On Wed, May 22, 2019 at 02:24 AM CEST, Daniel T. Lee wrote:
> This commit fixes style problem in samples/bpf/bpf_load.c
>
> Styles that have been changed are:
> - Magic string use of 'DEBUGFS'
> - Useless zero initialization of a global variable
> - Minor style fix with whitespace
>
> Signed-off
Hi all,
I'm currently working on a custom board with the imx6ull processor and the 6390
switching chip. This is our hardware setup.
- -MAC
| i.MX | MAC | PHY | PHY | PHY || 88E6390 |
| 6ULL |-|KSZ8081
The user can change the device_name with the IMSETDEVNAME ioctl, but we
need to ensure that the user's name is NUL terminated. Otherwise it
could result in a buffer overflow when we copy the name back to the user
with IMGETDEVINFO ioctl.
I also changed two strcpy() calls which handle the name to
On Tue, 2019-05-21 at 14:59 +0200, Lorenzo Bianconi wrote:
> Currently if we add a filter to the ingress qdisc (e.g matchall) the
> filter data are reported even in the egress path. The issue can be
> triggered with the following reproducer:
>
> $tc qdisc add dev lo ingress
> $tc filter add dev lo
Add three tests to test_verifier/basic_instr that make sure that the
high 32-bits of the destination register is cleared after an ALU32
and/or/xor.
Signed-off-by: Björn Töpel
---
.../selftests/bpf/verifier/basic_instr.c | 39 +++
1 file changed, 39 insertions(+)
diff --git
Once psock gets unlinked from its sock (sk_psock_drop), user-space can
still trigger a call to sk->sk_write_space by setting TCP_NOTSENT_LOWAT
socket option. This causes a null-ptr-deref because we try to read
psock->saved_write_space from sk_psock_write_space:
> On Tue, 2019-05-21 at 14:59 +0200, Lorenzo Bianconi wrote:
> > Currently if we add a filter to the ingress qdisc (e.g matchall) the
> > filter data are reported even in the egress path. The issue can be
> > triggered with the following reproducer:
> >
[...]
> > diff --git a/net/sched/sch_ingre
We are now providing business & personal loans:
-Rate starting at: 2.05%.
-Flexible repayment: up to 30 years.
For more information and application, please reply.
> To unsubscribe please reply with "unsubscribe" as subject.
On 5/21/2019 18:42, Juliana Rodrigueiro wrote:
So I ask myself, how actually feasible is it to gamble the usage of "ethtool"
to turn on or off TSO every time the network configuration changes?
Hello Juliana,
There are many PCH2 devices with different SKU's. Not all devices have
this problem (T
On 05/22/2019 12:20 PM, Lorenzo Bianconi wrote:
>> On Tue, 2019-05-21 at 14:59 +0200, Lorenzo Bianconi wrote:
>>> Currently if we add a filter to the ingress qdisc (e.g matchall) the
>>> filter data are reported even in the egress path. The issue can be
>>> triggered with the following reproducer:
Hi John,
On Tue, May 21, 2019 at 10:07 PM CEST, John Fastabend wrote:
> Jakub Sitnicki wrote:
>> Hi Daniel,
>>
>> On Tue, Feb 19, 2019 at 05:00 PM CET, Daniel Borkmann wrote:
>> > On 02/11/2019 10:09 AM, Jakub Sitnicki wrote:
>> >> Backlog work for psock (sk_psock_backlog) might sleep while waitin
Hi Jakub,
Thanks for all the fixes. This version is much better.
On 5/22/2019 4:57 AM, Jakub Kicinski wrote:
> Describe existing kernel TLS offload (added back in Linux 4.19) -
> the mechanism, the expected behavior and the notable corner cases.
>
> This documentation is mostly targeting hardwar
Our customer reports several kernel crashes all preceding with message
of tx queue timed out:
NETDEV WATCHDOG: eth2 (enic): transmit queue 0 timed out
Error message of one of those crashes:
BUG: unable to handle kernel paging request at a007e090
...
