Russell King says:
====================
net: phylink: improve PHY validation
One of the issues which has concerned me about the rate matching
implenentation that we have is that phy_get_rate_matching() returns
whether rate matching will be used for a particular interface, and we
enquire only for one interface.
Aquantia PHYs can be programmed with the rate matching and interface
mode settings on a per-media speed basis using the per-speed vendor 1
global configuration registers.
Thus, it is possible for the PHY to be configured to use rate matching
for 10G, 5G, 2.5G with 10GBASE-R, and then SGMII for the remaining
speeds. Therefore, it clearly doesn't make sense to enquire about rate
matching for just one interface mode.
Also, PHYs that change their interfaces are handled sub-optimally, in
that we validate all the interface modes that the host supports, rather
than the interface modes that the PHY will use.
This patch series changes the way we validate PHYs, but in order to do
so, we need to know exactly which interface modes will be used by the
PHY. So that phylib can convey this information, we add
"possible_interfaces" to struct phy_device.
possible_interfaces is to be filled in by a phylib driver once the PHY
is configured (in other words in the PHYs .config_init method) with the
interface modes that it will switch between. This then allows users of
phylib to know which interface modes will be used by the PHY.
This allows us to solve both these issues: where possible_interfaces is
provided, we can validate which ethtool link modes can be supported by
looking at which interface modes that both the PHY and host support,
and request rate matching information for each mode.
This should improve the accuracy of the validation.
Sending this out again without RFC as Jie Luo will need it for the
QCA8084 changes. No changes except to add the attributations already
received. Thanks!
====================
Link: https://lore.kernel.org/r/ZWCWn+uNkVLPaQhn@shell.armlinux.org.uk
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Some PHYs such as Aquantia, Broadcom 84881, and Marvell 88X33x0 can
switch between a set of interface types depending on the negotiated
media speed, or can use rate adaption for some or all of these
interface types.
We currently assume that these are Clause 45 PHYs that are configured
not to use a specific set of interface modes, which has worked so far,
but is just a work-around. In this workaround, we validate using all
interfaces that the MAC supports, which can lead to extra modes being
advertised that can not be supported.
To properly address this, switch to using the newly introduced PHY
possible_interfaces bitmap which indicates which interface modes will
be used by the PHY as configured. We calculate the union of the PHY's
possible interfaces and MACs supported interfaces, checking that is
non-empty. If the PHY is on a SFP, we further reduce the set by those
which can be used on a SFP module, again checking that is non-empty.
Finally, we validate the subset of interfaces, taking account of
whether rate matching will be used for each individual interface mode.
This becomes independent of whether the PHY is clause 22 or clause 45.
It is encouraged that all PHYs that switch interface modes or use
rate matching should populate phydev->possible_interfaces.
Tested-by: Luo Jie <quic_luoj@quicinc.com>
Signed-off-by: Russell King (Oracle) <rmk+kernel@armlinux.org.uk>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Link: https://lore.kernel.org/r/E1r6VIV-00DDMF-Pi@rmk-PC.armlinux.org.uk
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
When bringing up a PHY, we need to work out which ethtool link modes it
should support and advertise. Clause 22 PHYs operate in a single
interface mode, which can be easily dealt with. However, clause 45 PHYs
tend to switch interface mode depending on the media. We need more
flexible validation at this point, so this patch splits out that code
in preparation to changing it.
Tested-by: Luo Jie <quic_luoj@quicinc.com>
Signed-off-by: Russell King (Oracle) <rmk+kernel@armlinux.org.uk>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Link: https://lore.kernel.org/r/E1r6VIQ-00DDM9-LK@rmk-PC.armlinux.org.uk
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Add a possible_interfaces member to struct phy_device to indicate which
interfaces a clause 45 PHY may switch between depending on the media.
This must be populated by the PHY driver by the time the .config_init()
method completes according to the PHYs host-side configuration.
For example, the Marvell 88x3310 PHY can switch between 10GBASE-R,
5GBASE-R, 2500BASE-X, and SGMII on the host side depending on the media
side speed, so all these interface modes are set in the
possible_interfaces member.
This allows phylib users (such as phylink) to know in advance which
interface modes to expect, which allows them to appropriately restrict
the advertised link modes according to the capabilities of other parts
of the link.
