problem.
In order to fix it, set fn->leaf to NULL before calling
fib6_add_rt2node() when trying to add the first default route.
And reset fn->leaf to null_entry when adding fails or when deleting the
last default route.
syzkaller reported the following issue which is fixed by this commit:
WARNING: suspicious RCU usage
4.15.0-rc5+ #171 Not tainted
-----------------------------
net/ipv6/ip6_fib.c:1702 suspicious rcu_dereference_protected() usage!
other info that might help us debug this:
rcu_scheduler_active = 2, debug_locks = 1
4 locks held by swapper/0/0:
#0: ((&net->ipv6.ip6_fib_timer)){+.-.}, at: [<00000000d43f631b>] lockdep_copy_map include/linux/lockdep.h:178 [inline]
#0: ((&net->ipv6.ip6_fib_timer)){+.-.}, at: [<00000000d43f631b>] call_timer_fn+0x1c6/0x820 kernel/time/timer.c:1310
#1: (&(&net->ipv6.fib6_gc_lock)->rlock){+.-.}, at: [<000000002ff9d65c>] spin_lock_bh include/linux/spinlock.h:315 [inline]
#1: (&(&net->ipv6.fib6_gc_lock)->rlock){+.-.}, at: [<000000002ff9d65c>] fib6_run_gc+0x9d/0x3c0 net/ipv6/ip6_fib.c:2007
#2: (rcu_read_lock){....}, at: [<0000000091db762d>] __fib6_clean_all+0x0/0x3a0 net/ipv6/ip6_fib.c:1560
#3: (&(&tb->tb6_lock)->rlock){+.-.}, at: [<000000009e503581>] spin_lock_bh include/linux/spinlock.h:315 [inline]
#3: (&(&tb->tb6_lock)->rlock){+.-.}, at: [<000000009e503581>] __fib6_clean_all+0x1d0/0x3a0 net/ipv6/ip6_fib.c:1948
stack backtrace:
CPU: 0 PID: 0 Comm: swapper/0 Not tainted 4.15.0-rc5+ #171
Hardware name: Google Google Compute Engine/Google Compute Engine, BIOS Google 01/01/2011
Call Trace:
<IRQ>
__dump_stack lib/dump_stack.c:17 [inline]
dump_stack+0x194/0x257 lib/dump_stack.c:53
lockdep_rcu_suspicious+0x123/0x170 kernel/locking/lockdep.c:4585
fib6_del+0xcaa/0x11b0 net/ipv6/ip6_fib.c:1701
fib6_clean_node+0x3aa/0x4f0 net/ipv6/ip6_fib.c:1892
fib6_walk_continue+0x46c/0x8a0 net/ipv6/ip6_fib.c:1815
fib6_walk+0x91/0xf0 net/ipv6/ip6_fib.c:1863
fib6_clean_tree+0x1e6/0x340 net/ipv6/ip6_fib.c:1933
__fib6_clean_all+0x1f4/0x3a0 net/ipv6/ip6_fib.c:1949
fib6_clean_all net/ipv6/ip6_fib.c:1960 [inline]
fib6_run_gc+0x16b/0x3c0 net/ipv6/ip6_fib.c:2016
fib6_gc_timer_cb+0x20/0x30 net/ipv6/ip6_fib.c:2033
call_timer_fn+0x228/0x820 kernel/time/timer.c:1320
expire_timers kernel/time/timer.c:1357 [inline]
__run_timers+0x7ee/0xb70 kernel/time/timer.c:1660
run_timer_softirq+0x4c/0xb0 kernel/time/timer.c:1686
__do_softirq+0x2d7/0xb85 kernel/softirq.c:285
invoke_softirq kernel/softirq.c:365 [inline]
irq_exit+0x1cc/0x200 kernel/softirq.c:405
exiting_irq arch/x86/include/asm/apic.h:540 [inline]
smp_apic_timer_interrupt+0x16b/0x700 arch/x86/kernel/apic/apic.c:1052
apic_timer_interrupt+0xa9/0xb0 arch/x86/entry/entry_64.S:904
</IRQ>
Reported-by: syzbot <syzkaller@googlegroups.com>
Fixes: 66f5d6ce53 ("ipv6: replace rwlock with rcu and spinlock in fib6_table")
Signed-off-by: Wei Wang <weiwan@google.com>
Acked-by: Martin KaFai Lau <kafai@fb.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Sergei Shtylyov says:
====================
Ether fixes for the SolutionEngine771x boards
Here's the series of 2 patches against Linus' repo. This series should
(hoplefully) fix the Ether support on the SolutionEngine771x boards...
[1/2] SolutionEngine771x: fix Ether platform data
[2/2] SolutionEngine771x: add Ether TSU resource
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
After the Ether platform data is fixed, the driver probe() method would
still fail since the 'struct sh_eth_cpu_data' corresponding to SH771x
indicates the presence of TSU but the memory resource for it is absent.