Call Trace:
[] skb_release_head_state+0x
> On 05/22/2019 12:20 PM, Lorenzo Bianconi wrote:
> >> On Tue, 2019-05-21 at 14:59 +0200, Lorenzo Bianconi wrote:
> >>> Currently if we add a filter to the ingress qdisc (e.g matchall) the
> >>> filter data are reported even in the egress path. The issue can be
> >>> triggered with the following re
Hi,
what exactly do you mean with this DT binding ? i cant find this for
meson gxbb somewhere in the DTS ?
regards,
Simon
Am 13.05.2019 um 11:07 schrieb Jose Abreu:
> From: Simon Huelck
> Date: Sat, May 11, 2019 at 15:53:34
>
>> ethtool -S gave me some counts for mmc_rx_fifo_overflow, which i
Here's another attempt (the first, horribly broken one, here [1]) to
move the XDP_QUERY_PROG{,_HW} code out from the drivers to generic
netdev code.
The first patch in the series move the XDP query functionality, and
the second remove XDP_QUERY_PROG{,_HW} from all drivers.
Please refer to the ind
From: Björn Töpel
Remove all use of XDP_QUERY_PROG{,_HW}, since it was moved to the
generic code path.
Signed-off-by: Björn Töpel
---
drivers/net/ethernet/broadcom/bnxt/bnxt_xdp.c | 4
.../net/ethernet/cavium/thunder/nicvf_main.c | 3 ---
.../net/ethernet/freescale/dpaa2/dpaa2-eth.c |
From: Björn Töpel
All XDP capable drivers need to implement the XDP_QUERY_PROG{,_HW}
command of ndo_bpf. The query code is fairly generic. This commit
refactors the query code up from the drivers to the netdev level.
The struct net_device has gained four new members tracking the XDP
program, the
Björn Töpel writes:
> From: Björn Töpel
>
> All XDP capable drivers need to implement the XDP_QUERY_PROG{,_HW}
> command of ndo_bpf. The query code is fairly generic. This commit
> refactors the query code up from the drivers to the netdev level.
>
> The struct net_device has gained four new mem
This series add two improvements for the XSKMAP, used by AF_XDP
sockets.
1. Automatic cleanup when an AF_XDP socket goes out of scope. Instead
of manually cleaning out the "released" state socket from the map,
this is done automatically. This mimics the SOCKMAP behavior; Each
socket track
From: Björn Töpel
The XSKMAP did not honor the BPF_EXIST/BPF_NOEXIST flags when updating
an entry. This patch addressed that.
Signed-off-by: Björn Töpel
---
kernel/bpf/xskmap.c | 18 +++---
1 file changed, 15 insertions(+), 3 deletions(-)
diff --git a/kernel/bpf/xskmap.c b/kernel/
From: Björn Töpel
When an AF_XDP socket is released/closed the XSKMAP still holds a
reference to the socket in a "released" state. The socket will still
use the netdev queue resource, and block newly created sockets from
attaching to that queue, but no user application can access the
fill/complet
On 21/05/2019 19:52:55+0300, Claudiu Manoil wrote:
> Doing kmalloc in atomic context is always an issue,
> more so for a list that can grow significantly.
> Turns out that the driver only uses the duplicated
> list of multicast mac addresses to keep track of
> what addresses to delete from h/w befo
Don't disable the ipg clock in the regulator error path. The clock is
disable unconditionally two lines below the failed_regulator label.
Signed-off-by: Baruch Siach
---
drivers/net/ethernet/freescale/fec_main.c | 1 -
1 file changed, 1 deletion(-)
diff --git a/drivers/net/ethernet/freescale/fe
Hi Jose,
On 13/05/2019 11:07, Jose Abreu wrote:
> From: Simon Huelck
> Date: Sat, May 11, 2019 at 15:53:34
>
>> ethtool -S gave me some counts for mmc_rx_fifo_overflow, which i didnt
>> recognize before.
