* Re: nfc: trf7970a: Prevent repeated polling from crashing the kernel
From: Justin Bronder @ 2016-12-20 19:13 UTC (permalink / raw)
To: Mark Greer
Cc: Geoff Lansberry, linux-wireless-u79uwXL29TY76Z2rM5mHXA,
lauro.venancio-430g2QfJUUCGglJvpFV4uA,
aloisio.almeida-430g2QfJUUCGglJvpFV4uA,
sameo-VuQAYsv1563Yd54FQh9/CA, robh+dt-DgEjT+Ai2ygdnm+yROfE0A,
mark.rutland-5wv7dgnIgG8, netdev-u79uwXL29TY76Z2rM5mHXA,
devicetree-u79uwXL29TY76Z2rM5mHXA,
linux-kernel-u79uwXL29TY76Z2rM5mHXA, Jaret Cantu
In-Reply-To: <20161220185905.GA5867-luAo+O/VEmrlveNOaEYElw@public.gmane.org>
On 20/12/16 11:59 -0700, Mark Greer wrote:
> On Tue, Dec 20, 2016 at 11:16:32AM -0500, Geoff Lansberry wrote:
> > From: Jaret Cantu <jaret.cantu-jEh4hwF5bVhBDgjK7y7TUQ@public.gmane.org>
> >
> > Repeated polling attempts cause a NULL dereference error to occur.
> > This is because the state of the trf7970a is currently reading but
> > another request has been made to send a command before it has finished.
>
> How is this happening? Was trf7970a_abort_cmd() called and it didn't
> work right? Was it not called at all and there is a bug in the digital
> layer? More details please.
>
> > The solution is to properly kill the waiting reading (workqueue)
> > before failing on the send.
>
> If the bug is in the calling code, then that is what should get fixed.
> This seems to be a hack to work-around a digital layer bug.
One of our uses of NFC is to begin polling to read a tag and then stop polling
(in order to save power) until we know via user interaction that we need to poll
again. This is typically many minutes later so the power saving is pretty
significant. However, it's possible that a user will remove the tag before
reading has completed. We also detect this case and stop polling. I can go
more into this if necessary but that is what exposed a panic.
You can reproduce using neard and python, in our testing it was very likely to
occur in 10-100 iterations of the following.:
#!/usr/bin/python
import time
import dbus
bus = dbus.SystemBus()
nfc0 = bus.get_object('org.neard', '/org/neard/nfc0')
props = dbus.Interface(nfc0, 'org.freedesktop.DBus.Properties')
try:
props.Set('org.neard.Adapter', 'Powered', dbus.Boolean(1))
except:
pass
adapter = dbus.Interface(nfc0, 'org.neard.Adapter')
for i in range(1000):
adapter.StartPollLoop('Initiator')
time.sleep(0.1)
adapter.StopPollLoop()
print(i)
I believe the last time we tested this was around the 4.1 release.
--
Justin Bronder
--
To unsubscribe from this list: send the line "unsubscribe devicetree" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
^ permalink raw reply
* Re: [GIT PULL 00/29] perf/core improvements and fixes
From: Ingo Molnar @ 2016-12-20 19:15 UTC (permalink / raw)
To: Arnaldo Carvalho de Melo
Cc: Kim Phillips, Alexander Shishkin, Adrian Hunter, Andi Kleen,
Paul Mackerras, Jiri Olsa, Daniel Borkmann, Alexei Starovoitov,
Ravi Bangoria, Peter Zijlstra, Naveen N . Rao,
Markus Trippelsdorf, Taeung Song, Wang Nan, Joe Stringer,
Nicholas Piggin, Arnaldo Carvalho de Melo, Namhyung Kim,
Kyle McMartin, Kan Liang, Chris Riyder, linux-kernel,
Davidlohr Bueso, Masami
In-Reply-To: <20161220170358.4350-1-acme@kernel.org>
* Arnaldo Carvalho de Melo <acme@kernel.org> wrote:
> Hi Ingo,
>
> Please consider pulling, I had most of this queued before your first
> pull req to Linus for 4.10, most are fixes, with 'perf sched timehist --idle'
> as a followup new feature to the 'perf sched timehist' command introduced in
> this window.
>
> One other thing that delayed this was the samples/bpf/ switch to
> tools/lib/bpf/ that involved fixing up merge clashes with net.git and also
> to properly test it, after more rounds than antecipated, but all seems ok
> now and would be good to get this merge issues past us ASAP.
>
> - Arnaldo
>
> Test results at the end of this message, as usual.
>
> The following changes since commit e7aa8c2eb11ba69b1b69099c3c7bd6be3087b0ba:
>
> Merge tag 'docs-4.10' of git://git.lwn.net/linux (2016-12-12 21:58:13 -0800)
>
> are available in the git repository at:
>
> git://git.kernel.org/pub/scm/linux/kernel/git/acme/linux.git tags/perf-core-for-mingo-20161220
>
> for you to fetch changes up to 9899694a7f67714216665b87318eb367e2c5c901:
>
> samples/bpf: Move open_raw_sock to separate header (2016-12-20 12:00:40 -0300)
>
> ----------------------------------------------------------------
> perf/core improvements and fixes:
>
> New features:
>
> - Introduce 'perf sched timehist --idle', to analyse processes
> going to/from idle state (Namhyung Kim)
>
> Fixes:
>
> - Allow 'perf record -u user' to continue when facing races with threads
> going away after having scanned them via /proc (Jiri Olsa)
>
> - Fix 'perf mem' --all-user/--all-kernel options (Jiri Olsa)
>
> - Support jumps with multiple arguments (Ravi Bangoria)
>
> - Fix jumps to before the function where they are located (Ravi
> Bangoria)
>
> - Fix lock-pi help string (Davidlohr Bueso)
>
> - Fix build of 'perf trace' in odd systems such as a RHEL PPC one (Jiri Olsa)
>
> - Do not overwrite valid build id in 'perf diff' (Kan Liang)
>
> - Don't throw error for zero length symbols, allowing the use of the TUI
> in PowerPC, where such symbols became more common recently (Ravi Bangoria)
>
> Infrastructure:
>
> - Switch of samples/bpf/ to use tools/lib/bpf, removing libbpf
> duplication (Joe Stringer)
>
> - Move headers check into bash script (Jiri Olsa)
>
> Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
>
> ----------------------------------------------------------------
> Arnaldo Carvalho de Melo (3):
> perf tools: Remove some needless __maybe_unused
> samples/bpf: Make perf_event_read() static
> samples/bpf: Be consistent with bpf_load_program bpf_insn parameter
>
> Davidlohr Bueso (1):
> perf bench futex: Fix lock-pi help string
>
> Jiri Olsa (7):
> perf tools: Move headers check into bash script
> perf mem: Fix --all-user/--all-kernel options
> perf evsel: Use variable instead of repeating lengthy FD macro
> perf thread_map: Add thread_map__remove function
> perf evsel: Allow to ignore missing pid
> perf record: Force ignore_missing_thread for uid option
> perf trace: Check if MAP_32BIT is defined (again)
>
> Joe Stringer (8):
> tools lib bpf: Sync {tools,}/include/uapi/linux/bpf.h
> tools lib bpf: use __u32 from linux/types.h
> tools lib bpf: Add flags to bpf_create_map()
> samples/bpf: Make samples more libbpf-centric
> samples/bpf: Switch over to libbpf
> tools lib bpf: Add bpf_prog_{attach,detach}
> samples/bpf: Remove perf_event_open() declaration
> samples/bpf: Move open_raw_sock to separate header
>
> Kan Liang (1):
> perf diff: Do not overwrite valid build id
>
> Namhyung Kim (6):
> perf sched timehist: Split is_idle_sample()
> perf sched timehist: Introduce struct idle_time_data
> perf sched timehist: Save callchain when entering idle
> perf sched timehist: Skip non-idle events when necessary
> perf sched timehist: Add -I/--idle-hist option
> perf sched timehist: Show callchains for idle stat
>
> Ravi Bangoria (3):
> perf annotate: Support jump instruction with target as second operand
> perf annotate: Fix jump target outside of function address range
> perf annotate: Don't throw error for zero length symbols
>
> samples/bpf/Makefile | 70 +--
> samples/bpf/README.rst | 4 +-
> samples/bpf/bpf_load.c | 21 +-
> samples/bpf/bpf_load.h | 3 +
> samples/bpf/fds_example.c | 13 +-
> samples/bpf/lathist_user.c | 2 +-
> samples/bpf/libbpf.c | 176 -------
> samples/bpf/libbpf.h | 28 +-
> samples/bpf/lwt_len_hist_user.c | 6 +-
> samples/bpf/offwaketime_user.c | 8 +-
> samples/bpf/sampleip_user.c | 7 +-
> samples/bpf/sock_example.c | 14 +-
> samples/bpf/sock_example.h | 35 ++
> samples/bpf/sockex1_user.c | 7 +-
> samples/bpf/sockex2_user.c | 5 +-
> samples/bpf/sockex3_user.c | 5 +-
> samples/bpf/spintest_user.c | 8 +-
> samples/bpf/tc_l2_redirect_user.c | 4 +-
> samples/bpf/test_cgrp2_array_pin.c | 4 +-
> samples/bpf/test_cgrp2_attach.c | 12 +-
> samples/bpf/test_cgrp2_attach2.c | 8 +-
> samples/bpf/test_cgrp2_sock.c | 7 +-
> samples/bpf/test_current_task_under_cgroup_user.c | 8 +-
> samples/bpf/test_lru_dist.c | 32 +-
> samples/bpf/test_probe_write_user_user.c | 2 +-
> samples/bpf/trace_event_user.c | 23 +-
> samples/bpf/trace_output_user.c | 7 +-
> samples/bpf/tracex2_user.c | 10 +-
> samples/bpf/tracex3_user.c | 4 +-
> samples/bpf/tracex4_user.c | 4 +-
> samples/bpf/tracex6_user.c | 5 +-
> samples/bpf/xdp1_user.c | 2 +-
> samples/bpf/xdp_tx_iptunnel_user.c | 6 +-
> tools/include/uapi/linux/bpf.h | 593 +++++++++++++---------
> tools/lib/bpf/bpf.c | 30 +-
> tools/lib/bpf/bpf.h | 9 +-
> tools/lib/bpf/libbpf.c | 3 +-
> tools/perf/Documentation/perf-sched.txt | 4 +
> tools/perf/Makefile.perf | 94 +---
> tools/perf/bench/futex-lock-pi.c | 2 +-
> tools/perf/builtin-c2c.c | 13 +-
> tools/perf/builtin-mem.c | 4 +-
> tools/perf/builtin-record.c | 3 +
> tools/perf/builtin-report.c | 2 +-
> tools/perf/builtin-sched.c | 261 ++++++++--
> tools/perf/builtin-stat.c | 6 +-
> tools/perf/check-headers.sh | 59 +++
> tools/perf/perf.h | 1 +
> tools/perf/tests/builtin-test.c | 4 +
> tools/perf/tests/tests.h | 1 +
> tools/perf/tests/thread-map.c | 44 ++
> tools/perf/trace/beauty/mmap.c | 2 +
> tools/perf/ui/browsers/annotate.c | 5 +-
> tools/perf/util/annotate.c | 23 +-
> tools/perf/util/annotate.h | 5 +-
> tools/perf/util/evsel.c | 61 ++-
> tools/perf/util/evsel.h | 1 +
> tools/perf/util/symbol.c | 3 +-
> tools/perf/util/thread_map.c | 22 +
> tools/perf/util/thread_map.h | 1 +
> 60 files changed, 1075 insertions(+), 731 deletions(-)
> delete mode 100644 samples/bpf/libbpf.c
> create mode 100644 samples/bpf/sock_example.h
> create mode 100755 tools/perf/check-headers.sh
Pulled, thanks a lot Arnaldo!
Ingo
^ permalink raw reply
* Re: [PATCH] ethernet: sfc: Add Kconfig entry for vendor Solarflare
From: David Miller @ 2016-12-20 19:20 UTC (permalink / raw)
To: tklauser; +Cc: netdev, linux-net-drivers, ecree, bkenward
In-Reply-To: <20161220133826.1478-1-tklauser@distanz.ch>
From: Tobias Klauser <tklauser@distanz.ch>
Date: Tue, 20 Dec 2016 14:38:26 +0100
> Since commit
>
> 5a6681e22c14 ("sfc: separate out SFC4000 ("Falcon") support into new sfc-falcon driver")
>
> there are two drivers for Solarflare devices, but both still show up
> directly beneath "Ethernet driver support" in the Kconfig. Follow the
> pattern of other vendors and group them beneath an own vendor Kconfig
> entry for Solarflare.
>
> Cc: Edward Cree <ecree@solarflare.com>
> Signed-off-by: Tobias Klauser <tklauser@distanz.ch>
Applied.
^ permalink raw reply
* Re: [mm PATCH 0/3] Page fragment updates
From: Alexander Duyck @ 2016-12-20 19:22 UTC (permalink / raw)
To: Andrew Morton
Cc: linux-mm, Netdev, Eric Dumazet, David Miller, Jeff Kirsher,
linux-kernel@vger.kernel.org
In-Reply-To: <20161205121131.3c1d9ad8452d5e09247336e4@linux-foundation.org>
On Mon, Dec 5, 2016 at 12:11 PM, Andrew Morton
<akpm@linux-foundation.org> wrote:
> On Mon, 5 Dec 2016 09:01:12 -0800 Alexander Duyck <alexander.duyck@gmail.com> wrote:
>
>> On Tue, Nov 29, 2016 at 10:23 AM, Alexander Duyck
>> <alexander.duyck@gmail.com> wrote:
>> > This patch series takes care of a few cleanups for the page fragments API.
>> >
>> > ...
>>
>> It's been about a week since I submitted this series. Just wanted to
>> check in and see if anyone had any feedback or if this is good to be
>> accepted for 4.10-rc1 with the rest of the set?