Tested-by: Luo Jie <quic_luoj@quicinc.com>
Signed-off-by: Russell King (Oracle) <rmk+kernel@armlinux.org.uk>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Link: https://lore.kernel.org/r/E1r6VHk-00DDLN-I7@rmk-PC.armlinux.org.uk
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Pages belonging to a page_pool (PP) instance must be freed through the
PP APIs in-order to correctly release any DMA mappings and release
refcnt on the DMA device when freeing PP instance. When PP release a
page (page_pool_release_page) the page->pp_magic value is cleared.
This patch detect a leaked PP page in free_page_is_bad() via
unexpected state of page->pp_magic value being PP_SIGNATURE.
We choose to report and treat it as a bad page. It would be possible
to release the page via returning it to the PP instance as the
page->pp pointer is likely still valid.
Notice this code is only activated when either compiled with
CONFIG_DEBUG_VM or boot cmdline debug_pagealloc=on, and
CONFIG_PAGE_POOL.
Reduced example output of leak with PP_SIGNATURE = dead000000000040:
BUG: Bad page state in process swapper/4 pfn:141fa6
page:000000006dbf8062 refcount:0 mapcount:0 mapping:0000000000000000 index:0x141fa6000 pfn:0x141fa6
flags: 0x2fffff80000000(node=0|zone=2|lastcpupid=0x1fffff)
page_type: 0xffffffff()
raw: 002fffff80000000 dead000000000040 ffff88814888a000 0000000000000000
raw: 0000000141fa6000 0000000000000001 00000000ffffffff 0000000000000000
page dumped because: page_pool leak
[...]
Call Trace:
<IRQ>
dump_stack_lvl+0x32/0x50
bad_page+0x70/0xf0
free_unref_page_prepare+0x263/0x430
free_unref_page+0x34/0x130
mlx5e_free_rx_mpwqe+0x190/0x1c0 [mlx5_core]
mlx5e_post_rx_mpwqes+0x1ac/0x280 [mlx5_core]
mlx5e_napi_poll+0x12b/0x710 [mlx5_core]
? skb_free_head+0x4f/0x90
__napi_poll+0x2b/0x1c0
net_rx_action+0x27b/0x360
The advantage is the Call Trace directly points to the function
leaking the PP page, which in this case is an on purpose bug
introduced into the mlx5 driver to test this code change.
Currently PP will periodically in page_pool_release_retry()
printk warning "stalled pool shutdown" which cannot be directly
corrolated to leaking and might as well be a false positive
due to SKBs being stuck on a socket for an extended period.
After this patch we should be able to remove this printk.
Signed-off-by: Jesper Dangaard Brouer <hawk@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
RollBall SFP modules allow the access to PHY registers only after a
certain time has passed. Until then, the registers read 0xffff.
Currently we have quirks for modules where we need to wait either 25
seconds or 4 seconds, but recently I got hands on another module where
the wait is even shorter.
Instead of hardcoding different wait times, lets rework the code:
- increase the PHY retry count to 25
- when RollBall module is detected, increase the PHY retry time from
50ms to 1s
Signed-off-by: Marek Behún <kabel@kernel.org>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: David S. Miller <davem@davemloft.net>
This check can never be true for a firmware file with a correct format.
Existing checks in rtl_fw_data_ok() are sufficient, no problems with
invalid firmware files are known.
Signed-off-by: Heiner Kallweit <hkallweit1@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
The enum name used for id-to-str table does not handle
the enum-name override in the spec correctly.
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
If a new family is ever added with a dash in the name
the C codegen will break. Make sure we use the "safe"
form of the name consistently.
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
The commit dcd2cf5f2f ("net/smc: add autocorking support") adds an
atomic variable tx_pushing in smc_connection to make sure only one can
send to let it cork more and save CDC slot. since smc_tx_pending can be
called in the soft IRQ without checking sock_owned_by_user() at that
time, which would cause a race condition because bh_lock_sock() did
not honor sock_lock()
After commit 6b88af839d ("net/smc: don't send in the BH context if
sock_owned_by_user"), the transmission is deferred to when sock_lock()
is held by the user. Therefore, we no longer need tx_pending to hold
message.