>
> Flow Control can prevent this to happen. Please check if your DT FIFO
> bindings are >
Hi Abel,
I'm testing kernel v5.2-rc1 on my i.MX8MQ system, SolidRun Hummingboard
Pulse. The fec driver happens to probe before the gpio driver that we need for
the PHY reset. So fec_reset_phy() returns -EPROBE_DEFER. This triggers the
splat below when clk_ahb is disabled somewhere below the 'fai
Hi David,
On Wednesday, May 05/22/19, 2019 at 01:53:31 +0530, David Miller wrote:
> From: Vishal Kulkarni
> Date: Tue, 21 May 2019 09:12:02 +0530
>
> > This reverts commit 2391b0030e241386d710df10e53e2cfc3c5d4fc1
> > SGE's BAR2 Doorbell/GTS Page Size is now interpreted correctly in the
> > firmw
On Wednesday, May 05/22/19, 2019 at 01:54:10 +0530, David Miller wrote:
> From: Vishal Kulkarni
> Date: Tue, 21 May 2019 09:10:37 +0530
>
> > This patch enables hash filter along with offload
> >
> > Signed-off-by: Vishal Kulkarni
>
> This commit message is too terse.
>
> What is going on her
On Tue 21 May 2019 at 16:23, Vlad Buslov wrote:
> On Tue 21 May 2019 at 15:46, Jamal Hadi Salim wrote:
>> On 2019-05-20 5:12 p.m., Jamal Hadi Salim wrote:
>>> On 2019-05-20 2:36 p.m., Edward Cree wrote:
On 20/05/2019 17:29, Jamal Hadi Salim wrote:
>>
>>> Ok, so the "get" does it. Will try t
Function tcf_action_dump() relies on tc_action->order field when starting
nested nla to send action data to userspace. This approach breaks in
several cases:
- When multiple filters point to same shared action, tc_action->order field
is overwritten each time it is attached to filter. This causes
It's easy to have a mismatch of "intended to be public" vs really
exposed API functions. While Makefile does check for this mismatch, if
it actually occurs it's not trivial to determine which functions are
accidentally exposed. This patch dumps out a diff showing what's not
supposed to be exposed f
Hash (exact-match) filters used for offloading flows share the
same active region resources on the chip with upper layer drivers,
like iw_cxgb4, chcr, etc. Currently, only either Hash filters
or ULDs can use the active region resources, but not both. Hence,
use the new firmware configuration parame
On 5/22/19 9:15 AM, Andrii Nakryiko wrote:
> It's easy to have a mismatch of "intended to be public" vs really
> exposed API functions. While Makefile does check for this mismatch, if
> it actually occurs it's not trivial to determine which functions are
> accidentally exposed. This patch dumps out
On Wed, May 22, 2019 at 10:33:29AM +0200, Benjamin Beckmeyer wrote:
> Hi all,
>
> I'm currently working on a custom board with the imx6ull processor and the
> 6390
> switching chip. This is our hardware setup.
>
> - -MAC
> | i.MX
Hi, is there any update on the plan for the networking track at
Plumbers (LPC) in Lisbon this year? Last year there was a CFP, but if
there was one this year I can't find any mention or details.
Should we be assuming we are submitting through the normal LPC channels
and timelines for presentation
On 05/21, Yonghong Song wrote:
> This patch tries to solve the following specific use case.
>
> Currently, bpf program can already collect stack traces
> through kernel function get_perf_callchain()
> when certain events happens (e.g., cache miss counter or
> cpu clock counter overflows). But such
On 5/22/19 9:38 AM, Stanislav Fomichev wrote:
> On 05/21, Yonghong Song wrote:
>> This patch tries to solve the following specific use case.
>>
>> Currently, bpf program can already collect stack traces
>> through kernel function get_perf_callchain()
>> when certain events happens (e.g., cache miss
On 05/22, Alexei Starovoitov wrote:
> On 5/22/19 9:15 AM, Andrii Nakryiko wrote:
> > It's easy to have a mismatch of "intended to be public" vs really
> > exposed API functions. While Makefile does check for this mismatch, if
> > it actually occurs it's not trivial to determine which functions are
On Mon, Apr 29, 2019 at 11:34:41AM +0300, Leon Romanovsky wrote:
> From: Mark Zhang
>
> Add an API to support set/clear per-port auto mode.