>
> Looks good to me. I have it all queued for post-4.9 processing.
So I guess there is a small bug in the first patch in that I was
comparing a pointer to to 0 instead of NULL. Just wondering if I
should resubmit the first patch, the whole series, or if I need to
just submit an incremental patch.
Thanks.
- Alex
--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org. For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>
^ permalink raw reply
* Re: [PATCH] net_sched: sch_fq: use rb_entry()
From: David Miller @ 2016-12-20 19:23 UTC (permalink / raw)
To: geliangtang; +Cc: jhs, netdev, linux-kernel
In-Reply-To: <6aa180710d70b09d4f81d3d219b3161077e1ff11.1482204526.git.geliangtang@gmail.com>
From: Geliang Tang <geliangtang@gmail.com>
Date: Tue, 20 Dec 2016 22:02:15 +0800
> To make the code clearer, use rb_entry() instead of container_of() to
> deal with rbtree.
>
> Signed-off-by: Geliang Tang <geliangtang@gmail.com>
Applied.
^ permalink raw reply
* Re: [PATCH] net_sched: sch_netem: use rb_entry()
From: David Miller @ 2016-12-20 19:23 UTC (permalink / raw)
To: geliangtang; +Cc: stephen, jhs, netem, netdev, linux-kernel
In-Reply-To: <dc6e312e4e2ff3868d1791032c329477d27ba50b.1482204694.git.geliangtang@gmail.com>
From: Geliang Tang <geliangtang@gmail.com>
Date: Tue, 20 Dec 2016 22:02:16 +0800
> To make the code clearer, use rb_entry() instead of container_of() to
> deal with rbtree.
>
> Signed-off-by: Geliang Tang <geliangtang@gmail.com>
Applied.
^ permalink raw reply
* Re: [PATCH] RDS: use rb_entry()
From: David Miller @ 2016-12-20 19:23 UTC (permalink / raw)
To: geliangtang
Cc: santosh.shilimkar, netdev, linux-rdma, rds-devel, linux-kernel
In-Reply-To: <2cd84448fe04ffb7023be892c5ed04bbfc759c87.1482204342.git.geliangtang@gmail.com>
From: Geliang Tang <geliangtang@gmail.com>
Date: Tue, 20 Dec 2016 22:02:18 +0800
> To make the code clearer, use rb_entry() instead of container_of() to
> deal with rbtree.
>
> Signed-off-by: Geliang Tang <geliangtang@gmail.com>
Applied.
^ permalink raw reply
* Re: [PATCH] net/mlx5: use rb_entry()
From: David Miller @ 2016-12-20 19:23 UTC (permalink / raw)
To: geliangtang; +Cc: saeedm, matanb, leonro, netdev, linux-rdma, linux-kernel
In-Reply-To: <8443fa3fa03d82c2b829375d8020762e5236dc6d.1482203930.git.geliangtang@gmail.com>
From: Geliang Tang <geliangtang@gmail.com>
Date: Tue, 20 Dec 2016 22:02:14 +0800
> To make the code clearer, use rb_entry() instead of container_of() to
> deal with rbtree.
>
> Signed-off-by: Geliang Tang <geliangtang@gmail.com>
Applied.
^ permalink raw reply
* Re: [PATCH net] be2net: Increase skb headroom size to 256 bytes
From: David Miller @ 2016-12-20 19:30 UTC (permalink / raw)
To: suresh.reddy; +Cc: netdev, kalesh-anakkur.purayil
In-Reply-To: <20161220151430.11115-1-suresh.reddy@broadcom.com>
From: Suresh Reddy <suresh.reddy@broadcom.com>
Date: Tue, 20 Dec 2016 10:14:30 -0500
> From: Kalesh A P <kalesh-anakkur.purayil@broadcom.com>
>
> The driver currently allocates 128 bytes of skb headroom.
> This was found to be insufficient with some configurations
> like Geneve tunnels, which resulted in skb head reallocations.
>
> Increase the headroom to 256 bytes to fix this.
>
> Signed-off-by: Kalesh A P <kalesh-anakkur.purayil@broadcom.com>
> Signed-off-by: Suresh Reddy <suresh.reddy@broadcom.com>
Adding 128 bytes of headroom just for geneve seems excessive.
Do you really need to add that much?
^ permalink raw reply
* Re: ipv6: handle -EFAULT from skb_copy_bits
From: Cong Wang @ 2016-12-20 19:31 UTC (permalink / raw)
To: Dave Jones; +Cc: David Miller, Linux Kernel Network Developers
In-Reply-To: <20161220181728.dd2cynjwrceruwcu@codemonkey.org.uk>
On Tue, Dec 20, 2016 at 10:17 AM, Dave Jones <davej@codemonkey.org.uk> wrote:
> On Mon, Dec 19, 2016 at 08:36:23PM -0500, David Miller wrote:
> > From: Dave Jones <davej@codemonkey.org.uk>
> > Date: Mon, 19 Dec 2016 19:40:13 -0500
> >
> > > On Mon, Dec 19, 2016 at 07:31:44PM -0500, Dave Jones wrote:
> > >
> > > > Unfortunately, this made no difference. I spent some time today trying
> > > > to make a better reproducer, but failed. I'll revisit again tomorrow.
> > > >
> > > > Maybe I need >1 process/thread to trigger this. That would explain why
> > > > I can trigger it with Trinity.
> > >
> > > scratch that last part, I finally just repro'd it with a single process.
> >
> > Thanks for the info, I'll try to think about this some more.
>
> I threw in some debug printks right before that BUG_ON.
> it's always this:
>
> skb->len=31 skb->data_len=0 offset:30 total_len:9
Clearly we fail because 30 > 31 - 2, seems 'offset' is not correct here,
off-by-one?
^ permalink raw reply
* Re: ipv6: handle -EFAULT from skb_copy_bits
From: Dave Jones @ 2016-12-20 19:34 UTC (permalink / raw)
To: David Miller; +Cc: netdev
In-Reply-To: <20161220.132813.435056880928769245.davem@davemloft.net>
On Tue, Dec 20, 2016 at 01:28:13PM -0500, David Miller wrote:
> This has to do with the SKB buffer layout and geometry, not whether
> the packet is "fragmented" in the protocol sense.
>
> So no, this isn't a criteria for packets being filtered out by this
> point.
>
> Can you try to capture what sk->sk_socket->type and
> inet_sk(sk)->hdrincl are set to at the time of the crash?
>
type:3 hdrincl:0
Dave
^ permalink raw reply
* Re: nfc: trf7970a: Prevent repeated polling from crashing the kernel
From: Mark Greer @ 2016-12-20 19:56 UTC (permalink / raw)
To: Justin Bronder
Cc: Geoff Lansberry, linux-wireless-u79uwXL29TY76Z2rM5mHXA,
lauro.venancio-430g2QfJUUCGglJvpFV4uA,
aloisio.almeida-430g2QfJUUCGglJvpFV4uA,
sameo-VuQAYsv1563Yd54FQh9/CA, robh+dt-DgEjT+Ai2ygdnm+yROfE0A,
mark.rutland-5wv7dgnIgG8, netdev-u79uwXL29TY76Z2rM5mHXA,
devicetree-u79uwXL29TY76Z2rM5mHXA,
linux-kernel-u79uwXL29TY76Z2rM5mHXA, Jaret Cantu
In-Reply-To: <20161220191352.GB23496-WrO9gjaJVAZ9BIO3fBtL+FdjMaeQq/Z1QQ4Iyu8u01E@public.gmane.org>
On Tue, Dec 20, 2016 at 02:13:52PM -0500, Justin Bronder wrote:
> On 20/12/16 11:59 -0700, Mark Greer wrote:
> > On Tue, Dec 20, 2016 at 11:16:32AM -0500, Geoff Lansberry wrote:
> > > From: Jaret Cantu <jaret.cantu-jEh4hwF5bVhBDgjK7y7TUQ@public.gmane.org>
> > >
> > > Repeated polling attempts cause a NULL dereference error to occur.
> > > This is because the state of the trf7970a is currently reading but
> > > another request has been made to send a command before it has finished.
> >
> > How is this happening? Was trf7970a_abort_cmd() called and it didn't
> > work right? Was it not called at all and there is a bug in the digital
> > layer? More details please.
> >
> > > The solution is to properly kill the waiting reading (workqueue)
> > > before failing on the send.
> >
> > If the bug is in the calling code, then that is what should get fixed.
> > This seems to be a hack to work-around a digital layer bug.
>
> One of our uses of NFC is to begin polling to read a tag and then stop polling
> (in order to save power) until we know via user interaction that we need to poll
> again. This is typically many minutes later so the power saving is pretty
> significant. However, it's possible that a user will remove the tag before
> reading has completed. We also detect this case and stop polling. I can go
> more into this if necessary but that is what exposed a panic.
>
> You can reproduce using neard and python, in our testing it was very likely to
> occur in 10-100 iterations of the following.:
>
> #!/usr/bin/python
> import time
>
> import dbus
>
> bus = dbus.SystemBus()
> nfc0 = bus.get_object('org.neard', '/org/neard/nfc0')
> props = dbus.Interface(nfc0, 'org.freedesktop.DBus.Properties')
>
> try:
> props.Set('org.neard.Adapter', 'Powered', dbus.Boolean(1))
> except:
> pass
>
> adapter = dbus.Interface(nfc0, 'org.neard.Adapter')
>
> for i in range(1000):
> adapter.StartPollLoop('Initiator')
> time.sleep(0.1)
> adapter.StopPollLoop()
> print(i)
>
> I believe the last time we tested this was around the 4.1 release.
Thanks for the info, Justin, but I was also seeking more information
at the kernel NFC subsystem and trf7970a driver level. This patch
adds code inside an 'if' in the driver whose condition should never
be evaluate to true but apparently it did. How?
Thanks,
Mark
--
--
To unsubscribe from this list: send the line "unsubscribe devicetree" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
^ permalink raw reply
* [RFC][PATCH 0/5 net-next] Rework inet_csk_get_port
From: Josef Bacik @ 2016-12-20 20:06 UTC (permalink / raw)
To: davem, hannes, kraigatgoog, eric.dumazet, tom, netdev,
kernel-team
At some point recently the guys working on our load balancer added the ability
to use SO_REUSEPORT. When they restarted their app with this option enabled
they immediately hit a softlockup on what appeared to be the
inet_bind_bucket->lock. Eventually what all of our debugging and discussion led
us to was the fact that the application comes up without SO_REUSEPORT, shuts
down which creates around 100k twsk's, and then comes up and tries to open a
bunch of sockets using SO_REUSEPORT, which meant traversing the inet_bind_bucket
owners list under the lock. Since this lock is needed for dealing with the
twsk's and basically anything else related to connections we would softlockup,
and sometimes not ever recover.
To solve this problem I did what you see in Path 5/5. Once we have a
SO_REUSEPORT socket on the tb->owners list we know that the socket has no
conflicts with any of the other sockets on that list. So we can add a copy of
the sock_common (really all we need is the recv_saddr but it seemed ugly to copy
just the ipv6, ipv4, and flag to indicate if we were ipv6 only in there so I've
copied the whole common) in order to check subsequent SO_REUSEPORT sockets. If
they match the previous one then we can skip the expensive
inet_csk_bind_conflict check. This is what eliminated the soft lockup that we
were seeing.
Patches 1-4 are cleanups and re-workings. For instance when we specify port ==
0 we need to find an open port, but we would do two passes through
inet_csk_bind_conflict every time we found a possible port. We would also keep
track of the smallest_port value in order to try and use it if we found no
port our first run through. This however made no sense as it would have had to
fail the first pass through inet_csk_bind_conflict, so would not actually pass
the second pass through either. Finally I split the function into two functions
in order to make it easier to read and to distinguish between the two behaviors.
I have tested this on one of our load balancing boxes during peak traffic and it
hasn't fallen over. But this is not my area, so obviously feel free to point
out where I'm being stupid and I'll get it fixed up and retested. Thanks,
Josef
^ permalink raw reply
* [PATCH 2/5 net-next] inet: kill smallest_size and smallest_port
From: Josef Bacik @ 2016-12-20 20:07 UTC (permalink / raw)
To: davem, hannes, kraigatgoog, eric.dumazet, tom, netdev,
kernel-team
In-Reply-To: <1482264424-15439-1-git-send-email-jbacik@fb.com>
In inet_csk_get_port we seem to be using smallest_port to figure out where the
best place to look for a SO_REUSEPORT sk that matches with an existing set of
SO_REUSEPORT's. However if we get to the logic
if (smallest_size != -1) {
port = smallest_port;
goto have_port;
}
we will do a useless search, because we would have already done the
inet_csk_bind_conflict for that port and it would have returned 1, otherwise we
would have gone to found_tb and succeeded. Since this logic makes us do yet
another trip through inet_csk_bind_conflict for a port we know won't work just
delete this code and save us the time.
Signed-off-by: Josef Bacik <jbacik@fb.com>
---
net/ipv4/inet_connection_sock.c | 26 ++++----------------------
1 file changed, 4 insertions(+), 22 deletions(-)
diff --git a/net/ipv4/inet_connection_sock.c b/net/ipv4/inet_connection_sock.c
index 74f6a57..1a1a94bd 100644
--- a/net/ipv4/inet_connection_sock.c
+++ b/net/ipv4/inet_connection_sock.c
@@ -93,7 +93,6 @@ int inet_csk_get_port(struct sock *sk, unsigned short snum)
bool reuse = sk->sk_reuse && sk->sk_state != TCP_LISTEN;
struct inet_hashinfo *hinfo = sk->sk_prot->h.hashinfo;
int ret = 1, attempts = 5, port = snum;
- int smallest_size = -1, smallest_port;
struct inet_bind_hashbucket *head;
struct net *net = sock_net(sk);
int i, low, high, attempt_half;
@@ -103,7 +102,6 @@ int inet_csk_get_port(struct sock *sk, unsigned short snum)
bool reuseport_ok = !!snum;
if (port) {
-have_port:
head = &hinfo->bhash[inet_bhashfn(net, port,
hinfo->bhash_size)];
spin_lock_bh(&head->lock);
@@ -137,8 +135,6 @@ other_half_scan:
* We do the opposite to not pollute connect() users.