So remove atomic variable tx_pushing and its operation, and
smc_tx_sndbuf_nonempty becomes a wrapper of __smc_tx_sndbuf_nonempty,
so rename __smc_tx_sndbuf_nonempty back to smc_tx_sndbuf_nonempty
Suggested-by: Alexandra Winter <wintera@linux.ibm.com>
Co-developed-by: Dust Li <dust.li@linux.alibaba.com>
Signed-off-by: Dust Li <dust.li@linux.alibaba.com>
Signed-off-by: Li RongQing <lirongqing@baidu.com>
Reviewed-by: Alexandra Winter <wintera@linux.ibm.com>
diff v4: remove atomic variable tx_pushing
diff v3: improvements in the commit body and comments
diff v2: fix a typo in commit body and add net-next subject-prefix
net/smc/smc.h | 1 -
net/smc/smc_tx.c | 30 +-----------------------------
2 files changed, 1 insertion(+), 30 deletions(-)
Signed-off-by: David S. Miller <davem@davemloft.net>
32bit builds generate the following warning when we use a u32-max
in range validation:
warning: decimal constant 4294967295 is between LONG_MAX and ULONG_MAX. For C99 that means long long, C90 compilers are very likely to produce unsigned long (and a warning) here
The range values are u64, slap ULL/LL on all of them just
to avoid such noise.
There's currently no code using full range validation, but
it will matter in the upcoming page-pool introspection.
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
Alex Elder says:
====================
net: ipa: add IPA v5.5 support
This series adds IPA support for the Qualcomm SM8550 SoC, which uses
IPA v5.5.
The first patch adds a new compatible string for the SM8550. The
second cleans up "ipa_reg.h" a bit for consistency. The third patch
adds definitions and some minor code changes related to IPA v5.5.
The fourth defines IPA register offsets and fields used for IPA
v5.0; most--but not all--register definitions are the same as used
in IPA v5.0. The final patch adds configuration data used for IPA
v5.5 (here again this mostly duplicates IPA v5.0 definitions).
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
Add the configuration data required for IPA v5.5, which is used in
the Qualcomm SM8550 SoC. With that, the driver supports IPA v5.5.
Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
GSI register definitions for IPA v5.5 are the same as those used for
IPA v5.0.
Update ipa_reg_id_valid() to reflect that IPA v5.0+ supports source
and destination resource groups 4 through 7.
Add the definitions of IPA register offsets and fields for IPA v5.5.
Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
For IPA v5.5+, the QTIME_TIMESTAMP_CFG register no longer defines
two fields in the DPL timestamp. Make the code referencing those
fields in ipa_qtime_config() conditional based on IPA version.
IPA v5.0+ supports the IPA_MEM_AP_V4_FILTER and IPA_MEM_AP_V6_FILTER
memory regions. Update ipa_mem_id_valid() to reflect that.
IPA v5.5 no longer supports a few register fields, adds some others,
and removes support for a few IPA interrupt types. Update
"ipa_reg.h" to include information about IPA v5.5.
Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
Some definitions in "ipa_reg.h" are only valid for certain versions
of IPA. In such cases a comment indicates a version or range of
versions where the definition is (or is not) valid. Almost all such
cases look like "IPA vX.Y", but a few don't include the "IPA" tag.
Update these so they all consistently include "IPA". And replace
a few lines that talk about "the next bit" in the definition of the
ipa_irq_id enumerated type with a more concise comment using the
"IPA vX.Y" convention.
Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
Shinas Rasheed says:
====================
Get max rx packet length and solve
Patchsets which resolve observed style issues in control net
source files, and also implements get mtu control net api
to fetch max mtu value from firmware
Changes:
V2:
- Introduced a patch to resolve style issues as mentioned in V1
- Removed OCTEP_MAX_MTU macro, as it is redundant.
V1: https://lore.kernel.org/all/20231121191224.2489474-1-srasheed@marvell.com/
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
Max receive packet length can vary across SoCs, so
this needs to be queried from respective firmware and
filled by driver. A control net get mtu api should be
implemented to do the same.
Signed-off-by: Shinas Rasheed <srasheed@marvell.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Solve observed style issues for kernel documentation
and structure declarations in control net API definitions.
Signed-off-by: Shinas Rasheed <srasheed@marvell.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Guangguan Wang says:
===================
add two sysctl for SMC-R v2.1
This patch set add two sysctl for SMC-R v2.1:
net.smc.smcr_max_links_per_lgr is used to control the max links
per lgr.
net.smc.smcr_max_conns_per_lgr is used to control the max connections
per lgr.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
Add a new sysctl: net.smc.smcr_max_conns_per_lgr, which is
used to control the preferred max connections per lgr for
SMC-R v2.1. The default value of this sysctl is 255, and
the acceptable value ranges from 16 to 255.