>
> Signed-off-by: Mark Zhang
> Reviewed-by: Majd Dibbiny
> Signed-off-by: Leon Romanovsky
> drivers/infiniband/core/Makefile | 2 +-
> drivers/infin
On Thu, 09 May 2019 21:57:49 -0700, John Fastabend wrote:
> It is possible (via shutdown()) for TCP socks to go through TCP_CLOSE
> state via tcp_disconnect() without calling into close callback. This
> would allow a kTLS enabled socket to exist outside of ESTABLISHED
> state which is not supported
On Mon, Apr 29, 2019 at 11:34:49AM +0300, Leon Romanovsky wrote:
> From: Mark Zhang
>
> Since a QP can only be bound to one counter, then if it is bound to a
> separate counter, for backward compatibility purpose, the statistic
> value must be:
> * stat of default counter
> + stat of all running
From: Jesse Brandeburg
Date: Wed, 22 May 2019 09:36:44 -0700
> Hi, is there any update on the plan for the networking track at
> Plumbers (LPC) in Lisbon this year? Last year there was a CFP, but if
> there was one this year I can't find any mention or details.
We are assembling the tech commit
On 05/22, Alexei Starovoitov wrote:
> On 5/22/19 9:38 AM, Stanislav Fomichev wrote:
> > On 05/21, Yonghong Song wrote:
> >> This patch tries to solve the following specific use case.
> >>
> >> Currently, bpf program can already collect stack traces
> >> through kernel function get_perf_callchain()
On Mon, Apr 29, 2019 at 11:34:42AM +0300, Leon Romanovsky wrote:
> void rdma_counter_init(struct ib_device *dev)
> {
> struct rdma_port_counter *port_counter;
> diff --git a/drivers/infiniband/core/device.c
> b/drivers/infiniband/core/device.c
> index 9204b4251fc8..dfaa57de871f 100644
> +
On Mon, Apr 29, 2019 at 11:34:42AM +0300, Leon Romanovsky wrote:
> +/**
> + * rdma_counter_unbind_qp - Unbind a qp from a counter
> + * @force:
> + * true - Decrease the counter ref-count anyway (e.g., qp destroy)
> + */
> +int rdma_counter_unbind_qp(struct ib_qp *qp, bool force)
> +{
> + st
On Mon, Apr 29, 2019 at 11:34:47AM +0300, Leon Romanovsky wrote:
> From: Mark Zhang
>
> This patch adds the ability to return all available counters
> together with their properties and hwstats.
>
> Signed-off-by: Mark Zhang
> Reviewed-by: Majd Dibbiny
> Signed-off-by: Leon Romanovsky
> driv
On Mon, Apr 29, 2019 at 11:34:47AM +0300, Leon Romanovsky wrote:
> From: Mark Zhang
>
> This patch adds the ability to return all available counters
> together with their properties and hwstats.
>
> Signed-off-by: Mark Zhang
> Reviewed-by: Majd Dibbiny
> Signed-off-by: Leon Romanovsky
> driv
On 2019-05-22 11:08 a.m., Vlad Buslov wrote:
On Tue 21 May 2019 at 16:23, Vlad Buslov wrote:
It seems that culprit in this case is tc_action->order field. It is used
as nla attrtype when dumping actions. Initially it is set according to
ordering of actions of filter that creates them. Howev
On Mon, Apr 29, 2019 at 11:34:49AM +0300, Leon Romanovsky wrote:
> diff --git a/drivers/infiniband/core/device.c
> b/drivers/infiniband/core/device.c
> index c56ffc61ab1e..8ae4906a60e7 100644
> +++ b/drivers/infiniband/core/device.c
> @@ -1255,7 +1255,11 @@ int ib_register_device(struct ib_device
On Mon, Apr 29, 2019 at 11:34:53AM +0300, Leon Romanovsky wrote:
> From: Mark Zhang
>
> This patch adds the ability to return the hwstats of per-port default
> counters (which can also be queried through sysfs nodes).