*/
offset |= 1U;
- smallest_size = -1;
- smallest_port = low; /* avoid compiler warning */
other_parity_scan:
port = low + offset;
@@ -152,15 +148,6 @@ other_parity_scan:
spin_lock_bh(&head->lock);
inet_bind_bucket_for_each(tb, &head->chain)
if (net_eq(ib_net(tb), net) && tb->port == port) {
- if (((tb->fastreuse > 0 && reuse) ||
- (tb->fastreuseport > 0 &&
- sk->sk_reuseport &&
- !rcu_access_pointer(sk->sk_reuseport_cb) &&
- uid_eq(tb->fastuid, uid))) &&
- (tb->num_owners < smallest_size || smallest_size == -1)) {
- smallest_size = tb->num_owners;
- smallest_port = port;
- }
if (!inet_csk_bind_conflict(sk, tb, false, reuseport_ok))
goto tb_found;
goto next_port;
@@ -171,10 +158,6 @@ next_port:
cond_resched();
}
- if (smallest_size != -1) {
- port = smallest_port;
- goto have_port;
- }
offset--;
if (!(offset & 1))
goto other_parity_scan;
@@ -196,19 +179,18 @@ tb_found:
if (sk->sk_reuse == SK_FORCE_REUSE)
goto success;
- if (((tb->fastreuse > 0 && reuse) ||
+ if ((tb->fastreuse > 0 && reuse) ||
(tb->fastreuseport > 0 &&
!rcu_access_pointer(sk->sk_reuseport_cb) &&
- sk->sk_reuseport && uid_eq(tb->fastuid, uid))) &&
- smallest_size == -1)
+ sk->sk_reuseport && uid_eq(tb->fastuid, uid)))
goto success;
if (inet_csk_bind_conflict(sk, tb, true, reuseport_ok)) {
if ((reuse ||
(tb->fastreuseport > 0 &&
sk->sk_reuseport &&
!rcu_access_pointer(sk->sk_reuseport_cb) &&
- uid_eq(tb->fastuid, uid))) &&
- !snum && smallest_size != -1 && --attempts >= 0) {
+ uid_eq(tb->fastuid, uid))) && !snum &&
+ --attempts >= 0) {
spin_unlock_bh(&head->lock);
goto again;
}
--
2.9.3
^ permalink raw reply related
* [PATCH 1/5 net-next] inet: replace ->bind_conflict with ->rcv_saddr_equal
From: Josef Bacik @ 2016-12-20 20:07 UTC (permalink / raw)
To: davem, hannes, kraigatgoog, eric.dumazet, tom, netdev,
kernel-team
In-Reply-To: <1482264424-15439-1-git-send-email-jbacik@fb.com>
The only difference between inet6_csk_bind_conflict and inet_csk_bind_conflict
is how they check the rcv_saddr. Since we want to be able to check the saddr in
other places just drop the protocol specific ->bind_conflict and replace it with
->rcv_saddr_equal, then make inet_csk_bind_conflict the one true bind conflict
function.
Signed-off-by: Josef Bacik <jbacik@fb.com>
---
include/net/inet6_connection_sock.h | 5 -----
include/net/inet_connection_sock.h | 9 +++------
net/dccp/ipv4.c | 3 ++-
net/dccp/ipv6.c | 2 +-
net/ipv4/inet_connection_sock.c | 22 +++++++-------------
net/ipv4/tcp_ipv4.c | 3 ++-
net/ipv4/udp.c | 1 +
net/ipv6/inet6_connection_sock.c | 40 -------------------------------------
net/ipv6/tcp_ipv6.c | 4 ++--
9 files changed, 18 insertions(+), 71 deletions(-)
diff --git a/include/net/inet6_connection_sock.h b/include/net/inet6_connection_sock.h
index 3212b39..8ec87b6 100644
--- a/include/net/inet6_connection_sock.h
+++ b/include/net/inet6_connection_sock.h
@@ -15,16 +15,11 @@
#include <linux/types.h>
-struct inet_bind_bucket;
struct request_sock;
struct sk_buff;
struct sock;
struct sockaddr;
-int inet6_csk_bind_conflict(const struct sock *sk,
- const struct inet_bind_bucket *tb, bool relax,
- bool soreuseport_ok);
-
struct dst_entry *inet6_csk_route_req(const struct sock *sk, struct flowi6 *fl6,
const struct request_sock *req, u8 proto);
diff --git a/include/net/inet_connection_sock.h b/include/net/inet_connection_sock.h
index ec0479a..9cd43c5 100644
--- a/include/net/inet_connection_sock.h
+++ b/include/net/inet_connection_sock.h
@@ -62,9 +62,9 @@ struct inet_connection_sock_af_ops {
char __user *optval, int __user *optlen);
#endif
void (*addr2sockaddr)(struct sock *sk, struct sockaddr *);
- int (*bind_conflict)(const struct sock *sk,
- const struct inet_bind_bucket *tb,
- bool relax, bool soreuseport_ok);
+ int (*rcv_saddr_equal)(const struct sock *sk1,
+ const struct sock *sk2,
+ bool match_wildcard);
void (*mtu_reduced)(struct sock *sk);
};
@@ -261,9 +261,6 @@ inet_csk_rto_backoff(const struct inet_connection_sock *icsk,
struct sock *inet_csk_accept(struct sock *sk, int flags, int *err);
-int inet_csk_bind_conflict(const struct sock *sk,
- const struct inet_bind_bucket *tb, bool relax,
- bool soreuseport_ok);
int inet_csk_get_port(struct sock *sk, unsigned short snum);
struct dst_entry *inet_csk_route_req(const struct sock *sk, struct flowi4 *fl4,
diff --git a/net/dccp/ipv4.c b/net/dccp/ipv4.c
index 9c67a96..1931324 100644
--- a/net/dccp/ipv4.c
+++ b/net/dccp/ipv4.c
@@ -17,6 +17,7 @@
#include <linux/skbuff.h>
#include <linux/random.h>
+#include <net/addrconf.h>
#include <net/icmp.h>
#include <net/inet_common.h>
#include <net/inet_hashtables.h>
@@ -901,7 +902,7 @@ static const struct inet_connection_sock_af_ops dccp_ipv4_af_ops = {
.getsockopt = ip_getsockopt,
.addr2sockaddr = inet_csk_addr2sockaddr,
.sockaddr_len = sizeof(struct sockaddr_in),
- .bind_conflict = inet_csk_bind_conflict,
+ .rcv_saddr_equal = ipv4_rcv_saddr_equal,
#ifdef CONFIG_COMPAT
.compat_setsockopt = compat_ip_setsockopt,
.compat_getsockopt = compat_ip_getsockopt,
diff --git a/net/dccp/ipv6.c b/net/dccp/ipv6.c
index 4663a01..45242b8 100644
--- a/net/dccp/ipv6.c
+++ b/net/dccp/ipv6.c
@@ -926,7 +926,7 @@ static const struct inet_connection_sock_af_ops dccp_ipv6_af_ops = {
.getsockopt = ipv6_getsockopt,
.addr2sockaddr = inet6_csk_addr2sockaddr,
.sockaddr_len = sizeof(struct sockaddr_in6),
- .bind_conflict = inet6_csk_bind_conflict,
+ .rcv_saddr_equal = ipv6_rcv_saddr_equal,
#ifdef CONFIG_COMPAT
.compat_setsockopt = compat_ipv6_setsockopt,
.compat_getsockopt = compat_ipv6_getsockopt,
diff --git a/net/ipv4/inet_connection_sock.c b/net/ipv4/inet_connection_sock.c
index 5f44fa1..74f6a57 100644
--- a/net/ipv4/inet_connection_sock.c
+++ b/net/ipv4/inet_connection_sock.c
@@ -44,9 +44,9 @@ void inet_get_local_port_range(struct net *net, int *low, int *high)
}
EXPORT_SYMBOL(inet_get_local_port_range);
-int inet_csk_bind_conflict(const struct sock *sk,
- const struct inet_bind_bucket *tb, bool relax,
- bool reuseport_ok)
+static int inet_csk_bind_conflict(const struct sock *sk,
+ const struct inet_bind_bucket *tb,
+ bool relax, bool reuseport_ok)
{
struct sock *sk2;
bool reuse = sk->sk_reuse;
@@ -62,7 +62,6 @@ int inet_csk_bind_conflict(const struct sock *sk,
sk_for_each_bound(sk2, &tb->owners) {
if (sk != sk2 &&
- !inet_v6_ipv6only(sk2) &&
(!sk->sk_bound_dev_if ||
!sk2->sk_bound_dev_if ||
sk->sk_bound_dev_if == sk2->sk_bound_dev_if)) {
@@ -72,23 +71,18 @@ int inet_csk_bind_conflict(const struct sock *sk,
rcu_access_pointer(sk->sk_reuseport_cb) ||
(sk2->sk_state != TCP_TIME_WAIT &&
!uid_eq(uid, sock_i_uid(sk2))))) {
-
- if (!sk2->sk_rcv_saddr || !sk->sk_rcv_saddr ||
- sk2->sk_rcv_saddr == sk->sk_rcv_saddr)
+ if (inet_csk(sk)->icsk_af_ops->rcv_saddr_equal(sk, sk2, true))
break;
}
if (!relax && reuse && sk2->sk_reuse &&
sk2->sk_state != TCP_LISTEN) {
-
- if (!sk2->sk_rcv_saddr || !sk->sk_rcv_saddr ||
- sk2->sk_rcv_saddr == sk->sk_rcv_saddr)
+ if (inet_csk(sk)->icsk_af_ops->rcv_saddr_equal(sk, sk2, true))
break;
}
}
}
return sk2 != NULL;
}
-EXPORT_SYMBOL_GPL(inet_csk_bind_conflict);
/* Obtain a reference to a local port for the given sock,
* if snum is zero it means select any available local port.
@@ -167,8 +161,7 @@ other_parity_scan:
smallest_size = tb->num_owners;
smallest_port = port;
}
- if (!inet_csk(sk)->icsk_af_ops->bind_conflict(sk, tb, false,
- reuseport_ok))
+ if (!inet_csk_bind_conflict(sk, tb, false, reuseport_ok))
goto tb_found;
goto next_port;
}
@@ -209,8 +202,7 @@ tb_found:
sk->sk_reuseport && uid_eq(tb->fastuid, uid))) &&
smallest_size == -1)
goto success;
- if (inet_csk(sk)->icsk_af_ops->bind_conflict(sk, tb, true,
- reuseport_ok)) {
+ if (inet_csk_bind_conflict(sk, tb, true, reuseport_ok)) {
if ((reuse ||
(tb->fastreuseport > 0 &&
sk->sk_reuseport &&
diff --git a/net/ipv4/tcp_ipv4.c b/net/ipv4/tcp_ipv4.c
index 029708f..7608012 100644
--- a/net/ipv4/tcp_ipv4.c
+++ b/net/ipv4/tcp_ipv4.c
@@ -63,6 +63,7 @@
#include <linux/times.h>
#include <linux/slab.h>
+#include <net/addrconf.h>
#include <net/net_namespace.h>
#include <net/icmp.h>
#include <net/inet_hashtables.h>
@@ -1781,7 +1782,7 @@ const struct inet_connection_sock_af_ops ipv4_specific = {
.getsockopt = ip_getsockopt,
.addr2sockaddr = inet_csk_addr2sockaddr,
.sockaddr_len = sizeof(struct sockaddr_in),
- .bind_conflict = inet_csk_bind_conflict,
+ .rcv_saddr_equal = ipv4_rcv_saddr_equal,
#ifdef CONFIG_COMPAT
.compat_setsockopt = compat_ip_setsockopt,
.compat_getsockopt = compat_ip_getsockopt,
diff --git a/net/ipv4/udp.c b/net/ipv4/udp.c
index 2a70c05..6089ea8 100644
--- a/net/ipv4/udp.c
+++ b/net/ipv4/udp.c
@@ -374,6 +374,7 @@ int ipv4_rcv_saddr_equal(const struct sock *sk1, const struct sock *sk2,
}
return 0;
}
+EXPORT_SYMBOL(ipv4_rcv_saddr_equal);
static u32 udp4_portaddr_hash(const struct net *net, __be32 saddr,
unsigned int port)
diff --git a/net/ipv6/inet6_connection_sock.c b/net/ipv6/inet6_connection_sock.c
index 71939a2..7538715 100644
--- a/net/ipv6/inet6_connection_sock.c
+++ b/net/ipv6/inet6_connection_sock.c
@@ -28,46 +28,6 @@
#include <net/inet6_connection_sock.h>
#include <net/sock_reuseport.h>
-int inet6_csk_bind_conflict(const struct sock *sk,
- const struct inet_bind_bucket *tb, bool relax,
- bool reuseport_ok)
-{
- const struct sock *sk2;
- bool reuse = !!sk->sk_reuse;
- bool reuseport = !!sk->sk_reuseport && reuseport_ok;
- kuid_t uid = sock_i_uid((struct sock *)sk);
-
- /* We must walk the whole port owner list in this case. -DaveM */
- /*
- * See comment in inet_csk_bind_conflict about sock lookup
- * vs net namespaces issues.