Signed-off-by: Guangguan Wang <guangguan.wang@linux.alibaba.com>
Reviewed-by: Dust Li <dust.li@linux.alibaba.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Add a new sysctl: net.smc.smcr_max_links_per_lgr, which is
used to control the preferred max links per lgr for SMC-R
v2.1. The default value of this sysctl is 2, and the acceptable
value ranges from 1 to 2.
Signed-off-by: Guangguan Wang <guangguan.wang@linux.alibaba.com>
Reviewed-by: Dust Li <dust.li@linux.alibaba.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Adds tc offload support for matching on ICMP type and code.
Example usage:
To enable adding tc ingress rules
tc qdisc add dev eth0 ingress
TC rule drop the ICMP echo reply:
tc filter add dev eth0 protocol ip parent ffff: \
flower ip_proto icmp type 8 code 0 skip_sw action drop
TC rule to drop ICMPv6 echo reply:
tc filter add dev eth0 protocol ipv6 parent ffff: flower \
indev eth0 ip_proto icmpv6 type 128 code 0 action drop
Signed-off-by: Geetha sowjanya <gakula@marvell.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Introduced by commit 6e2d85ec05 ("net: phy: Stop with excessive soft
reset").
soft_reset call for phy_init_hw had multiple revision across the years
and the implementation goes back to 2014. Originally was a simple call
to write the generic PHY reset BIT, it was then moved to a dedicated
function. It was then added the option for PHY driver to define their
own special way to reset the PHY. Till this change, checking for ret was
correct as it was always filled by either the generic reset or the
custom implementation. This changed tho with commit 6e2d85ec05 ("net:
phy: Stop with excessive soft reset"), as the generic reset call to PHY
was dropped but the ret check was never made entirely optional and
dependent whether soft_reset was defined for the PHY driver or not.
Luckly nothing was ever added before the soft_reset call so the ret
check (in the case where a PHY didn't had soft_reset defined) although
wrong, never caused problems as ret was init 0 at the start of
phy_init_hw.
To prevent any kind of problem and to make the function cleaner and more
robust, correctly move the ret check if the soft_reset section making it
optional and needed only with the function defined.
Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
Reviewed-by: Larysa Zaremba <larysa.zaremba@intel.com>
Reviewed-by: Simon Horman <horms@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
This is a simple script that parses the Netlink YAML spec files
(Documentation/netlink/specs/), and generates RST files to be rendered
in the Network -> Netlink Specification documentation page.
Create a python script that is invoked during 'make htmldocs', reads the
YAML specs input file and generate the correspondent RST file.
Create a new Documentation/networking/netlink_spec index page, and
reference each Netlink RST file that was processed above in this main
index.rst file.
In case of any exception during the parsing, dump the error and skip
the file.
Do not regenerate the RST files if the input files (YAML) were not
changed in-between invocations.
Suggested-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: Breno Leitao <leitao@debian.org>
----
Changelog:
V3:
* Do not regenerate the RST files if the input files were not
changed. In order to do it, a few things changed:
- Rely on Makefile more to find what changed, and trigger
individual file processing
- The script parses file by file now (instead of batches)
- Create a new option to generate the index file
V2:
* Moved the logic from a sphinx extension to a external script
* Adjust some formatting as suggested by Donald Hunter and Jakub
* Auto generating all the rsts instead of having stubs
* Handling error gracefully
Reviewed-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
Pull networking fixes from Jakub Kicinski:
"Including fixes from bpf.
Current release - regressions:
- Revert "net: r8169: Disable multicast filter for RTL8168H and
RTL8107E"
- kselftest: rtnetlink: fix ip route command typo
Current release - new code bugs:
- s390/ism: make sure ism driver implies smc protocol in kconfig
- two build fixes for tools/net
Previous releases - regressions:
- rxrpc: couple of ACK/PING/RTT handling fixes
Previous releases - always broken:
- bpf: verify bpf_loop() callbacks as if they are called unknown
number of times
- improve stability of auto-bonding with Hyper-V
- account BPF-neigh-redirected traffic in interface statistics
Misc:
- net: fill in some more MODULE_DESCRIPTION()s"
* tag 'net-6.7-rc3' of git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net: (58 commits)
tools: ynl: fix duplicate op name in devlink
tools: ynl: fix header path for nfsd
net: ipa: fix one GSI register field width
tls: fix NULL deref on tls_sw_splice_eof() with empty record
net: axienet: Fix check for partial TX checksum
vsock/test: fix SEQPACKET message bounds test
i40e: Fix adding unsupported cloud filters
ice: restore timestamp configuration after device reset
ice: unify logic for programming PFINT_TSYN_MSK
ice: remove ptp_tx ring parameter flag
amd-xgbe: propagate the correct speed and duplex status
amd-xgbe: handle the corner-case during tx completion
amd-xgbe: handle corner-case during sfp hotplug
net: veth: fix ethtool stats reporting
octeontx2-pf: Fix ntuple rule creation to direct packet to VF with higher Rx queue than its PF
net: usb: qmi_wwan: claim interface 4 for ZTE MF290
Revert "net: r8169: Disable multicast filter for RTL8168H and RTL8107E"
net/smc: avoid data corruption caused by decline
nfc: virtual_ncidev: Add variable to check if ndev is running
dpll: Fix potential msg memleak when genlmsg_put_reply failed
...