Add the missing TSU resource to both Ether devices and fix the harmless
off-by-one error in the main memory resources, while at it...
Fixes: 4986b99688 ("net: sh_eth: remove the SH_TSU_ADDR")
Signed-off-by: Sergei Shtylyov <sergei.shtylyov@cogentembedded.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
The 'sh_eth' driver's probe() method would fail on the SolutionEngine7710
board and crash on SolutionEngine7712 board as the platform code is
hopelessly behind the driver's platform data -- it passes the PHY address
instead of 'struct sh_eth_plat_data *'; pass the latter to the driver in
order to fix the bug...
Fixes: 71557a37ad ("[netdrvr] sh_eth: Add SH7619 support")
Signed-off-by: Sergei Shtylyov <sergei.shtylyov@cogentembedded.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Fix the following 'make htmldocs' complaint:
Documentation/networking/msg_zerocopy.rst:: WARNING: document isn't included in any toctree.
Signed-off-by: Mike Rapoport <rppt@linux.vnet.ibm.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Commit 8f659a03a0 ("net: ipv4: fix for a race condition in
raw_sendmsg") fixed the issue of possibly inconsistent ->hdrincl handling
due to concurrent updates by reading this bit-field member into a local
variable and using the thus stabilized value in subsequent tests.
However, aforementioned commit also adds the (correct) comment that
/* hdrincl should be READ_ONCE(inet->hdrincl)
* but READ_ONCE() doesn't work with bit fields
*/
because as it stands, the compiler is free to shortcut or even eliminate
the local variable at its will.
Note that I have not seen anything like this happening in reality and thus,
the concern is a theoretical one.
However, in order to be on the safe side, emulate a READ_ONCE() on the
bit-field by doing it on the local 'hdrincl' variable itself:
int hdrincl = inet->hdrincl;
hdrincl = READ_ONCE(hdrincl);
This breaks the chain in the sense that the compiler is not allowed
to replace subsequent reads from hdrincl with reloads from inet->hdrincl.
Fixes: 8f659a03a0 ("net: ipv4: fix for a race condition in raw_sendmsg")
Signed-off-by: Nicolai Stange <nstange@suse.de>
Reviewed-by: Stefano Brivio <sbrivio@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Tariq Toukan says:
====================
ethtool ringparam upper bound
This patchset by Jenny adds sanity checks in ethtool ringparam
operation for input upper bounds, similarly to what's done in
ethtool_set_channels.
The checks are added in patch 1, using a call to get_ringparam
prior to calling set_ringparam NDO.
Patch 2 changes the function's behavior in mlx4_en, so that
it returns an error for out-of-range input, instead of rounding
it to closest valid, similar to mlx5e.
Patch 3 removes the upper bound checks in mlx5e_ethtool_set_ringparam
as it becomes redundant.
Series generated against net-next commit:
f66faae2f8 Merge branch 'ipv6-ipv4-nexthop-align'
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
Since the checks are done in upper layer ethtool code,
checks in driver are not needed any more.
Signed-off-by: Eugenia Emantayev <eugenia@mellanox.com>
Signed-off-by: Tariq Toukan <tariqt@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
In current implementation, any requested RX/TX ring size value
that is less than minimum is silently casted to nearest valid value.
Update this behavior to align with mlx5 behavior by printing warning
in dmesg and remaining the size unchanged.
Kernel is responsible for verifying against the maximum.
Signed-off-by: Eugenia Emantayev <eugenia@mellanox.com>
Signed-off-by: Tariq Toukan <tariqt@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Add a sanity check to ensure that all requested ring parameters
are within bounds, which should reduce errors in driver implementation.
Signed-off-by: Eugenia Emantayev <eugenia@mellanox.com>
Signed-off-by: Tariq Toukan <tariqt@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
gcc-8 reports
net/caif/caif_dev.c: In function 'caif_enroll_dev':
./include/linux/string.h:245:9: warning: '__builtin_strncpy' output may
be truncated copying 15 bytes from a string of length 15
[-Wstringop-truncation]
net/caif/cfctrl.c: In function 'cfctrl_linkup_request':
./include/linux/string.h:245:9: warning: '__builtin_strncpy' output may
be truncated copying 15 bytes from a string of length 15
[-Wstringop-truncation]
net/caif/cfcnfg.c: In function 'caif_connect_client':
./include/linux/string.h:245:9: warning: '__builtin_strncpy' output may
be truncated copying 15 bytes from a string of length 15
[-Wstringop-truncation]
The compiler require that the input param 'len' of strncpy() should be
greater than the length of the src string, so that '\0' is copied as
well. We can just use strlcpy() to avoid this warning.
Signed-off-by: Xiongfeng Wang <xiongfeng.wang@linaro.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