- */
- sk_for_each_bound(sk2, &tb->owners) {
- if (sk != sk2 &&
- (!sk->sk_bound_dev_if ||
- !sk2->sk_bound_dev_if ||
- sk->sk_bound_dev_if == sk2->sk_bound_dev_if)) {
- if ((!reuse || !sk2->sk_reuse ||
- sk2->sk_state == TCP_LISTEN) &&
- (!reuseport || !sk2->sk_reuseport ||
- rcu_access_pointer(sk->sk_reuseport_cb) ||
- (sk2->sk_state != TCP_TIME_WAIT &&
- !uid_eq(uid,
- sock_i_uid((struct sock *)sk2))))) {
- if (ipv6_rcv_saddr_equal(sk, sk2, true))
- break;
- }
- if (!relax && reuse && sk2->sk_reuse &&
- sk2->sk_state != TCP_LISTEN &&
- ipv6_rcv_saddr_equal(sk, sk2, true))
- break;
- }
- }
-
- return sk2 != NULL;
-}
-EXPORT_SYMBOL_GPL(inet6_csk_bind_conflict);
-
struct dst_entry *inet6_csk_route_req(const struct sock *sk,
struct flowi6 *fl6,
const struct request_sock *req,
diff --git a/net/ipv6/tcp_ipv6.c b/net/ipv6/tcp_ipv6.c
index bee59a6..2f40b98 100644
--- a/net/ipv6/tcp_ipv6.c
+++ b/net/ipv6/tcp_ipv6.c
@@ -1603,7 +1603,7 @@ static const struct inet_connection_sock_af_ops ipv6_specific = {
.getsockopt = ipv6_getsockopt,
.addr2sockaddr = inet6_csk_addr2sockaddr,
.sockaddr_len = sizeof(struct sockaddr_in6),
- .bind_conflict = inet6_csk_bind_conflict,
+ .rcv_saddr_equal = ipv6_rcv_saddr_equal,
#ifdef CONFIG_COMPAT
.compat_setsockopt = compat_ipv6_setsockopt,
.compat_getsockopt = compat_ipv6_getsockopt,
@@ -1634,7 +1634,7 @@ static const struct inet_connection_sock_af_ops ipv6_mapped = {
.getsockopt = ipv6_getsockopt,
.addr2sockaddr = inet6_csk_addr2sockaddr,
.sockaddr_len = sizeof(struct sockaddr_in6),
- .bind_conflict = inet6_csk_bind_conflict,
+ .rcv_saddr_equal = ipv6_rcv_saddr_equal,
#ifdef CONFIG_COMPAT
.compat_setsockopt = compat_ipv6_setsockopt,
.compat_getsockopt = compat_ipv6_getsockopt,
--
2.9.3
^ permalink raw reply related
* [PATCH 3/5 net-next] inet: don't check for bind conflicts twice when searching for a port
From: Josef Bacik @ 2016-12-20 20:07 UTC (permalink / raw)
To: davem, hannes, kraigatgoog, eric.dumazet, tom, netdev,
kernel-team
In-Reply-To: <1482264424-15439-1-git-send-email-jbacik@fb.com>
This is just wasted time, we've already found a tb that doesn't have a bind
conflict, and we don't drop the head lock so scanning again isn't going to give
us a different answer. Instead move the tb->reuse setting logic outside of the
found_tb path and put it in the success: path. Then make it so that we don't
goto again if we find a bind conflict in the found_tb path as we won't reach
this anymore when we are scanning for an ephemeral port.
Signed-off-by: Josef Bacik <jbacik@fb.com>
---
net/ipv4/inet_connection_sock.c | 39 ++++++++++++++++++---------------------
1 file changed, 18 insertions(+), 21 deletions(-)
diff --git a/net/ipv4/inet_connection_sock.c b/net/ipv4/inet_connection_sock.c
index 1a1a94bd..fc9bfe1 100644
--- a/net/ipv4/inet_connection_sock.c
+++ b/net/ipv4/inet_connection_sock.c
@@ -92,7 +92,7 @@ int inet_csk_get_port(struct sock *sk, unsigned short snum)
{
bool reuse = sk->sk_reuse && sk->sk_state != TCP_LISTEN;
struct inet_hashinfo *hinfo = sk->sk_prot->h.hashinfo;
- int ret = 1, attempts = 5, port = snum;
+ int ret = 1, port = snum;
struct inet_bind_hashbucket *head;
struct net *net = sock_net(sk);
int i, low, high, attempt_half;
@@ -100,6 +100,7 @@ int inet_csk_get_port(struct sock *sk, unsigned short snum)
kuid_t uid = sock_i_uid(sk);
u32 remaining, offset;
bool reuseport_ok = !!snum;
+ bool empty_tb = true;
if (port) {
head = &hinfo->bhash[inet_bhashfn(net, port,
@@ -111,7 +112,6 @@ int inet_csk_get_port(struct sock *sk, unsigned short snum)
goto tb_not_found;
}
-again:
attempt_half = (sk->sk_reuse == SK_CAN_REUSE) ? 1 : 0;
other_half_scan:
inet_get_local_port_range(net, &low, &high);
@@ -148,8 +148,12 @@ other_parity_scan:
spin_lock_bh(&head->lock);
inet_bind_bucket_for_each(tb, &head->chain)
if (net_eq(ib_net(tb), net) && tb->port == port) {
- if (!inet_csk_bind_conflict(sk, tb, false, reuseport_ok))
- goto tb_found;
+ if (hlist_empty(&tb->owners))
+ goto success;
+ if (!inet_csk_bind_conflict(sk, tb, false, reuseport_ok)) {
+ empty_tb = false;
+ goto success;
+ }
goto next_port;
}
goto tb_not_found;
@@ -184,23 +188,12 @@ tb_found:
!rcu_access_pointer(sk->sk_reuseport_cb) &&
sk->sk_reuseport && uid_eq(tb->fastuid, uid)))
goto success;
- if (inet_csk_bind_conflict(sk, tb, true, reuseport_ok)) {
- if ((reuse ||
- (tb->fastreuseport > 0 &&
- sk->sk_reuseport &&
- !rcu_access_pointer(sk->sk_reuseport_cb) &&
- uid_eq(tb->fastuid, uid))) && !snum &&
- --attempts >= 0) {
- spin_unlock_bh(&head->lock);
- goto again;
- }
+ if (inet_csk_bind_conflict(sk, tb, true, reuseport_ok))
goto fail_unlock;
- }
- if (!reuse)
- tb->fastreuse = 0;
- if (!sk->sk_reuseport || !uid_eq(tb->fastuid, uid))
- tb->fastreuseport = 0;
- } else {
+ empty_tb = false;
+ }
+success:
+ if (empty_tb) {
tb->fastreuse = reuse;
if (sk->sk_reuseport) {
tb->fastreuseport = 1;
@@ -208,8 +201,12 @@ tb_found:
} else {
tb->fastreuseport = 0;
}
+ } else {
+ if (!reuse)
+ tb->fastreuse = 0;
+ if (!sk->sk_reuseport || !uid_eq(tb->fastuid, uid))
+ tb->fastreuseport = 0;
}
-success:
if (!inet_csk(sk)->icsk_bind_hash)
inet_bind_hash(sk, tb, port);
WARN_ON(inet_csk(sk)->icsk_bind_hash != tb);
--
2.9.3
^ permalink raw reply related
* [PATCH 4/5 net-next] inet: split inet_csk_get_port into two functions
From: Josef Bacik @ 2016-12-20 20:07 UTC (permalink / raw)
To: davem, hannes, kraigatgoog, eric.dumazet, tom, netdev,
kernel-team
In-Reply-To: <1482264424-15439-1-git-send-email-jbacik@fb.com>
inet_csk_get_port does two different things, it either scans for an open port,
or it tries to see if the specified port is available for use. Since these two
operations have different rules and are basically independent lets split them
into two different functions to make them both more readable.
Signed-off-by: Josef Bacik <jbacik@fb.com>
---
net/ipv4/inet_connection_sock.c | 72 +++++++++++++++++++++++++++--------------
1 file changed, 47 insertions(+), 25 deletions(-)
diff --git a/net/ipv4/inet_connection_sock.c b/net/ipv4/inet_connection_sock.c
index fc9bfe1..d3ccf62 100644
--- a/net/ipv4/inet_connection_sock.c
+++ b/net/ipv4/inet_connection_sock.c
@@ -84,34 +84,21 @@ static int inet_csk_bind_conflict(const struct sock *sk,
return sk2 != NULL;
}
-/* Obtain a reference to a local port for the given sock,
- * if snum is zero it means select any available local port.
- * We try to allocate an odd port (and leave even ports for connect())
+/*
+ * Find an open port number for the socket. Returns with the
+ * inet_bind_hashbucket lock held.
*/
-int inet_csk_get_port(struct sock *sk, unsigned short snum)
+static struct inet_bind_hashbucket *
+inet_csk_find_open_port(struct sock *sk, struct inet_bind_bucket **tb_ret, int *port_ret)
{
- bool reuse = sk->sk_reuse && sk->sk_state != TCP_LISTEN;
struct inet_hashinfo *hinfo = sk->sk_prot->h.hashinfo;
- int ret = 1, port = snum;
+ int port = 0;
struct inet_bind_hashbucket *head;
struct net *net = sock_net(sk);
int i, low, high, attempt_half;
struct inet_bind_bucket *tb;
- kuid_t uid = sock_i_uid(sk);
u32 remaining, offset;
- bool reuseport_ok = !!snum;
- bool empty_tb = true;
- if (port) {
- head = &hinfo->bhash[inet_bhashfn(net, port,
- hinfo->bhash_size)];
- spin_lock_bh(&head->lock);
- inet_bind_bucket_for_each(tb, &head->chain)
- if (net_eq(ib_net(tb), net) && tb->port == port)
- goto tb_found;
-
- goto tb_not_found;
- }
attempt_half = (sk->sk_reuse == SK_CAN_REUSE) ? 1 : 0;
other_half_scan:
inet_get_local_port_range(net, &low, &high);
@@ -150,13 +137,12 @@ other_parity_scan:
if (net_eq(ib_net(tb), net) && tb->port == port) {
if (hlist_empty(&tb->owners))
goto success;
- if (!inet_csk_bind_conflict(sk, tb, false, reuseport_ok)) {
- empty_tb = false;
+ if (!inet_csk_bind_conflict(sk, tb, false, false))
goto success;
- }
goto next_port;
}
- goto tb_not_found;
+ tb = NULL;
+ goto success;
next_port:
spin_unlock_bh(&head->lock);
cond_resched();
@@ -171,8 +157,44 @@ next_port:
attempt_half = 2;
goto other_half_scan;
}
- return ret;
+ return NULL;
+success:
+ *port_ret = port;
+ *tb_ret = tb;
+ return head;
+}
+/* Obtain a reference to a local port for the given sock,
+ * if snum is zero it means select any available local port.
+ * We try to allocate an odd port (and leave even ports for connect())
+ */
+int inet_csk_get_port(struct sock *sk, unsigned short snum)
+{
+ bool reuse = sk->sk_reuse && sk->sk_state != TCP_LISTEN;
+ struct inet_hashinfo *hinfo = sk->sk_prot->h.hashinfo;
+ int ret = 1, port = snum;
+ struct inet_bind_hashbucket *head;
+ struct net *net = sock_net(sk);
+ struct inet_bind_bucket *tb = NULL;
+ kuid_t uid = sock_i_uid(sk);
+ bool empty_tb = true;
+
+ if (!port) {
+ head = inet_csk_find_open_port(sk, &tb, &port);
+ if (!head)
+ return 1;
+ if (!tb)
+ goto tb_not_found;
+ if (!hlist_empty(&tb->owners))
+ empty_tb = false;
+ goto success;
+ }
+ head = &hinfo->bhash[inet_bhashfn(net, port,
+ hinfo->bhash_size)];
+ spin_lock_bh(&head->lock);
+ inet_bind_bucket_for_each(tb, &head->chain)
+ if (net_eq(ib_net(tb), net) && tb->port == port)
+ goto tb_found;
tb_not_found:
tb = inet_bind_bucket_create(hinfo->bind_bucket_cachep,
net, head, port);
@@ -188,7 +210,7 @@ tb_found:
!rcu_access_pointer(sk->sk_reuseport_cb) &&
sk->sk_reuseport && uid_eq(tb->fastuid, uid)))
goto success;
- if (inet_csk_bind_conflict(sk, tb, true, reuseport_ok))
+ if (inet_csk_bind_conflict(sk, tb, true, true))
goto fail_unlock;
empty_tb = false;
}
--
2.9.3
^ permalink raw reply related
* [PATCH 5/5 net-next] inet: reset tb->fastreuseport when adding a reuseport sk
From: Josef Bacik @ 2016-12-20 20:07 UTC (permalink / raw)
To: davem, hannes, kraigatgoog, eric.dumazet, tom, netdev,
kernel-team
In-Reply-To: <1482264424-15439-1-git-send-email-jbacik@fb.com>
If we have non reuseport sockets on a tb we will set tb->fastreuseport to 0 and
never set it again. Which means that in the future if we end up adding a bunch
of reuseport sk's to that tb we'll have to do the expensive scan every time.
Instead add a sock_common to the tb so we know what reuseport sk succeeded last.
Once one sk has made it onto the list we know that there are no potential bind
conflicts on the owners list that match that sk's rcv_addr. So copy the sk's
common into our tb->fastsock and set tb->fastruseport to FASTREUSESOCK_STRICT so
we know we have to do an extra check for subsequent reuseport sockets and skip
the expensive bind conflict check.