We don't support CRUD-inspired message types in YNL too well.
One aspect that currently trips us up is the fact that single
message ID can be used in multiple commands (as the response).
This leads to duplicate entries in the id-to-string tables:
devlink-user.c:19:34: warning: initialized field overwritten [-Woverride-init]
19 | [DEVLINK_CMD_PORT_NEW] = "port-new",
| ^~~~~~~~~~
devlink-user.c:19:34: note: (near initialization for ‘devlink_op_strmap[7]’)
Fixes tag points at where the code was generated, the "real" problem
is that the code generator does not support CRUD.
Fixes: f2f9dd164d ("netlink: specs: devlink: add the remaining command to generate complete split_ops")
Link: https://lore.kernel.org/r/20231123030558.1611831-1-kuba@kernel.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
syzkaller discovered that if tls_sw_splice_eof() is executed as part of
sendfile() when the plaintext/ciphertext sk_msg are empty, the send path
gets confused because the empty ciphertext buffer does not have enough
space for the encryption overhead. This causes tls_push_record() to go on
the `split = true` path (which is only supposed to be used when interacting
with an attached BPF program), and then get further confused and hit the
tls_merge_open_record() path, which then assumes that there must be at
least one populated buffer element, leading to a NULL deref.
It is possible to have empty plaintext/ciphertext buffers if we previously
bailed from tls_sw_sendmsg_locked() via the tls_trim_both_msgs() path.
tls_sw_push_pending_record() already handles this case correctly; let's do
the same check in tls_sw_splice_eof().
Fixes: df720d288d ("tls/sw: Use splice_eof() to flush")
Cc: stable@vger.kernel.org
Reported-by: syzbot+40d43509a099ea756317@syzkaller.appspotmail.com
Signed-off-by: Jann Horn <jannh@google.com>
Link: https://lore.kernel.org/r/20231122214447.675768-1-jannh@google.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Tune message length calculation to make this test work on machines
where 'getpagesize()' returns >32KB. Now maximum message length is not
hardcoded (on machines above it was smaller than 'getpagesize()' return
value, thus we get negative value and test fails), but calculated at
runtime and always bigger than 'getpagesize()' result. Reproduced on
aarch64 with 64KB page size.
Fixes: 5c338112e4 ("test/vsock: rework message bounds test")
Signed-off-by: Arseniy Krasnov <avkrasnov@salutedevices.com>
Reported-by: Bogdan Marcynkov <bmarcynk@redhat.com>
Reviewed-by: Stefano Garzarella <sgarzare@redhat.com>
Link: https://lore.kernel.org/r/20231121211642.163474-1-avkrasnov@salutedevices.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
If a VF tries to add unsupported cloud filter through virtchnl
then i40e_add_del_cloud_filter(_big_buf) returns -ENOTSUPP but
this error code is stored in 'ret' instead of 'aq_ret' that
is used as error code sent back to VF. In this scenario where
one of the mentioned functions fails the value of 'aq_ret'
is zero so the VF will incorrectly receive a 'success'.
Use 'aq_ret' to store return value and remove 'ret' local
variable. Additionally fix the issue when filter allocation
fails, in this case no notification is sent back to the VF.
Fixes: e284fc2804 ("i40e: Add and delete cloud filter")
Reviewed-by: Simon Horman <horms@kernel.org>
Signed-off-by: Ivan Vecera <ivecera@redhat.com>
Tested-by: Rafal Romanowski <rafal.romanowski@intel.com>
Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>
Link: https://lore.kernel.org/r/20231121211338.3348677-1-anthony.l.nguyen@intel.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