>
> Signed-off-by: Mark Zhang
> Signed-off-by: Leon Romanovsky
> drivers/in
On Mon, Apr 29, 2019 at 11:34:36AM +0300, Leon Romanovsky wrote:
> From: Leon Romanovsky
>
> Changelog:
> v1 -> v2:
> * Rebased to latest rdma-next
> v0 -> v1:
> * Changed wording of counter comment
> * Removed unneeded assignments
> * Added extra patch to present global counters
>
> * I
From: Russell King - ARM Linux admin
Date: Mon, 20 May 2019 16:21:34 +0100
> I realise that net-next probably isn't open yet, but I believe folk
> will find these patches "interesting" so I'm sending them to share
> them with people working on this code, rather than expecting them to
> be picked
On Wed, 22 May 2019 14:53:50 +0200, Björn Töpel wrote:
> Shout out to all XDP driver hackers to check that the second patch
> doesn't break anything (especially Jakub). I've only been able to test
> on the Intel NICs.
Please test XDP offload on netdevsim, that's why we have it! :)
At the minimum p
On Wed, May 22, 2019 at 9:21 AM Alexei Starovoitov wrote:
>
> On 5/22/19 9:15 AM, Andrii Nakryiko wrote:
> > It's easy to have a mismatch of "intended to be public" vs really
> > exposed API functions. While Makefile does check for this mismatch, if
> > it actually occurs it's not trivial to deter
On Wed, May 22, 2019 at 9:46 AM Stanislav Fomichev wrote:
>
> On 05/22, Alexei Starovoitov wrote:
> > On 5/22/19 9:15 AM, Andrii Nakryiko wrote:
> > > It's easy to have a mismatch of "intended to be public" vs really
> > > exposed API functions. While Makefile does check for this mismatch, if
> >
On Wed 22 May 2019 at 20:24, Jamal Hadi Salim wrote:
> On 2019-05-22 11:08 a.m., Vlad Buslov wrote:
>>
>> On Tue 21 May 2019 at 16:23, Vlad Buslov wrote:
>
>>
>> It seems that culprit in this case is tc_action->order field. It is used
>> as nla attrtype when dumping actions. Initially it is set
It's easy to have a mismatch of "intended to be public" vs really
exposed API functions. While Makefile does check for this mismatch, if
it actually occurs it's not trivial to determine which functions are
accidentally exposed. This patch dumps out a diff showing what's not
supposed to be exposed f
On 05/22, Andrii Nakryiko wrote:
> On Wed, May 22, 2019 at 9:46 AM Stanislav Fomichev wrote:
> >
> > On 05/22, Alexei Starovoitov wrote:
> > > On 5/22/19 9:15 AM, Andrii Nakryiko wrote:
> > > > It's easy to have a mismatch of "intended to be public" vs really
> > > > exposed API functions. While M
On Wed, May 22, 2019 at 2:25 AM Björn Töpel wrote:
>
> Add three tests to test_verifier/basic_instr that make sure that the
> high 32-bits of the destination register is cleared after an ALU32
> and/or/xor.
>
> Signed-off-by: Björn Töpel
I think the patch intends for bpf-next, right? The patch i
On 2019-05-22 1:49 p.m., Vlad Buslov wrote:
On Wed 22 May 2019 at 20:24, Jamal Hadi Salim wrote:
Ok, thanks for chasing this. A test case i had in mind is to
maybe have 3 actions. Add the drop in the middle for one
and at the begging for another and see if they are visible
with the patch.
If
On 2019-05-22 2:23 p.m., Jamal Hadi Salim wrote:
On 2019-05-22 1:49 p.m., Vlad Buslov wrote:
On Wed 22 May 2019 at 20:24, Jamal Hadi Salim wrote:
Ok, thanks for chasing this. A test case i had in mind is to
maybe have 3 actions. Add the drop in the middle for one
and at the begging for anot
On Wed, 22 May 2019 14:53:51 +0200, Björn Töpel wrote:
> From: Björn Töpel
>
> All XDP capable drivers need to implement the XDP_QUERY_PROG{,_HW}
> command of ndo_bpf. The query code is fairly generic. This commit
> refactors the query code up from the drivers to the netdev level.