Signed-off-by: Josef Bacik <jbacik@fb.com>
---
include/net/inet_hashtables.h | 4 ++++
net/ipv4/inet_connection_sock.c | 53 +++++++++++++++++++++++++++++++++++++----
2 files changed, 53 insertions(+), 4 deletions(-)
diff --git a/include/net/inet_hashtables.h b/include/net/inet_hashtables.h
index 50f635c..b776401 100644
--- a/include/net/inet_hashtables.h
+++ b/include/net/inet_hashtables.h
@@ -74,12 +74,16 @@ struct inet_ehash_bucket {
* users logged onto your box, isn't it nice to know that new data
* ports are created in O(1) time? I thought so. ;-) -DaveM
*/
+#define FASTREUSEPORT_ANY 1
+#define FASTREUSEPORT_STRICT 2
+
struct inet_bind_bucket {
possible_net_t ib_net;
unsigned short port;
signed char fastreuse;
signed char fastreuseport;
kuid_t fastuid;
+ struct sock_common fastsock;
int num_owners;
struct hlist_node node;
struct hlist_head owners;
diff --git a/net/ipv4/inet_connection_sock.c b/net/ipv4/inet_connection_sock.c
index d3ccf62..9e29fad 100644
--- a/net/ipv4/inet_connection_sock.c
+++ b/net/ipv4/inet_connection_sock.c
@@ -164,6 +164,32 @@ success:
return head;
}
+static inline int sk_reuseport_match(struct inet_bind_bucket *tb,
+ struct sock *sk)
+{
+ struct sock *sk2 = (struct sock *)&tb->fastsock;
+ kuid_t uid = sock_i_uid(sk);
+
+ if (tb->fastreuseport <= 0)
+ return 0;
+ if (!sk->sk_reuseport)
+ return 0;
+ if (rcu_access_pointer(sk->sk_reuseport_cb))
+ return 0;
+ if (!uid_eq(tb->fastuid, uid))
+ return 0;
+ /* We only need to check the rcv_saddr if this tb was once marked
+ * without fastreuseport and then was reset, as we can only know that
+ * the fastsock has no potential bind conflicts with the rest of the
+ * possible socks on the owners list.
+ */
+ if (tb->fastreuseport == FASTREUSEPORT_ANY)
+ return 1;
+ if (!inet_csk(sk)->icsk_af_ops->rcv_saddr_equal(sk, sk2, true))
+ return 0;
+ return 1;
+}
+
/* Obtain a reference to a local port for the given sock,
* if snum is zero it means select any available local port.
* We try to allocate an odd port (and leave even ports for connect())
@@ -206,9 +232,7 @@ tb_found:
goto success;
if ((tb->fastreuse > 0 && reuse) ||
- (tb->fastreuseport > 0 &&
- !rcu_access_pointer(sk->sk_reuseport_cb) &&
- sk->sk_reuseport && uid_eq(tb->fastuid, uid)))
+ sk_reuseport_match(tb, sk))
goto success;
if (inet_csk_bind_conflict(sk, tb, true, true))
goto fail_unlock;
@@ -220,14 +244,35 @@ success:
if (sk->sk_reuseport) {
tb->fastreuseport = 1;
tb->fastuid = uid;
+ memcpy(&tb->fastsock, &sk->__sk_common,
+ sizeof(struct sock_common));
} else {
tb->fastreuseport = 0;
}
} else {
if (!reuse)
tb->fastreuse = 0;
- if (!sk->sk_reuseport || !uid_eq(tb->fastuid, uid))
+ if (sk->sk_reuseport) {
+ /* We didn't match or we don't have fastreuseport set on
+ * the tb, but we have sk_reuseport set on this socket
+ * and we know that there are no bind conflicts with
+ * this socket in this tb, so reset our tb's reuseport
+ * settings so that any subsequent sockets that match
+ * our current socket will be put on the fast path.
+ *
+ * If we reset we need to set FASTREUSEPORT_STRICT so we
+ * do extra checking for all subsequent sk_reuseport
+ * socks.
+ */
+ if (!sk_reuseport_match(tb, sk)) {
+ tb->fastreuseport = FASTREUSEPORT_STRICT;
+ tb->fastuid = uid;
+ memcpy(&tb->fastsock, &sk->__sk_common,
+ sizeof(struct sock_common));
+ }
+ } else {
tb->fastreuseport = 0;
+ }
}
if (!inet_csk(sk)->icsk_bind_hash)
inet_bind_hash(sk, tb, port);
--
2.9.3
^ permalink raw reply related
* [ANNOUNCE] nftables 0.7 release
From: Pablo Neira Ayuso @ 2016-12-20 20:46 UTC (permalink / raw)
To: netfilter-devel; +Cc: netdev, netfilter, netfilter-announce, lwn
[-- Attachment #1: Type: text/plain, Size: 10356 bytes --]
Hi!
The Netfilter project proudly presents:
nftables 0.7
This release contains many accumulated bug fixes and new features
available up to the (upcoming) Linux 4.10-rc1 kernel release.
* Facilitate migration from iptables to nftables:
At compilation time, you have to pass this option.
# ./configure --with-xtables
And libxtables needs to be installed in your system. This allows you
to list a ruleset containing xt extensions loaded through
iptables-compat-restore tool. The nft tool provides a native
translation for iptables extensions (if available).
* Add new fib expression, which can be used to obtain the output
interface from the route table based on either source or destination
address of a packet. This can be used to e.g. add reverse path
filtering, eg. drop if not coming from the same interface packet
arrived on:
# nft add rule x prerouting fib saddr . iif oif eq 0 drop
Accept only if from eth:
# nft add rule x prerouting fib saddr . iif oif eq "eth0" accept
Accept if from any valid interface:
# nft add rule x prerouting fib saddr oif accept
Querying of address type is also supported, this can be used
to only accept packets to addresses configured in the same
interface, eg.
# nft add rule x prerouting fib daddr . iif type local accept
Its also possible to use mark and verdict map, eg,
# nft add rule x prerouting \
meta mark set 0xdead fib daddr . mark type vmap {
blackhole : drop,
prohibit : drop,
unicast : accept
}
* Support hashing of any arbitrary key combination, eg.
# nft add rule x y \
dnat to jhash ip saddr . tcp dport mod 2 map { \
0 : 192.168.20.100, \
1 : 192.168.30.100 \
}
Another usecase: Set packet marks based on any arbitrary hashing.
* Add number generation support. Useful for round-robin packet mark
setting, eg.
# nft add rule filter prerouting meta mark set numgen inc mod 2
You can also specify an offset to indicate from what value you want
to start from.
The modulus provides the scale of the counting sequence. You can
also use this from maps, eg.
# nft add rule nat prerouting \
dnat to numgen inc mod 2 map { 0 : 192.168.10.100, 1 : 192.168.20.200 }
So this is distributing new connections in a round-robin fashion
between 192.168.10.100 and 192.168.20.200. Don't forget the special NAT
chain semantics: Only the first packet evaluates the rule, follow up
packets rely on conntrack to apply the NAT information.
You can also emulate flow distribution with different backend weights
using intervals, eg.
# nft add rule nat prerouting \
dnat to numgen inc mod 10 map { 0-5 : 192.168.10.100, 6-9 : 192.168.20.200 }
* Add quota support, eg.
# nft add rule filter input \
flow table http { ip saddr timeout 60s quota over 50 mbytes } drop
This creates a flow table, where every flow gets a quota of 50
mbytes. You can also from use simple rules too to enforce quotas, of
course.
* Introduce routing expression, for routing related data with support
for nexthop (i.e. the directly connected IP address that an outgoing
packet is sent to), which can be used either for matching or accounting, eg.
# nft add rule filter postrouting \
ip daddr 192.168.1.0/24 rt nexthop != 192.168.0.1 drop
This will drop any traffic to 192.168.1.0/24 that is not routed via
192.168.0.1.
# nft add rule filter postrouting \
flow table acct { rt nexthop timeout 600s counter }
# nft add rule ip6 filter postrouting \
flow table acct { rt nexthop timeout 600s counter }
These rules count outgoing traffic per nexthop. Note that the timeout
releases an entry if no traffic is seen for this nexthop within 10
minutes.
* Notrack support, to explicitly skip connection tracking for matching
packets, eg.
# nft add rule ip raw prerouting tcp dport { 80, 443 } notrack
So you can skip tracking for http and https traffic.
* Support to set non-byte bound packet header fields, including
checksum adjustment, eg. ip6 ecn set 1.
* Add 'create set' and 'create element' commands, eg.
# nft add set x y { type ipv4_addr\; }
# nft create set x y { type ipv4_addr\; }
<cmdline>:1:1-35: Error: Could not process rule: File exists
create set x y { type ipv4_addr; }
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
# nft add set x y { type ipv4_addr\; }
#
So 'create' bails out if the set already exists, while 'add'
doesn't, for more ergonomic usage as several users requested on
the mailing list.
* Allow to use variable reference for set element definitions, eg.
# cat ruleset.nft
define s-ext-2-int = { 10.10.10.10 . 25, 10.10.10.10 . 143 }
table inet forward {
set s-ext-2-int {
type ipv4_addr . inet_service
elements = $s-ext-2-int
}
}
# nft -f ruleset.nft
Useful to improve ruleset maintainability, as you can split out
variable and set definitions from the filtering policy itself.
* Allow to use variable definitions from element commands, eg.
define whitelist_v4 = { 1.1.1.1 }
table inet filter {
set whitelist_v4 { type ipv4_addr; }
}
add element inet filter whitelist_v4 $whitelist_v4
* Add support to flush set. You can use this new command to remove all
existing elements in a set, eg.
# nft flush set filter xyz
Note that this requires (upcoming) Linux kernel 4.10-rc versions.
* Inverted set lookups, eg. tcp dport != { 80, 443 }.
* Honor absolute and relative paths via include file, where:
include "./ruleset.nft"
refers to a file in the working directory.
include "ruleset.nft"
refers to a file in the nftables root path (via sysconfdir), and:
include "/etc/nftables/ruleset.nft"
provides an absolute reference to the file that need to be included.
This also solves an ambiguity if the same file name is used both under
sysconfdir and the current working directory.
* Support log flags, to enable logging TCP sequence and options:
# nft add rule x y log flags tcp sequence,options
... IP options, eg:
# nft add rule x y log flags ip options
... socket UID, eg.
# nft add rule x y log flags skuid
... decide ethernet link layer address, eg.
# nft add rule x y log flags ether
... or simply set on all flags:
# nft add rule x y log flags all
* tc classid parser support, eg.
nft add rule filter forward meta priority abcd:1234
* Allow numeric connlabels, so if connlabel still works with undefined
labels, eg. ct label set 2.
* Document log, reject, counter, meta, limit, nat, ct, payload and
queue statements from nft(8) manpage.
Bugfixes
========
Not strictly limited to this list below, but some highlights:
* Allow split table definitions, eg.
# cat ruleset.nft
table inet filter {
chain ssh {
type filter hook input priority 0; policy accept;
tcp dport ssh accept;
}
}
table inet filter {
chain input {
type filter hook input priority 1; policy drop;
}
}
# nft -f ruleset.nft
* Use new range expression to represent inverted intervals, eg.
ip saddr != 1.1.1.1-2.2.2.2, since previously generated bytecode was
not correct.
* Solve endianness problems with link layer address.
* Fix parser to keep map flag around on definition.
* Skip timeout attribute in dynamic set updates, other kernel bails
out with EINVAL.
* Restore parsing of dynamic set element updates.
* The time datatype now uses milliseconds, as the kernel expects.
* Allow numeric interface index numbers, eg. in meta iif, oif.
* Fix monitor trace crash with netdev family.
* Flow table with concatenation fixes.
* Keep element comments around when using set intervals.
* Fixed memory corruption in userspace when deleting lots of elements
in one go via nft -f.
* Several nft internal cache fixes, including cache reset on 'flush
ruleset'.
* Restore parens on right-hand side of relational expression.
* Replace getnameinfo() by internal lookup table, so we don't rely on
/etc/services anymore for service names, so we restrict them to
a well-known set that is supported by our scanner. You can list
service names via 'nft describe tcp dport'.
* Display symbol table values in the right hostbyte order and
decimal/hexadecimal representation.
* Fix a nasty bug in the set interval code triggering huge memory
consumption in userspace for set and map intervals with runtime
updates.
We also got lots more tests added to our infrastructure to catch up
regressions.
Syntax updates
==============
Several minor syntax updates, although previous syntax has been
preserved by now to facilitate transition, the new one is prefered:
* Consistency grammar fixes: 'snat' and 'dnat' now require 'to', eg.
snat to 1.2.3.4. For consistency with existing statements such as
redirect, masquerade, dup and fwd. Moreover, add colon after 'to' in
'redirect' for consistency with nat and masq statements.
* Allow ct l3proto/protocol without direction since they are unrelated
to the direction.
* Explicit ruleset exportation, eg. nft export ruleset json, for
consistency with other existing ruleset commands.
* Always quote user-defined strings from rules when listing them.
* Support for RFC2732 IPv6 address format with brackets, eg.
dnat to [2001:838:35f:1::]:80
* Allow strings starting by underscores and dots in user-define
strings, conforming with POSIX.1-2008 (which is simultaneously IEEE
Std 1003.1-2008).
Resources
=========
The nftables code can be obtained from:
* http://netfilter.org/projects/nftables/downloads.html
* ftp://ftp.netfilter.org/pub/nftables
* git://git.netfilter.org/nftables
To build the code, libnftnl 1.0.7 and libmnl >= 1.0.2 are required:
* http://netfilter.org/projects/libnftnl/index.html
* http://netfilter.org/projects/libmnl/index.html
Visit our wikipage for user documentation at:
* http://wiki.nftables.org
For the manpage reference, check man(8) nft.
In case of bugs and feature request, file them via:
* https://bugzilla.netfilter.org
Make sure you create no duplicates already, thanks!
Happy holidays!
[-- Attachment #2: changes-nftables-0.7.txt --]
[-- Type: text/plain, Size: 9551 bytes --]
Anatole Denis (7):
evaluate: Add set to cache only when well-formed
tests: Add regression test for malformed sets
Revert "evaluate: check for NULL datatype in rhs in lookup expr"
src: Interpret OP_NEQ against a set as OP_LOOKUP
tests/py: Unmask negative set lookup
rule: Introduce helper function cache_flush
evaluate: Update cache on flush ruleset
Anders K. Pedersen (4):
rt: introduce routing expression
Replace tests/files/expr-rt with Python based tests, and replace ether type with meta nfproto, which generates a bit fewer instructions.
evaluate: Allow concatenation of rt nexthop etc.
doc: fix synopsis for ct expression
Arturo Borrero (3):
tests: shell: delete unused variable in run-tests.sh
tests: shell: cleanup tempfile handling in testcases/sets/cache_handling_0
tests: shell: run-tests.sh: use src/nft binary by default
Arturo Borrero Gonzalez (12):
tests: shell: update kernel modules to clean
xt: update Arturo Borrero Gonzalez email address
tests: shell: delete useless stderr output in testcase
tests: shell: introduce the cache testcases directory
tests: shell: add a new testcase for ruleset loading bug
tests: shell: add testcases for comments in set elements
tests: shell: allow to execute a single testcase
tests: shell: testcase for adding many set elements
tests: shell: testcase for deleting many set elements
tests: shell: another testcase for deleting many set elements
tests: shell: add a testcase for many defines
tests: shell: add testcase for different defines usage
Carlos Falgueras García (1):
src: Simplify parser rule_spec tree
Elise Lennion (4):
datatype: Replace getnameinfo() by internal lookup table
datatype: Display pre-defined inet_service values in host byte order
datatype: Display pre-defined inet_service values in decimal base
expression: Show the base which pre-defined constants are displayed
Florian Westphal (30):
payload: don't update protocol context if we can't find a description
meta: add random support
meta: add tests for meta random
ct: use nftables sysconf location for connlabel configuration
tests: add basic payload tests
tests: add ether payload set test
netlink: add __binop_adjust helper
payload: print base and raw values for unknown payloads
evaluate: add small helper to check if payload expr needs binop adjustment
evaluate: add support to set IPv6 non-byte header fields
netlink: decode payload statment
tests: ip6 dscp, flowlabel and ecn test cases
netlink: make checksum fixup work with odd-sized header fields
tests: ip payload set support for ecn and dscp
ct: allow numeric conntrack labels
ct: display bit number instead of raw value
doc: update meta expression
doc: payload and conntrack statement
datatype: ll: use big endian byte ordering
tests: catch ordering issue w. ether set
payload: remove byteorder conversion
meta: permit numeric interface type
netlink: fix monitor trace crash with netdev family
meta: fix pkttype name and add 'other' symbol
utils: provide snprintf helper macro
ct: allow resolving ct keys at run time
meta: allow resolving meta keys at run time
src: add fib expression
Revert "tests: py: nft-tests.py: Add function for loading and removing kernel modules"
bison: remove old log level tokens
Jon Jensen (1):
Correct description of -n/--numeric option
Laura Garcia Liebana (5):
doc: Update datatypes
src: add offset attribute for numgen expression
netlink: fix linearize numgen type
src: make hash seed attribute optional
src: add offset attribute for hash expression
Liping Zhang (14):
tests: shell: make testcases which using tcp/udp port more rubost
tests: shell: add endless jump loop tests
parser_bison: keep snat/dnat existing syntax unchanged
tests: shell: add testcase for reject expr
meta: fix memory leak in tc classid parser
tests: py: replace "eth0" with "lo" in dup expr tests
src: fix compile error due to _UNTIL renamed to _MODULUS in libnftnl
tests: py: add more test cases for queue expr
tests: py: fix numgen case failed due to changes in libnftnl
src: support ct l3proto/protocol without direction syntax
ct: fix "ct l3proto/protocol" syntax broken
log: rename the log level "warning" to "warn"
src: add log flags syntax support
tests: shell: add test case for inserting element into verdict map
Manuel Johannes Messner (3):
tests: py: nft-tests.py: Add function for loading and removing kernel modules
tests: py: any: Make tests more generic by using other interfaces
tests: py: any: Remove duplicate tests
Nicholas Vinson (1):
nft: configure.ac: Replace magic dblatex dep.
Pablo Neira (2):
src: expose delinearize/linearize structures and stmt_error()
src: trigger layer 4 checksum when pseudoheader fields are modified
Pablo Neira Ayuso (71):
src: use new definitions from libnftnl
segtree: don't check for overlaps if set definition is empty
tests: shell: cover transactions via nft -f using flat syntax
datatype: time_type should send milliseconds to userspace
parser_bison: restore parsing of dynamic set element updates
netlink_linearize: skip NFTNL_EXPR_DYNSET_TIMEOUT attribute if timeout is unset
include: cache ip_tables.h, ip6_tables.h, arp_tables.h and ebtables.h
src: add xt compat support
parser_bison: fix typo in symbol redefinition error reporting
tests: shell: make sure split table definition works via nft -f
xt: use struct xt_xlate_{mt,tg}_params
parser_bison: keep map flag around when flags are specified
scanner: honor absolute and relative paths via include file
scanner: don't fall back on current directory if include is not found
scanner: don't break line on include error message
tests: tests to include files
ct: add missing slash to connlabel path
ct: release ct_label table on exit
src: quote user-defined strings when used from rule selectors
src: add 'to' for snat and dnat
src: support for RFC2732 IPv6 address format with brackets
parser_bison: missing token string in QUOTED_ASTERISK and ASTERISK_STRING
scanner: allow strings starting by underscores and dots
scanner: remove range expression
src: rename datatype name from tc_handle to classid
src: simplify classid printing using %x instead of %04x
src: meta priority support using tc classid
parser_bison: redirect to :port for consistency with nat/masq statement
parser_bison: explicit indication on export ruleset
src: add create set command
tests: shell: cover add and create set command
src: create element command
tests: shell: cover add and create set command
include: refresh uapi/linux/netfilter/nf_tables.h copy
tests: py: adapt it to new add element command semantics
src: add quota statement
src: add numgen expression
src: add hash expression
evaluate: add expr_evaluate_integer()
evaluate: validate maximum hash and numgen value
parser_bison: add variable_expr rule
parser_bison: allow variable references in set elements definition
tests: py: adapt netlink bytecode output of numgen and hash
evaluate: display expression, statement and command name on debug
netlink_delinearize: Avoid potential null pointer deref
doc: nft: add my copyright statement to the manpage
doc: nft: document log, reject, counter, meta, limit, nat and queue statements
src: use new range expression for != [a,b] intervals
parser_bison: allow to use variable to add/create/delete elements
src: don't need keyword for log level
parser: add offset keyword and parser rule
tests/py: add missing payload test for numgen offset
netlink_linearize: skip set element expression in flow table key
segtree: keep element comments in set intervals
tests: py: add some testcases for log flags
tests: py: missing range conversion in icmpv6
src: add notrack support
mnl: use nftnl_set_elems_nlmsg_build_payload_iter() when deleting elements
include: refresh nf_tables.h header
datatype: honor -nn option from inet_service_type_print()
evaluate: return ctx->table from table_lookup_global()
src: add support to flush sets
segtree: wrong prefix expression length on interval_map_decompose()
segtree: don't trigger error on exact overlaps
mnl: don't send empty set elements netlink message to kernel
tests: py: update quota and payload
netlink_linearize: fix IPv6 layer 4 checksum mangling
mnl: add mnl_nft_setelem_batch_flush() and use it from netlink_flush_setelems()
xt: use NFTNL_* definitions
configure: Bump version to v0.7
include: Missing noinst_HEADERS updates
Phil Sutter (5):
evaluate: Fix datalen checks in expr_evaluate_string()
evaluate: reject: Have a generic fix for missing network context
evaluate: Avoid undefined behaviour in concat_subtype_id()
parser_bison: Allow parens on RHS of relational_expr
tests: py: Test TCP flags match with parentheses
^ permalink raw reply
* [GIT] Networking
From: David Miller @ 2016-12-20 21:02 UTC (permalink / raw)
To: torvalds; +Cc: akpm, netdev, linux-kernel
1) Use rb_entry() instead of hardcoded container_of(), from Geliang Tang.
2) Use correct memory barriers in stammac driver, from Pavel Machek.
3) Fix assoc bind address handling in SCTP, from Xin Long.
4) Make the length check for UFO handling consistent between
__ip_append_data() and ip_finish_output(), from Zheng Li.
5) HSI driver compatible strings were busted fro hix5hd2, from Dongpo
Li.
6) Handle devm_ioremap() errors properly in cavium driver, from Arvind
Yadav.
Please pull, thanks a lot!
The following changes since commit 52f40e9d657cc126b766304a5dd58ad73b02ff46:
Merge git://git.kernel.org/pub/scm/linux/kernel/git/davem/net (2016-12-17 20:17:04 -0800)
are available in the git repository at:
git://git.kernel.org/pub/scm/linux/kernel/git/davem/net.git
for you to fetch changes up to a763f78cea845c91b8d91f93dabf70c407635dc5:
RDS: use rb_entry() (2016-12-20 14:22:49 -0500)
----------------------------------------------------------------
Arvind Yadav (1):
net: ethernet: cavium: octeon: octeon_mgmt: Handle return NULL error from devm_ioremap
David S. Miller (4):
Merge branch 'phy-broken-modes'
Merge branch 'fsl-fixes'
Merge branch 'hix5hd2_gmac-compatible-string'
Merge branch 'sctp-fixes'
Dongpo Li (2):
net: hix5hd2_gmac: fix compatible strings name
ARM: dts: hix5hd2: don't change the existing compatible string
Geliang Tang (4):
net/mlx5: use rb_entry()
net_sched: sch_fq: use rb_entry()
net_sched: sch_netem: use rb_entry()
RDS: use rb_entry()
Jarno Rajahalme (1):
openvswitch: Add a missing break statement.
Madalin Bucur (4):
fsl/fman: fix 1G support for QSGMII interfaces
powerpc: fsl/fman: remove fsl,fman from of_device_ids[]
fsl/fman: A007273 only applies to PPC SoCs
fsl/fman: enable compilation on ARM64
Pavel Machek (1):
stmmac: fix memory barriers
Tobias Klauser (1):
ethernet: sfc: Add Kconfig entry for vendor Solarflare
WingMan Kwok (2):
net: netcp: ethss: fix errors in ethtool ops
net: netcp: ethss: fix 10gbe host port tx pri map configuration
Xin Long (2):
sctp: reduce indent level in sctp_copy_local_addr_list
sctp: not copying duplicate addrs to the assoc's bind address list
jbrunet (3):
net: phy: fix sign type error in genphy_config_eee_advert
net: phy: use boolean dt properties for eee broken modes
dt: bindings: net: use boolean dt properties for eee broken modes
zheng li (1):
ipv4: Should use consistent conditional judgement for ip fragment in __ip_append_data and ip_finish_output
Documentation/devicetree/bindings/net/hisilicon-hix5hd2-gmac.txt | 13 ++++++++-----
Documentation/devicetree/bindings/net/phy.txt | 10 ++++++++--
arch/arm/boot/dts/hisi-x5hd2.dtsi | 4 ++--
arch/powerpc/platforms/85xx/corenet_generic.c | 3 ---
drivers/net/ethernet/Kconfig | 1 -
drivers/net/ethernet/cavium/octeon/octeon_mgmt.c | 6 ++++++
drivers/net/ethernet/freescale/fman/Kconfig | 2 +-
drivers/net/ethernet/freescale/fman/fman.c | 15 +++++++++++++++
drivers/net/ethernet/freescale/fman/mac.c | 1 +
drivers/net/ethernet/hisilicon/hix5hd2_gmac.c | 13 +++++++------
drivers/net/ethernet/mellanox/mlx5/core/fs_counters.c | 2 +-
drivers/net/ethernet/sfc/Kconfig | 21 +++++++++++++++++++++
drivers/net/ethernet/stmicro/stmmac/dwmac4_descs.c | 4 ++--
drivers/net/ethernet/stmicro/stmmac/enh_desc.c | 2 +-
drivers/net/ethernet/stmicro/stmmac/stmmac_main.c | 8 ++++----
drivers/net/ethernet/ti/netcp_ethss.c | 24 ++++++++++++++++++------
drivers/net/phy/phy_device.c | 22 +++++++++++++++++-----
include/dt-bindings/net/mdio.h | 19 -------------------
net/ipv4/ip_output.c | 2 +-
net/openvswitch/flow_netlink.c | 1 +
net/rds/rdma.c | 2 +-
net/sched/sch_fq.c | 14 +++++++-------
net/sched/sch_netem.c | 2 +-
net/sctp/bind_addr.c | 3 +++
net/sctp/protocol.c | 40 ++++++++++++++++++++++------------------
25 files changed, 148 insertions(+), 86 deletions(-)
delete mode 100644 include/dt-bindings/net/mdio.h
^ permalink raw reply
* Re: HalfSipHash Acceptable Usage
From: Theodore Ts'o @ 2016-12-20 21:36 UTC (permalink / raw)
To: Jason A. Donenfeld
Cc: Jean-Philippe Aumasson, Hannes Frederic Sowa, LKML, Eric Biggers,
Daniel J . Bernstein, David Laight, David Miller, Andi Kleen,
George Spelvin, kernel-hardening, Andy Lutomirski,
Linux Crypto Mailing List, Tom Herbert, Vegard Nossum, Netdev,
Linus Torvalds
In-Reply-To: <CAHmME9rPmH=wP_eHYopt8ZPG9TSN7bos3fGOuqKL2HjQW-2SWA@mail.gmail.com>
On Mon, Dec 19, 2016 at 06:32:44PM +0100, Jason A. Donenfeld wrote:
> 1) Anything that requires actual long-term security will use
> SipHash2-4, with the 64-bit output and the 128-bit key. This includes
> things like TCP sequence numbers. This seems pretty uncontroversial to
> me. Seem okay to you?
Um, why do TCP sequence numbers need long-term security? So long as
you rekey every 5 minutes or so, TCP sequence numbers don't need any
more security than that, since even if you break the key used to
generate initial sequence numbers seven a minute or two later, any
pending TCP connections will have timed out long before.