>
> The struct
T1 PHYs make use of a single twisted pair, rather than the traditional
2 pair for 100BaseT or 4 pair for 1000BaseT. This patchset adds link
modes for 100BaseT1 and 1000BaseT1, and them makes use of 100BaseT1 in
the list of PHY features used by current T1 drivers.
---
The NXP PHY TJA1100 and TJA11
On Tue, May 21, 2019 at 10:40 PM Yonghong Song wrote:
>
> The test covered both nmi and tracepoint perf events.
> $ ./test_send_signal_user
> test_send_signal (tracepoint): OK
> test_send_signal (perf_event): OK
>
> Signed-off-by: Yonghong Song
> ---
> tools/testing/selftests/bpf/Makefile
The code was assuming the reset default of the delay control register
was to have delay disabled. This is what the datasheet shows as the
register's initial value. However, that's not actually true: the
default is controlled by the PHY's pin strapping.
If the interface mode is selected as RX or
The clock output is generally only used for testing and development and
not used to daisy-chain PHYs. It's just a source of RF noise afterward.
Add a mux value for "off". I've added it as another enumeration to the
output property. In the actual PHY, the mux and the output enable are
independen
Add a note to make it more clear how the driver behaves when "rgmii" vs
"rgmii-id", "rgmii-idrx", or "rgmii-idtx" interface modes are selected.
Cc: Rob Herring
Cc: Mark Rutland
Signed-off-by: Trent Piepho
---
Notes:
Changes from v1:
Clarify behavior may change to enforce no delay in
Generally, the output clock pin is only used for testing and only serves
as a source of RF noise after this. It could be used to daisy-chain
PHYs, but this is uncommon. Since the PHY can disable the output, make
doing so an option. I do this by adding another enumeration to the
allowed values of
Now that there is a link mode for 100BaseT1, use it in
phy_basic_t1_features so T1 PHY drivers will indicate this mode via
the Ethtool API.
Signed-off-by: Andrew Lunn
---
drivers/net/phy/phy_device.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/drivers/net/phy/phy_device.c
The variables used to store u32 DT properties were signed ints. This
doesn't work properly if the value of the property were to overflow.
Use unsigned variables so this doesn't happen.
Cc: Andrew Lunn
Cc: Florian Fainelli
Cc: Heiner Kallweit
Signed-off-by: Trent Piepho
---
drivers/net/phy/dp
Insure property is in valid range and fail when reading DT if it is not.
Also add error message for existing failure if required property is not
present.
Cc: Andrew Lunn
Cc: Florian Fainelli
Cc: Heiner Kallweit
Signed-off-by: Trent Piepho
---
Notes:
Changes from v1:
New patch in ser
This was being done in config the first time the phy was configured.
Should be in the probe method.
Cc: Andrew Lunn
Cc: Florian Fainelli
Cc: Heiner Kallweit
Signed-off-by: Trent Piepho
---
Notes:
Changes from v1:
New patch in series v2
drivers/net/phy/dp83867.c | 33 ++
The driver would only set the IO impedance value when RGMII internal
delays were enabled. There is no reason for this. Move the IO
impedance block out of the RGMII delay block.
Cc: Andrew Lunn
Cc: Florian Fainelli
Cc: Heiner Kallweit
Signed-off-by: Trent Piepho
---
Notes:
Changes from v
Add link modes for 100Mbps and 1Gbps over a single pair.
Signed-off-by: Andrew Lunn
---
drivers/net/phy/phy-core.c | 4 +++-
include/uapi/linux/ethtool.h | 2 ++
2 files changed, 5 insertions(+), 1 deletion(-)
diff --git a/drivers/net/phy/phy-core.c b/drivers/net/phy/phy-core.c
index 3daf0214
On Wed, May 22, 2019 at 06:43:22PM +, Trent Piepho wrote:
> Generally, the output clock pin is only used for testing and only serves
> as a source of RF noise after this. It could be used to daisy-chain
> PHYs, but this is uncommon. Since the PHY can disable the output, make
> doing so an opt
On Mon, 20 May 2019 11:56:52 +0200
Paolo Abeni wrote:
> The mirred act admits an optional control action, defaulting
> to TC_ACT_PIPE. The parsing code currently emits an error message
> if the control action is not provided on the command line, even
> if the command itself completes with no erro
On Wed, May 22, 2019 at 06:43:23PM +, Trent Piepho wrote:
> The code was assuming the reset default of the delay control register
> was to have delay disabled. This is what the datasheet shows as the
> register's initial value. However, that's not actually true: the
> default is controlled by
After previous patches, verifier will mark a insn if it really needs zero
extension on dst_reg.
It is then for back-ends to decide how to use such information to eliminate
unnecessary zero extension code-gen during JIT compilation.
One approach is verifier insert explicit zero extension for those
The encoding for this new variant is based on BPF_X format. "imm" field was
0 only, now it could be 1 which means doing zero extension unconditionally
.code = BPF_ALU | BPF_MOV | BPF_X
.dst_reg = DST
.src_reg = SRC
.imm = 1
We use this new form for doing zero extension for which verifier
v7:
- Drop the first patch in v6, the one adding 32-bit return value and
argument type. (Alexei)
- Rename bpf_jit_hardware_zext to bpf_jit_needs_zext. (Alexei)
- Use mov32 with imm == 1 to indicate it is zext. (Alexei)
- JIT back-ends peephole next insn to optimize out unnecessary zext
eBPF ISA specification requires high 32-bit cleared when low 32-bit
sub-register is written. This applies to destination register of ALU32 etc.
JIT back-ends must guarantee this semantic when doing code-gen. x86_64 and
AArch64 ISA has the same semantics, so the corresponding JIT back-end
doesn't ne
This patch randomizes high 32-bit of a definition when BPF_F_TEST_RND_HI32
is set.
Suggested-by: Alexei Starovoitov
Signed-off-by: Jiong Wang
---
kernel/bpf/verifier.c | 68 ++-
1 file changed, 57 insertions(+), 11 deletions(-)
diff --git a/kerne
Cc: Wang YanQing
Tested-by: Wang YanQing
Signed-off-by: Jiong Wang
---
arch/x86/net/bpf_jit_comp32.c | 83 +--
1 file changed, 56 insertions(+), 27 deletions(-)
diff --git a/arch/x86/net/bpf_jit_comp32.c b/arch/x86/net/bpf_jit_comp32.c
index b29e82f..133
This patch eliminate zero extension code-gen for instructions including
both alu and load/store. The only exception is for ctx load, because
offload target doesn't go through host ctx convert logic so we do
customized load and ignores zext flag set by verifier.
Cc: Jakub Kicinski
Reviewed-by: Jak
Patched insns do not go through generic verification, therefore doesn't has
zero extension information collected during insn walking.
We don't bother analyze them at the moment, for any sub-register def comes
from them, just conservatively mark it as needing zero extension.
Signed-off-by: Jiong W
libbpf doesn't allow passing "prog_flags" during bpf program load in a
couple of load related APIs, "bpf_load_program_xattr", "load_program" and
"bpf_prog_load_xattr".
It makes sense to allow passing "prog_flags" which is useful for
customizing program loading.
Reviewed-by: Jakub Kicinski
Signed
Cc: David S. Miller
Signed-off-by: Jiong Wang
---
arch/sparc/net/bpf_jit_comp_64.c | 29 +++--
1 file changed, 27 insertions(+), 2 deletions(-)
diff --git a/arch/sparc/net/bpf_jit_comp_64.c b/arch/sparc/net/bpf_jit_comp_64.c
index 65428e7..3364e2a 100644
--- a/arch/sparc
Cc: Martin Schwidefsky
Cc: Heiko Carstens
Signed-off-by: Jiong Wang
---
arch/s390/net/bpf_jit_comp.c | 41 ++---
1 file changed, 34 insertions(+), 7 deletions(-)
diff --git a/arch/s390/net/bpf_jit_comp.c b/arch/s390/net/bpf_jit_comp.c
index 5e7c630..e636728
x86_64 and AArch64 perhaps are two arches that running bpf testsuite
frequently, however the zero extension insertion pass is not enabled for
them because of their hardware support.
It is critical to guarantee the pass correction as it is supposed to be
enabled at default for a couple of other arc
- bpf_fill_ld_abs_vlan_push_pop:
Prevent zext happens inside PUSH_CNT loop. This could happen because
of BPF_LD_ABS (32-bit def) + BPF_JMP (64-bit use), or BPF_LD_ABS +
EXIT (64-bit use of R0). So, change BPF_JMP to BPF_JMP32 and redefine
R0 at exit path to cut off the data-flow f
The previous libbpf patch allows user to specify "prog_flags" to bpf
program load APIs. To enable high 32-bit randomization for a test, we need
to set BPF_F_TEST_RND_HI32 in "prog_flags".
To enable such randomization for all tests, we need to make sure all places
are passing BPF_F_TEST_RND_HI32. C
Cc: Naveen N. Rao
Cc: Sandipan Das
Signed-off-by: Jiong Wang
---
arch/powerpc/net/bpf_jit_comp64.c | 36 +---
1 file changed, 33 insertions(+), 3 deletions(-)
diff --git a/arch/powerpc/net/bpf_jit_comp64.c
b/arch/powerpc/net/bpf_jit_comp64.c
index 21a1dcd..0ebd
Cc: Shubham Bansal
Signed-off-by: Jiong Wang
---
arch/arm/net/bpf_jit_32.c | 42 +++---
1 file changed, 31 insertions(+), 11 deletions(-)
diff --git a/arch/arm/net/bpf_jit_32.c b/arch/arm/net/bpf_jit_32.c
index c8bfbbf..97a6b4b 100644
--- a/arch/arm/net/bpf_j
Cc: Björn Töpel
Acked-by: Björn Töpel
Tested-by: Björn Töpel
Signed-off-by: Jiong Wang
---
arch/riscv/net/bpf_jit_comp.c | 43 ++-
1 file changed, 30 insertions(+), 13 deletions(-)
diff --git a/arch/riscv/net/bpf_jit_comp.c b/arch/riscv/net/bpf_jit_comp
On Wed, 22 May 2019 11:25:02 +, Boris Pismenny wrote:
> > +Performance metrics
> > +===
> > +
> > +TLS offload can be characterized by the following basic metrics:
> > +
> > + * max connection count
> > + * connection installation rate
> > + * connection installation latency
> >
On Wed, May 22, 2019 at 06:43:24PM +, Trent Piepho wrote:
> The variables used to store u32 DT properties were signed ints. This
> doesn't work properly if the value of the property were to overflow.
> Use unsigned variables so this doesn't happen.
>
> Cc: Andrew Lunn
> Cc: Florian Fainelli
On Wed, May 22, 2019 at 06:43:25PM +, Trent Piepho wrote:
> The driver would only set the IO impedance value when RGMII internal
> delays were enabled. There is no reason for this. Move the IO
> impedance block out of the RGMII delay block.
>
> Cc: Andrew Lunn
> Cc: Florian Fainelli
> Cc:
From: David Ahern
This set exports ipv4 and ipv6 fib functions for use by the nexthop
code. It also adds new ones to send route notifications if a nexthop
configuration changes.
v2
- repost of patches dropped at the end of the last dev window
added patch 8 which exports nh_update_mtu since it
From: David Ahern
Add hook to ipv6 stub to bump the sernum up to the root node for a
route. This is needed by the nexthop code when a nexthop config changes.
Signed-off-by: David Ahern
---
include/net/ip6_fib.h| 1 +
include/net/ipv6_stubs.h | 1 +
net/ipv6/af_inet6.c | 1 +
net/ipv6/
From: David Ahern
Add fib_info_notify_update to walk the fib and send RTM_NEWROUTE
notifications with NLM_F_REPLACE set for entries linked to a fib_info
that have nh_updated flag set. This helper will be used by the nexthop
code to notify userspace of routes that are impacted when a nexthop
confi
1 - 100 of 271 matches
Mail list logo