See the security analysis done in RFC 6528[1], where among other
things, it points out why MD5 is acceptable with periodic rekeying,
although there is the concern that this could break certain hueristics
used when establishing new connections during the TIME-WAIT state.
[1] https://tools.ietf.org/html/rfc6528
- Ted
^ permalink raw reply
* [PATCH net-next 00/10] netcp: enhancements and minor fixes
From: Murali Karicheri @ 2016-12-20 22:09 UTC (permalink / raw)
To: netdev-u79uwXL29TY76Z2rM5mHXA, linux-omap-u79uwXL29TY76Z2rM5mHXA,
grygorii.strashko-l0cyMroinI0, mugunthanvnm-l0cyMroinI0,
linux-kernel-u79uwXL29TY76Z2rM5mHXA, arnd-r2nGTMty4D4,
davem-fT/PcQaiUtIeIZ0/mPfg9Q, devicetree-u79uwXL29TY76Z2rM5mHXA,
mark.rutland-5wv7dgnIgG8, robh+dt-DgEjT+Ai2ygdnm+yROfE0A
This series is for net-next. This propagates enhancements and minor
bug fixes from internal version of the driver to keep the upstream
in sync. Please review and apply if this looks good.
Tested on all of K2HK/E/L boards.
Thanks
Murali Karicheri
Michael Scherban (1):
net: netcp: store network statistics in 64 bits
Murali Karicheri (7):
net: netcp: extract eflag from desc for rx_hook handling
net: netcp: remove the redundant memmov()
net: netcp: ethss: get phy-handle only if link interface is MAC-to-PHY
net: netcp: use hw capability to remove FCS word from rx packets
net: netcp: ale: update to support unknown vlan controls for NU switch
net: netcp: ale: use ale_status to size the ale table
net: netcp: ale: add proper ale entry mask bits for netcp switch ALE
WingMan Kwok (2):
net: netcp: ethss: add support of subsystem register region regmap
net: netcp: ethss: add support of 10gbe pcsr link status
.../devicetree/bindings/net/keystone-netcp.txt | 19 +-
drivers/net/ethernet/ti/cpsw_ale.c | 180 ++++++++++++++++---
drivers/net/ethernet/ti/cpsw_ale.h | 17 +-
drivers/net/ethernet/ti/netcp.h | 21 +++
drivers/net/ethernet/ti/netcp_core.c | 102 ++++++++---
drivers/net/ethernet/ti/netcp_ethss.c | 200 +++++++++++++++++----
include/linux/soc/ti/knav_dma.h | 2 +
7 files changed, 456 insertions(+), 85 deletions(-)
--
1.9.1
--
To unsubscribe from this list: send the line "unsubscribe devicetree" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
^ permalink raw reply
* [PATCH net-next 01/10] net: netcp: ethss: add support of subsystem register region regmap
From: Murali Karicheri @ 2016-12-20 22:09 UTC (permalink / raw)
To: netdev, linux-omap, grygorii.strashko, mugunthanvnm, linux-kernel,
arnd, davem, devicetree, mark.rutland, robh+dt
In-Reply-To: <1482271793-7671-1-git-send-email-m-karicheri2@ti.com>
From: WingMan Kwok <w-kwok2@ti.com>
10gbe phy driver needs to access the 10gbe subsystem control
register during phy initialization. To facilitate the shared
access of the subsystem register region between the 10gbe Ethernet
driver and the phy driver, this patch adds support of the
subsystem register region defined by a syscon node in the dts.
Although there is no shared access to the gbe subsystem register
region, using syscon for that is for the sake of consistency.
This change is backward compatible with previously released gbe
devicetree bindings.
Signed-off-by: WingMan Kwok <w-kwok2@ti.com>
Signed-off-by: Murali Karicheri <m-karicheri2@ti.com>
Signed-off-by: Sekhar Nori <nsekhar@ti.com>
---
.../devicetree/bindings/net/keystone-netcp.txt | 16 ++-
drivers/net/ethernet/ti/netcp_ethss.c | 140 +++++++++++++++++----
2 files changed, 127 insertions(+), 29 deletions(-)
diff --git a/Documentation/devicetree/bindings/net/keystone-netcp.txt b/Documentation/devicetree/bindings/net/keystone-netcp.txt
index 04ba1dc..0854a73 100644
--- a/Documentation/devicetree/bindings/net/keystone-netcp.txt
+++ b/Documentation/devicetree/bindings/net/keystone-netcp.txt
@@ -72,20 +72,24 @@ Required properties:
"ti,netcp-gbe-2" for 1GbE N NetCP 1.5 (N=2)
"ti,netcp-xgbe" for 10 GbE
+- syscon-subsys: phandle to syscon node of the switch
+ subsystem registers.
+
- reg: register location and the size for the following register
regions in the specified order.
- switch subsystem registers
+ - sgmii module registers
- sgmii port3/4 module registers (only for NetCP 1.4)
- switch module registers
- serdes registers (only for 10G)
NetCP 1.4 ethss, here is the order
- index #0 - switch subsystem registers
+ index #0 - sgmii module registers
index #1 - sgmii port3/4 module registers
index #2 - switch module registers
NetCP 1.5 ethss 9 port, 5 port and 2 port
- index #0 - switch subsystem registers
+ index #0 - sgmii module registers
index #1 - switch module registers
index #2 - serdes registers
@@ -145,6 +149,11 @@ Optional properties:
Example binding:
+gbe_subsys: subsys@2090000 {
+ compatible = "syscon";
+ reg = <0x02090000 0x100>;
+};
+
netcp: netcp@2000000 {
reg = <0x2620110 0x8>;
reg-names = "efuse";
@@ -163,7 +172,8 @@ netcp: netcp@2000000 {
ranges;
gbe@90000 {
label = "netcp-gbe";
- reg = <0x90000 0x300>, <0x90400 0x400>, <0x90800 0x700>;
+ syscon-subsys = <&gbe_subsys>;
+ reg = <0x90100 0x200>, <0x90400 0x200>, <0x90800 0x700>;
/* enable-ale; */
tx-queue = <648>;
tx-channel = <8>;
diff --git a/drivers/net/ethernet/ti/netcp_ethss.c b/drivers/net/ethernet/ti/netcp_ethss.c
index c7e547e..473edda1 100644
--- a/drivers/net/ethernet/ti/netcp_ethss.c
+++ b/drivers/net/ethernet/ti/netcp_ethss.c
@@ -19,9 +19,11 @@
*/
#include <linux/io.h>
+#include <linux/mfd/syscon.h>
#include <linux/module.h>
#include <linux/of_mdio.h>
#include <linux/of_address.h>
+#include <linux/regmap.h>
#include <linux/if_vlan.h>
#include <linux/ptp_classify.h>
#include <linux/net_tstamp.h>
@@ -43,7 +45,10 @@
#define GBE_MODULE_NAME "netcp-gbe"
#define GBE_SS_VERSION_14 0x4ed21104
+/* for devicetree backward compatible only */
#define GBE_SS_REG_INDEX 0
+
+#define GBE_SGMII_REG_INDEX 0
#define GBE_SGMII34_REG_INDEX 1
#define GBE_SM_REG_INDEX 2
/* offset relative to base of GBE_SS_REG_INDEX */
@@ -71,9 +76,11 @@
#define IS_SS_ID_NU(d) \
(GBE_IDENT((d)->ss_version) == GBE_SS_ID_NU)
-#define GBENU_SS_REG_INDEX 0
+#define GBENU_SGMII_REG_INDEX 0
#define GBENU_SM_REG_INDEX 1
+/* offset relative to base of GBE_SS_REG_INDEX */
#define GBENU_SGMII_MODULE_OFFSET 0x100
+/* offset relative to base of GBENU_SM_REG_INDEX */
#define GBENU_HOST_PORT_OFFSET 0x1000
#define GBENU_SLAVE_PORT_OFFSET 0x2000
#define GBENU_EMAC_OFFSET 0x2330
@@ -82,13 +89,12 @@
#define GBENU_ALE_OFFSET 0x1e000
#define GBENU_HOST_PORT_NUM 0
#define GBENU_NUM_ALE_ENTRIES 1024
-#define GBENU_SGMII_MODULE_SIZE 0x100
/* 10G Ethernet SS defines */
#define XGBE_MODULE_NAME "netcp-xgbe"
#define XGBE_SS_VERSION_10 0x4ee42100
-#define XGBE_SS_REG_INDEX 0
+#define XGBE_SGMII_REG_INDEX 0
#define XGBE_SM_REG_INDEX 1
#define XGBE_SERDES_REG_INDEX 2
@@ -173,6 +179,7 @@
#define XGBE_SET_REG_OFS(p, rb, rn) p->rb##_ofs.rn = \
offsetof(struct xgbe##_##rb, rn)
#define GBE_REG_ADDR(p, rb, rn) (p->rb + p->rb##_ofs.rn)
+#define GBE_REG_OFS(p, rb, rn) ((p)->rb##_ofs.rn)
#define HOST_TX_PRI_MAP_DEFAULT 0x00000000
@@ -225,6 +232,7 @@
/* The PTP event messages - Sync, Delay_Req, Pdelay_Req, and Pdelay_Resp. */
#define EVENT_MSG_BITS (BIT(0) | BIT(1) | BIT(2) | BIT(3))
#endif /* CONFIG_TI_CPTS */
+#define SGMII_MODULE_SIZE 0x100
struct xgbe_ss_regs {
u32 id_ver;
@@ -716,7 +724,9 @@ struct gbe_priv {
u32 ss_version;
u32 stats_en_mask;
- void __iomem *ss_regs;
+ struct regmap *ss_regmap;
+ struct regmap *pcsr_regmap;
+ void __iomem *ss_regs;
void __iomem *switch_regs;
void __iomem *host_port_regs;
void __iomem *ale_reg;
@@ -2192,7 +2202,7 @@ static void gbe_port_config(struct gbe_priv *gbe_dev, struct gbe_slave *slave,
int max_rx_len)
{
void __iomem *rx_maxlen_reg;
- u32 xgmii_mode;
+ int ret;
if (max_rx_len > NETCP_MAX_FRAME_SIZE)
max_rx_len = NETCP_MAX_FRAME_SIZE;
@@ -2200,9 +2210,16 @@ static void gbe_port_config(struct gbe_priv *gbe_dev, struct gbe_slave *slave,
/* Enable correct MII mode at SS level */
if ((gbe_dev->ss_version == XGBE_SS_VERSION_10) &&
(slave->link_interface >= XGMII_LINK_MAC_PHY)) {
- xgmii_mode = readl(GBE_REG_ADDR(gbe_dev, ss_regs, control));
- xgmii_mode |= (1 << slave->slave_num);
- writel(xgmii_mode, GBE_REG_ADDR(gbe_dev, ss_regs, control));
+ ret = regmap_update_bits(gbe_dev->ss_regmap,
+ GBE_REG_OFS(gbe_dev, ss_regs, control),
+ 1 << slave->slave_num,
+ 1 << slave->slave_num);
+
+ if (ret) {
+ dev_err(gbe_dev->dev,
+ "regmap update xgmii mode bit Failed\n");
+ return;
+ }
}
if (IS_SS_ID_MU(gbe_dev))
@@ -3127,35 +3144,46 @@ static int set_xgbe_ethss10_priv(struct gbe_priv *gbe_dev,
void __iomem *regs;
int ret, i;
- ret = of_address_to_resource(node, XGBE_SS_REG_INDEX, &res);
+ gbe_dev->ss_regmap = syscon_regmap_lookup_by_phandle(node,
+ "syscon-subsys");
+
+ if (IS_ERR(gbe_dev->ss_regmap)) {
+ dev_err(gbe_dev->dev,
+ "subsys regmap lookup failed: %ld\n",
+ PTR_ERR(gbe_dev->ss_regmap));
+ return PTR_ERR(gbe_dev->ss_regmap);
+ }
+
+ ret = of_address_to_resource(node, XGBE_SM_REG_INDEX, &res);
if (ret) {
dev_err(gbe_dev->dev,
- "Can't xlate xgbe of node(%s) ss address at %d\n",
- node->name, XGBE_SS_REG_INDEX);
+ "Can't xlate xgbe of node(%s) sm address at %d\n",
+ node->name, XGBE_SM_REG_INDEX);
return ret;
}
regs = devm_ioremap_resource(gbe_dev->dev, &res);
if (IS_ERR(regs)) {
- dev_err(gbe_dev->dev, "Failed to map xgbe ss register base\n");
+ dev_err(gbe_dev->dev, "Failed to map xgbe sm register base\n");
return PTR_ERR(regs);
}
- gbe_dev->ss_regs = regs;
+ gbe_dev->switch_regs = regs;
- ret = of_address_to_resource(node, XGBE_SM_REG_INDEX, &res);
+ ret = of_address_to_resource(node, XGBE_SGMII_REG_INDEX, &res);
if (ret) {
dev_err(gbe_dev->dev,
- "Can't xlate xgbe of node(%s) sm address at %d\n",
- node->name, XGBE_SM_REG_INDEX);
+ "Can't xlate xgbe of node(%s) sgmii address at %d\n",
+ node->name, XGBE_SGMII_REG_INDEX);
return ret;
}
regs = devm_ioremap_resource(gbe_dev->dev, &res);
if (IS_ERR(regs)) {
- dev_err(gbe_dev->dev, "Failed to map xgbe sm register base\n");
+ dev_err(gbe_dev->dev,
+ "Failed to map xgbe sgmii register base\n");
return PTR_ERR(regs);
}
- gbe_dev->switch_regs = regs;
+ gbe_dev->sgmii_port_regs = regs;
ret = of_address_to_resource(node, XGBE_SERDES_REG_INDEX, &res);
if (ret) {
@@ -3171,6 +3199,8 @@ static int set_xgbe_ethss10_priv(struct gbe_priv *gbe_dev,
return PTR_ERR(regs);
}
gbe_dev->xgbe_serdes_regs = regs;
+ gbe_dev->sgmii_port34_regs = gbe_dev->sgmii_port_regs +
+ (2 * SGMII_MODULE_SIZE);
gbe_dev->num_stats_mods = gbe_dev->max_num_ports;
gbe_dev->et_stats = xgbe10_et_stats;
@@ -3195,9 +3225,9 @@ static int set_xgbe_ethss10_priv(struct gbe_priv *gbe_dev,
}
gbe_dev->ss_version = XGBE_SS_VERSION_10;
- gbe_dev->sgmii_port_regs = gbe_dev->ss_regs +
- XGBE10_SGMII_MODULE_OFFSET;
- gbe_dev->host_port_regs = gbe_dev->ss_regs + XGBE10_HOST_PORT_OFFSET;
+
+ gbe_dev->host_port_regs = gbe_dev->switch_regs +
+ XGBE10_HOST_PORT_OFFSET;
for (i = 0; i < gbe_dev->max_num_ports; i++)
gbe_dev->hw_stats_regs[i] = gbe_dev->switch_regs +
@@ -3228,8 +3258,8 @@ static int set_xgbe_ethss10_priv(struct gbe_priv *gbe_dev,
return 0;
}
-static int get_gbe_resource_version(struct gbe_priv *gbe_dev,
- struct device_node *node)
+static int get_gbe_resource_version_ss_regs(struct gbe_priv *gbe_dev,
+ struct device_node *node)
{
struct resource res;
void __iomem *regs;
@@ -3248,8 +3278,27 @@ static int get_gbe_resource_version(struct gbe_priv *gbe_dev,
dev_err(gbe_dev->dev, "Failed to map gbe register base\n");
return PTR_ERR(regs);
}
+
gbe_dev->ss_regs = regs;
gbe_dev->ss_version = readl(gbe_dev->ss_regs);
+ gbe_dev->ss_regmap = NULL;
+ return 0;
+}
+
+static int get_gbe_resource_version(struct gbe_priv *gbe_dev,
+ struct device_node *node)
+{
+ gbe_dev->ss_regmap = syscon_regmap_lookup_by_phandle(node,
+ "syscon-subsys");
+ if (IS_ERR(gbe_dev->ss_regmap)) {
+ dev_dbg(gbe_dev->dev,
+ "subsys regmap lookup failed: %ld. try reg property\n",
+ PTR_ERR(gbe_dev->ss_regmap));
+ return get_gbe_resource_version_ss_regs(gbe_dev, node);
+ }
+
+ regmap_read(gbe_dev->ss_regmap, 0, &gbe_dev->ss_version);
+ gbe_dev->ss_regs = NULL;
return 0;
}
@@ -3260,6 +3309,27 @@ static int set_gbe_ethss14_priv(struct gbe_priv *gbe_dev,
void __iomem *regs;
int i, ret;
+ if (gbe_dev->ss_regs) {
+ gbe_dev->sgmii_port_regs = gbe_dev->ss_regs +
+ GBE13_SGMII_MODULE_OFFSET;
+ } else {
+ ret = of_address_to_resource(node, GBE_SGMII_REG_INDEX, &res);
+ if (ret) {
+ dev_err(gbe_dev->dev,
+ "Can't translate of gbe node(%s) address at index %d\n",
+ node->name, GBE_SGMII_REG_INDEX);
+ return ret;
+ }
+
+ regs = devm_ioremap_resource(gbe_dev->dev, &res);
+ if (IS_ERR(regs)) {
+ dev_err(gbe_dev->dev,
+ "Failed to map gbe sgmii port register base\n");
+ return PTR_ERR(regs);
+ }
+ gbe_dev->sgmii_port_regs = regs;
+ }
+
ret = of_address_to_resource(node, GBE_SGMII34_REG_INDEX, &res);
if (ret) {
dev_err(gbe_dev->dev,
@@ -3314,7 +3384,6 @@ static int set_gbe_ethss14_priv(struct gbe_priv *gbe_dev,
return -ENOMEM;
}
- gbe_dev->sgmii_port_regs = gbe_dev->ss_regs + GBE13_SGMII_MODULE_OFFSET;
gbe_dev->host_port_regs = gbe_dev->switch_regs + GBE13_HOST_PORT_OFFSET;
/* K2HK has only 2 hw stats modules visible at a time, so
@@ -3402,14 +3471,33 @@ static int set_gbenu_ethss_priv(struct gbe_priv *gbe_dev,
}
gbe_dev->switch_regs = regs;
- gbe_dev->sgmii_port_regs = gbe_dev->ss_regs + GBENU_SGMII_MODULE_OFFSET;
+ if (gbe_dev->ss_regs) {
+ gbe_dev->sgmii_port_regs = gbe_dev->ss_regs +
+ GBENU_SGMII_MODULE_OFFSET;
+ } else {
+ ret = of_address_to_resource(node, GBENU_SGMII_REG_INDEX, &res);
+ if (ret) {
+ dev_err(gbe_dev->dev,
+ "Can't translate of gbenu node(%s) addr at index %d\n",
+ node->name, GBENU_SGMII_REG_INDEX);
+ return ret;
+ }
+
+ regs = devm_ioremap_resource(gbe_dev->dev, &res);
+ if (IS_ERR(regs)) {
+ dev_err(gbe_dev->dev,
+ "Failed to map gbenu sgmii port register base\n");
+ return PTR_ERR(regs);
+ }
+ gbe_dev->sgmii_port_regs = regs;
+ }
/* Although sgmii modules are mem mapped to one contiguous
* region on GBENU devices, setting sgmii_port34_regs allows
* consistent code when accessing sgmii api
*/
gbe_dev->sgmii_port34_regs = gbe_dev->sgmii_port_regs +
- (2 * GBENU_SGMII_MODULE_SIZE);
+ (2 * SGMII_MODULE_SIZE);
gbe_dev->host_port_regs = gbe_dev->switch_regs + GBENU_HOST_PORT_OFFSET;
--
1.9.1
^ permalink raw reply related
* [PATCH net-next 02/10] net: netcp: ethss: add support of 10gbe pcsr link status
From: Murali Karicheri @ 2016-12-20 22:09 UTC (permalink / raw)
To: netdev, linux-omap, grygorii.strashko, mugunthanvnm, linux-kernel,
arnd, davem, devicetree, mark.rutland, robh+dt
In-Reply-To: <1482271793-7671-1-git-send-email-m-karicheri2@ti.com>
From: WingMan Kwok <w-kwok2@ti.com>
The 10GBASE-R Physical Coding Sublayer (PCS-R) module provides
functionality of a physical coding sublayer (PCS) on data being
transferred between a demuxed XGMII and SerDes supporting a 16
or 32 bit interface. From the driver point of view, whether
a ethernet link is up or not depends also on the status of the
block-lock bit of the PCSR. This patch adds the checking of that
bit in order to determine the link status.
Signed-off-by: WingMan Kwok <w-kwok2@ti.com>
Signed-off-by: Murali Karicheri <m-karicheri2@ti.com>
Signed-off-by: Sekhar Nori <nsekhar@ti.com>
---
.../devicetree/bindings/net/keystone-netcp.txt | 3 ++
drivers/net/ethernet/ti/netcp_ethss.c | 37 ++++++++++++++++++++--
2 files changed, 37 insertions(+), 3 deletions(-)
diff --git a/Documentation/devicetree/bindings/net/keystone-netcp.txt b/Documentation/devicetree/bindings/net/keystone-netcp.txt
index 0854a73..57fc13f 100644
--- a/Documentation/devicetree/bindings/net/keystone-netcp.txt
+++ b/Documentation/devicetree/bindings/net/keystone-netcp.txt
@@ -75,6 +75,9 @@ Required properties:
- syscon-subsys: phandle to syscon node of the switch
subsystem registers.
+- syscon-pcsr: (10gbe only) phandle to syscon node of the
+ switch PCSR registers.
+
- reg: register location and the size for the following register
regions in the specified order.
- switch subsystem registers
diff --git a/drivers/net/ethernet/ti/netcp_ethss.c b/drivers/net/ethernet/ti/netcp_ethss.c
index 473edda1..cb48f88 100644
--- a/drivers/net/ethernet/ti/netcp_ethss.c
+++ b/drivers/net/ethernet/ti/netcp_ethss.c
@@ -63,6 +63,12 @@
#define GBE13_ALE_OFFSET 0x600
#define GBE13_HOST_PORT_NUM 0
#define GBE13_NUM_ALE_ENTRIES 1024
+/* offset relative to PCSR regmap */
+#define XGBE10_PCSR_OFFSET(x) ((x) * 0x80)
+#define XGBE10_PCSR_RX_STATUS(x) (XGBE10_PCSR_OFFSET(x) + 0x0C)
+
+#define XGBE10_PCSR_BLOCK_LOCK_MASK BIT(30)
+#define XGBE10_PCSR_BLOCK_LOCK_SHIFT 30
/* 1G Ethernet NU SS defines */
#define GBENU_MODULE_NAME "netcp-gbenu"
@@ -2111,6 +2117,10 @@ static void netcp_ethss_link_state_action(struct gbe_priv *gbe_dev,
if (phy)
phy_print_status(phy);
+ else if (slave->link_interface == XGMII_LINK_MAC_MAC_FORCED) {
+ netdev_printk(KERN_INFO, ndev,
+ "Link is %s\n", (up ? "Up" : "Down"));
+ }
}
static bool gbe_phy_link_status(struct gbe_slave *slave)
@@ -2123,18 +2133,29 @@ static void netcp_ethss_update_link_state(struct gbe_priv *gbe_dev,
struct net_device *ndev)
{
int sp = slave->slave_num;
- int phy_link_state, sgmii_link_state = 1, link_state;
+ int phy_link_state, sw_link_state = 1, link_state, ret;
+ u32 pcsr_rx_stat;
if (!slave->open)
return;
if (!SLAVE_LINK_IS_XGMII(slave)) {
- sgmii_link_state =
+ sw_link_state =
netcp_sgmii_get_port_link(SGMII_BASE(gbe_dev, sp), sp);
+ } else if (slave->link_interface == XGMII_LINK_MAC_MAC_FORCED) {
+ /* read status from pcsr status reg */
+ ret = regmap_read(gbe_dev->pcsr_regmap,
+ XGBE10_PCSR_RX_STATUS(sp), &pcsr_rx_stat);
+
+ if (ret)
+ return;
+
+ sw_link_state = (pcsr_rx_stat & XGBE10_PCSR_BLOCK_LOCK_MASK) >>
+ XGBE10_PCSR_BLOCK_LOCK_SHIFT;
}
phy_link_state = gbe_phy_link_status(slave);
- link_state = phy_link_state & sgmii_link_state;
+ link_state = phy_link_state & sw_link_state;
if (atomic_xchg(&slave->link_state, link_state) != link_state)
netcp_ethss_link_state_action(gbe_dev, ndev, slave,
@@ -3154,6 +3175,16 @@ static int set_xgbe_ethss10_priv(struct gbe_priv *gbe_dev,
return PTR_ERR(gbe_dev->ss_regmap);
}
+ gbe_dev->pcsr_regmap = syscon_regmap_lookup_by_phandle(node,
+ "syscon-pcsr");
+
+ if (IS_ERR(gbe_dev->pcsr_regmap)) {
+ dev_err(gbe_dev->dev,
+ "pcsr regmap lookup failed: %ld\n",
+ PTR_ERR(gbe_dev->pcsr_regmap));
+ return PTR_ERR(gbe_dev->pcsr_regmap);
+ }
+
ret = of_address_to_resource(node, XGBE_SM_REG_INDEX, &res);
if (ret) {
dev_err(gbe_dev->dev,
--
1.9.1
^ permalink raw reply related
* [PATCH net-next 04/10] net: netcp: remove the redundant memmov()
From: Murali Karicheri @ 2016-12-20 22:09 UTC (permalink / raw)
To: netdev, linux-omap, grygorii.strashko, mugunthanvnm, linux-kernel,
arnd, davem, devicetree, mark.rutland, robh+dt
In-Reply-To: <1482271793-7671-1-git-send-email-m-karicheri2@ti.com>
The psdata is populated with command data by netcp modules
to the tail of the buffer and set_words() copy the same
to the front of the psdata. So remove the redundant memmov
function call.
Signed-off-by: Murali Karicheri <m-karicheri2@ti.com>
---
drivers/net/ethernet/ti/netcp_core.c | 6 +++---
1 file changed, 3 insertions(+), 3 deletions(-)
diff --git a/drivers/net/ethernet/ti/netcp_core.c b/drivers/net/ethernet/ti/netcp_core.c
index a136c56..286fd8d 100644
--- a/drivers/net/ethernet/ti/netcp_core.c
+++ b/drivers/net/ethernet/ti/netcp_core.c
@@ -1226,9 +1226,9 @@ static int netcp_tx_submit_skb(struct netcp_intf *netcp,
/* psdata points to both native-endian and device-endian data */
__le32 *psdata = (void __force *)p_info.psdata;
- memmove(p_info.psdata, p_info.psdata + p_info.psdata_len,
- p_info.psdata_len);
- set_words(p_info.psdata, p_info.psdata_len, psdata);
+ set_words((u32 *)psdata +
+ (KNAV_DMA_NUM_PS_WORDS - p_info.psdata_len),
+ p_info.psdata_len, psdata);
tmp |= (p_info.psdata_len & KNAV_DMA_DESC_PSLEN_MASK) <<
KNAV_DMA_DESC_PSLEN_SHIFT;
}
--
1.9.1
^ permalink raw reply related
page: next (older) | prev (newer) | latest
- recent:[subjects (threaded)|topics (new)|topics (active)]
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox