David Marchand [Tue, 16 Jun 2020 09:47:00 +0000 (11:47 +0200)]
net/mvpp2: fix non-EAL thread support
Caught by code inspection, for a non-EAL thread identified with
rte_lcore_id() == LCORE_ID_ANY, the code currently arbitrarily uses
lcore 0 while there is no guarantee this lcore is used.
After enabling promiscuous mode all packets whose destination MAC
address is a multicast address were being dropped. This fix configures
H/W to receive all traffic in promiscuous mode. Promiscuous mode also
overrides allmulticast mode on/off status.
Fixes: 40e9f6fc1558 ("net/qede: enable VF-VF traffic with unmatched dest address") Cc: stable@dpdk.org Signed-off-by: Devendra Singh Rawat <dsinghrawat@marvell.com> Signed-off-by: Igor Russkikh <irusskikh@marvell.com> Signed-off-by: Rasesh Mody <rmody@marvell.com>
net/mlx5: fix host physical function representor naming
The new kernel adds the names like "pf0" for Host PCI physical
function representor on Bluefield SmartNIC hosts. This patch
provides correct HPF representor recognition over the kernel
versions 5.7 and laters.
The following port naming formats are supported:
- missing physical port name (no sysfs/netlink key) at all,
master is assumed
- decimal digits (for example "12"), representor is
assumed, the value is the index of attached VF
- "p" followed by decimal digits, for example "p2", master
is assumed
- "pf" followed by PF index, for example "pf0", Host PF
representor is assumed on SmartNIC systems.
- "pf" followed by PF index concatenated with "vf" followed by
VF index, for example "pf0vf1", representor is assumed.
If index of VF is "-1" it is a special case of Host PF
representor, this representor must be indexed in devargs
as 65535, for example representor=[0-3,65535] will
allow representors for VF0, VF1, VF2, VF3 and for host PF.
Fixes: 79aa430721b1 ("common/mlx5: split common file under Linux directory") Cc: stable@dpdk.org Signed-off-by: Viacheslav Ovsiienko <viacheslavo@mellanox.com> Acked-by: Matan Azrad <matan@mellanox.com>
Junyu Jiang [Wed, 24 Jun 2020 02:09:39 +0000 (02:09 +0000)]
net/ice: initialize and update RSS based on user config
Initialize and update RSS configure based on user request
(rte_eth_rss_conf) from dev_configure and .rss_hash_update ops.
All previous default configure has been removed.
Ophir Munk [Fri, 19 Jun 2020 07:30:08 +0000 (07:30 +0000)]
common/mlx5: move some getter functions from net driver
Getter functions such as: 'mlx5_os_get_ctx_device_name',
'mlx5_os_get_ctx_device_path', 'mlx5_os_get_dev_device_name',
'mlx5_os_get_umem_id' are implemented under net directory. To enable
additional devices (e.g. regex, vdpa) to access these getter functions
they are moved under common directory.
As part of this commit string sizes DEV_SYSFS_NAME_MAX and
DEV_SYSFS_PATH_MAX are increased by 1 to make sure that the destination
string size in strncpy() function is bigger than the source string size.
This update will avoid GCC version 8 error -Werror=stringop-truncation.
Suanming Mou [Thu, 18 Jun 2020 08:12:50 +0000 (16:12 +0800)]
net/mlx5: optimize free counter lookup
Currently, when allocate a new counter, it needs loop the whole
container pool list to get a free counter.
In the case with millions of counters allocated, and all the pools
are empty, allocate the new counter will still need to loop the
whole container pool list first, then allocate a new pool to get a
free counter. It wastes the cycles during the pool list traversal.
Add a global free counter list in the container helps to get the free
counters more efficiently.
Suanming Mou [Thu, 18 Jun 2020 07:24:44 +0000 (15:24 +0800)]
net/mlx5: optimize single counter pool search
For single counter, when allocate a new counter, it needs to find the pool
it belongs in order to do the query together.
Once there are millions of counters allocated, the pool array in the
counter container will become very large. In this case, the pool search
from the pool array will become extremely slow.
Save the minimum and maximum counter ID to have a quick check of current
counter ID range. And start searching the pool from the last pool in the
container will mostly get the needed pool since counter ID increases
sequentially.
Suanming Mou [Thu, 18 Jun 2020 07:24:43 +0000 (15:24 +0800)]
net/mlx5: manage shared counters in three-level table
Currently, to check if any shared counter with same ID existing, it will
have to loop the counter pools to search for the counter. Even add the
counter to the list will also not so helpful while there are thousands
of shared counters in the list.
Change Three-Level table to look up the counter index saved in the
relevant table entry will be more efficient.
This patch introduces the Three-level table to save the ID relevant
counter index in the table. Then the next while the same ID comes, just
check the table entry of this ID will get the counter index directly.
No search will be needed.
Suanming Mou [Thu, 18 Jun 2020 07:24:42 +0000 (15:24 +0800)]
net/mlx5: add three-level table utility
For the case which data is linked with sequence increased index, the
array table will be more efficient than hash table once need to search
one data entry in large numbers of entries. Since the traditional hash
tables has fixed table size, when huge numbers of data saved to the hash
table, it also comes lots of hash conflict.
But simple array table also has fixed size, allocates all the needed
memory at once will waste lots of memory. For the case don't know the
exactly number of entries will be impossible to allocate the array.
Then the multiple level table helps to balance the two disadvantages.
Allocate a global high level table with sub table entries at first,
the global table contains the sub table entries, and the sub table will
be allocated only once the corresponding index entry need to be saved.
e.g. for up to 32-bits index, three level table with 10-10-12 splitting,
with sequence increased index, the memory grows with every 4K entries.
The currently implementation introduces 10-10-12 32-bits splitting
Three-Level table to help the cases which have millions of entries to
save. The index entries can be addressed directly by the index, no
search will be needed.
Matan Azrad [Thu, 18 Jun 2020 18:59:42 +0000 (18:59 +0000)]
common/mlx5: support DevX virtq stats operations
Add DevX API to create and query virtio queue statistics
from the HW. The next counters are supported by the HW per
virtio queue:
received_desc.
completed_desc.
error_cqes.
bad_desc_errors.
exceed_max_chain.
invalid_buffer.
Matan Azrad [Thu, 18 Jun 2020 18:59:41 +0000 (18:59 +0000)]
vhost: introduce operation to get vDPA queue stats
The vDPA device offloads all the datapath of the vhost
device to the HW device.
In order to expose to the user traffic information this
patch introduces new 3 APIs to get traffic statistics, the
device statistics name and to reset the statistics per
virtio queue.
The statistics are taken directly from the vDPA driver
managing the HW device and can be different for each vendor
driver.
Signed-off-by: Matan Azrad <matan@mellanox.com> Reviewed-by: Maxime Coquelin <maxime.coquelin@redhat.com>
Maxime Coquelin [Thu, 28 May 2020 09:03:47 +0000 (11:03 +0200)]
vhost: enable reply-ack systematically
As announced during v20.05 release cycle, this
patch makes reply-ack protocol feature to be enabled
unconditionally.
This protocol feature makes the communication between the
master and the slave more robust, avoiding for example
possible undefined behaviour with VHOST_USER_SET_MEM_TABLE.
Also, reply-ack support will be required for upcoming
VHOST_USER_SET_STATUS request.
Note that this protocol feature was disabled by default
because Qemu version 2.7.0 to 2.9.0 had a bug causing a
deadlock when reply-ack was negotiated and multiqueue
enabled. These Qemu version are now very old and no more
maintained, so we can reasonably consider we no more
support them.
Signed-off-by: Maxime Coquelin <maxime.coquelin@redhat.com> Reviewed-by: Chenbo Xia <chenbo.xia@intel.com>
Qi Zhang [Thu, 11 Jun 2020 08:43:30 +0000 (16:43 +0800)]
net/ice/base: replace RSS profile locks
Replacing flow profile locks with RSS profile locks in the function to
remove all RSS rules for a given VSI. This is to align the locks used
for RSS rule addition to VSI and removal during VSI teardown to avoid
a race condition owing to several iterations of the above operations.
In function to get RSS rules for given VSI and protocol header replacing
the pointer reference of the RSS entry with a copy of hash value to
ensure thread safety.
Signed-off-by: Vignesh Sridhar <vignesh.sridhar@intel.com> Signed-off-by: Paul M Stillwell Jr <paul.m.stillwell.jr@intel.com> Signed-off-by: Qi Zhang <qi.z.zhang@intel.com> Acked-by: Qiming Yang <qiming.yang@intel.com>
Qi Zhang [Thu, 11 Jun 2020 08:43:29 +0000 (16:43 +0800)]
net/ice/base: fix VSI ID mask to 10 bits
set_rss_lut failed due to incorrect vsi_id mask. vsi_id is 10 bit
but mask was 0x1FF whereas it should be 0x3FF.
For vsi_num >= 512, FW set_rss_lut has been failing with return code
EACCESS (vsi ownership issue) because software was providing
incorrect vsi_num (dropping 10th bit due to incorrect mask) for
set_rss_lut admin command
Fixes: a90fae1d0755 ("net/ice/base: add admin queue structures and commands") Cc: stable@dpdk.org Signed-off-by: Kiran Patil <kiran.patil@intel.com> Signed-off-by: Paul M Stillwell Jr <paul.m.stillwell.jr@intel.com> Signed-off-by: Qi Zhang <qi.z.zhang@intel.com> Acked-by: Qiming Yang <qiming.yang@intel.com>
Qi Zhang [Thu, 11 Jun 2020 08:43:28 +0000 (16:43 +0800)]
net/ice/base: choose TCP dummy packet by protocol
In order to find proper dummy packets for switch filter,
it need to check ipv4 next protocol number, if it is 0x06,
which means next payload is TCP, we need to use TCP
format dummy packet.
Signed-off-by: Wei Zhao <wei.zhao1@intel.com> Signed-off-by: Paul M Stillwell Jr <paul.m.stillwell.jr@intel.com> Signed-off-by: Qi Zhang <qi.z.zhang@intel.com> Acked-by: Qiming Yang <qiming.yang@intel.com>
Qi Zhang [Thu, 11 Jun 2020 08:43:27 +0000 (16:43 +0800)]
net/ice/base: get tunnel type for recipe
This patch add support to get tunnel type of recipe
after get recipe from FW. This will fix the issue in
function ice_find_recp() for tunnel type comparing.
Signed-off-by: Wei Zhao <wei.zhao1@intel.com> Signed-off-by: Paul M Stillwell Jr <paul.m.stillwell.jr@intel.com> Signed-off-by: Qi Zhang <qi.z.zhang@intel.com> Acked-by: Qiming Yang <qiming.yang@intel.com>
Qi Zhang [Thu, 11 Jun 2020 08:43:26 +0000 (16:43 +0800)]
net/ice/base: support flow director for GTPU with outer IPv6
Add FDIR support for MAC_IPV6_GTPU type with outer IPv6 address, teid
and qfi fields matching.
Signed-off-by: Junfeng Guo <junfeng.guo@intel.com> Signed-off-by: Paul M Stillwell Jr <paul.m.stillwell.jr@intel.com> Signed-off-by: Qi Zhang <qi.z.zhang@intel.com> Acked-by: Qiming Yang <qiming.yang@intel.com>
Qi Zhang [Thu, 11 Jun 2020 08:43:25 +0000 (16:43 +0800)]
net/ice/base: rename misleading variable
The grst_delay variable in ice_check_reset contains the maximum time
(in 100 msec units) that the driver will wait for a reset event to
transition to the Device Active state. The value is the sum of three
separate components:
1) The maximum time it may take for the firmware to process its
outstanding command before handling the reset request.
2) The value in RSTCTL.GRSTDEL (the delay firmware inserts between first
seeing the driver reset request and the actual hardware assertion).
3) The maximum expected reset processing time in hardware.
Referring to this total time as "grst_delay" is misleading and
potentially confusing to someone checking the code and cross-referencing
the hardware specification.
Fix this by renaming the variable to "grst_timeout", which is more
descriptive of its actual use.
Signed-off-by: Nick Nunley <nicholas.d.nunley@intel.com> Signed-off-by: Paul M Stillwell Jr <paul.m.stillwell.jr@intel.com> Signed-off-by: Qi Zhang <qi.z.zhang@intel.com> Acked-by: Qiming Yang <qiming.yang@intel.com>
Qi Zhang [Thu, 11 Jun 2020 08:43:24 +0000 (16:43 +0800)]
net/ice/base: add commands for system diagnostic
System diagnostic solution extend the ability to fetch FW
internal status data and error indication.
Signed-off-by: Sharon Haroni <sharon.haroni@intel.com> Signed-off-by: Paul M Stillwell Jr <paul.m.stillwell.jr@intel.com> Signed-off-by: Qi Zhang <qi.z.zhang@intel.com> Acked-by: Qiming Yang <qiming.yang@intel.com>
Qi Zhang [Thu, 11 Jun 2020 08:43:23 +0000 (16:43 +0800)]
net/ice/base: support flow director for outer IP of GTPU
Add outer IP address fields while generating the training packets for
GTPU, so that we can support FDIR based on outer IP of GTPU.
Signed-off-by: Junfeng Guo <junfeng.guo@intel.com> Signed-off-by: Paul M Stillwell Jr <paul.m.stillwell.jr@intel.com> Signed-off-by: Qi Zhang <qi.z.zhang@intel.com> Acked-by: Qiming Yang <qiming.yang@intel.com>
Qi Zhang [Thu, 11 Jun 2020 08:43:22 +0000 (16:43 +0800)]
net/ice/base: refactor to avoid need to retry
The ice_discover_caps function is used to read the device and function
capabilities, updating the hardware capabilities structures with
relevant data.
The exact number of capabilities returned by the hardware is unknown
ahead of time. The AdminQ command will report the total number of
capabilities in the return buffer.
The current implementation involves requesting capabilities once,
reading this returned size, and then re-requested with that size.
This isn't really necessary. The firmware interface has a maximum size
of ICE_AQ_MAX_BUF_LEN. Firmware can never return more than
ICE_AQ_MAX_BUF_LEN / sizeof(struct ice_aqc_list_caps_elem) capabilities.
Avoid the retry loop by simply allocating a buffer of size
ICE_AQ_MAX_BUF_LEN. This is significantly simpler than retrying. The
extra allocation isn't a big deal, as it will be released after we
finish parsing the capabilities.
Signed-off-by: Jacob Keller <jacob.e.keller@intel.com> Signed-off-by: Paul M Stillwell Jr <paul.m.stillwell.jr@intel.com> Signed-off-by: Qi Zhang <qi.z.zhang@intel.com> Acked-by: Qiming Yang <qiming.yang@intel.com>
Qi Zhang [Thu, 11 Jun 2020 08:43:21 +0000 (16:43 +0800)]
net/ice/base: adjust profile ID map locks
The profile id map lock should be held till the caller completes
all references of that profile entries.
The current code releases the lock right after the match search.
This caused a driver issue when the profile map entries were
referenced after it was freed in other thread after the lock was
released earlier.
Also return type of get/set profile functions were changed to
return the ice status instead of the profile entry pointer.
This will prevent the caller referencing the profile fields
outside the lock.
Signed-off-by: Victor Raj <victor.raj@intel.com> Signed-off-by: Paul M Stillwell Jr <paul.m.stillwell.jr@intel.com> Signed-off-by: Qi Zhang <qi.z.zhang@intel.com> Acked-by: Qiming Yang <qiming.yang@intel.com>
Thomas Monjalon [Sun, 14 Jun 2020 22:18:44 +0000 (00:18 +0200)]
devtools: allow non-standard toolchain in meson test
If a compiler is not found in $PATH, the compilation test is skipped.
In some cases, the compiler could be found after extending $PATH
in an environment configuration script (called by load-devel-config).
The decision to skip is deferred to a later stage, after loading the
configuration script.
In such case, the variable DPDK_TARGET, used by the configuration script
as input, is the compiler name.
Signed-off-by: Thomas Monjalon <thomas@monjalon.net> Reviewed-by: David Christensen <drc@linux.vnet.ibm.com>
Thomas Monjalon [Sun, 14 Jun 2020 22:03:29 +0000 (00:03 +0200)]
devtools: shrink cross-compilation test definition
Each cross-compilation case needs to define the target compiler
and the meson cross file.
Given the compiler is already defined in the cross file,
the latter is enough.
The function "build" is changed to accept a cross file alternatively
to the compiler name. In the case of a file (detected if readable),
the compiler is extracted with sed and tr, and the option --cross-file
is automatically added.
Signed-off-by: Thomas Monjalon <thomas@monjalon.net> Reviewed-by: David Christensen <drc@linux.vnet.ibm.com>
Tal Shnaiderman [Mon, 29 Jun 2020 12:37:40 +0000 (15:37 +0300)]
bus/pci: support Windows with bifurcated drivers
Uses SetupAPI.h functions to scan PCI tree.
Uses DEVPKEY_Device_Numa_Node to get the PCI NUMA node.
Uses SPDRP_BUSNUMBER and SPDRP_BUSNUMBER to get the BDF.
scanning currently supports types RTE_KDRV_NONE.
Signed-off-by: Tal Shnaiderman <talshn@mellanox.com>
Tal Shnaiderman [Mon, 29 Jun 2020 12:37:36 +0000 (15:37 +0300)]
pci: fix address domain format size
the struct rte_pci_addr defines domain as uint32_t variable however
the PCI_PRI_FMT macro used for logging the struct sets the format
of domain to uint16_t.
The mismatch causes the following warning messages
in Windows clang build:
format specifies type 'unsigned short' but the argument
has type 'uint32_t' (aka 'unsigned int') [-Wformat]
Fixes: af75078fece3 ("first public release") Cc: stable@dpdk.org Signed-off-by: Tal Shnaiderman <talshn@mellanox.com>
Tal Shnaiderman [Mon, 29 Jun 2020 12:37:35 +0000 (15:37 +0300)]
pci: build on Windows
Added <sys/types.h> in rte_pci header file
to include off_t type since it is missing for Windows.
Define the implementation of the Linux function rte_pci_get_sysfs_path
in pci_common.c for Linux OS only as it is unneeded for other OSs
and to avoid the warning on deprecated call to getenv() on Windows:
"warning: 'getenv' is deprecated: This function or variable may be unsafe.
Consider using _dupenv_s instead."
Signed-off-by: Tal Shnaiderman <talshn@mellanox.com>
Tal Shnaiderman [Mon, 29 Jun 2020 12:37:41 +0000 (15:37 +0300)]
build: generate version map file for MinGW
The MinGW build for Windows has special cases where exported
function contain additional prefix:
__emutls_v.per_lcore__*
To avoid adding those prefixed functions to the version.map file
the map_to_def.py script was modified to create a map file for MinGW
with the needed changed.
The file name was changed to map_to_win.py and lib/meson.build map output
was unified with drivers/meson.build output
Signed-off-by: Tal Shnaiderman <talshn@mellanox.com>
Thomas Monjalon [Wed, 17 Jun 2020 23:52:32 +0000 (01:52 +0200)]
mk: add a paused deprecation warning before each build
DPDK 20.05 had some deprecation notes after "make config"
and after the build.
For DPDK 20.08, the config note is replaced with a warning
before the config and before the build.
After the warning, there is a pause which can be skipped
with the variable MAKE_PAUSE.
This deprecation process was discussed in the Technical Board:
http://mails.dpdk.org/archives/dev/2020-April/162839.html
Signed-off-by: Thomas Monjalon <thomas@monjalon.net> Acked-by: Bruce Richardson <bruce.richardson@intel.com> Acked-by: David Marchand <david.marchand@redhat.com>
Thomas Monjalon [Wed, 17 Jun 2020 21:41:25 +0000 (23:41 +0200)]
doc: update build instructions in the Linux guide
Before removing the "make" build system completely,
the Linux guide instructions are made more concise and accurate.
Some detailed explanations are also available in
doc/guides/prog_guide/dev_kit_root_make_help.rst
This is the swan song for makefile system,
in order to have accurate information backported in LTS.
Cc: stable@dpdk.org Signed-off-by: Thomas Monjalon <thomas@monjalon.net> Acked-by: Bruce Richardson <bruce.richardson@intel.com> Acked-by: David Marchand <david.marchand@redhat.com>
Thomas Monjalon [Wed, 17 Jun 2020 21:37:51 +0000 (23:37 +0200)]
doc: remove some build instructions where unneeded
The build should be described only in few places,
in order to maintain up-to-date, accurate and detailed instructions.
This change is removing some of the unneeded repetitions.
Signed-off-by: Thomas Monjalon <thomas@monjalon.net> Acked-by: Ajit Khaparde <ajit.khaparde@broadcom.com> Acked-by: Jay Zhou <jianjay.zhou@huawei.com> Acked-by: Bruce Richardson <bruce.richardson@intel.com> Acked-by: David Marchand <david.marchand@redhat.com>
Thomas Monjalon [Wed, 17 Jun 2020 21:29:32 +0000 (23:29 +0200)]
doc: remove outdated guidelines for library addition
There was a doc about how to extend DPDK by adding a library.
It could have been useful but was never updated,
so it is lacking a lot of explanations about doxygen,
meson, versioning, maintainership, etc.
Anyway such guidelines should fit in the contributors guide.
Better to completely remove this obsolete document.
Signed-off-by: Thomas Monjalon <thomas@monjalon.net> Acked-by: Bruce Richardson <bruce.richardson@intel.com> Acked-by: David Marchand <david.marchand@redhat.com>
Wisam Jaddo [Thu, 4 Jun 2020 13:34:59 +0000 (13:34 +0000)]
app/flow-perf: add insertion rate calculation
Add insertion rate calculation feature into flow
performance application.
The application now provide the ability to test
insertion rate of specific rte_flow rule, by
stressing it to the NIC, and calculate the
insertion rate.
The application offers some options in the command
line, to configure which rule to apply.
After that the application will start producing
rules with same pattern but increasing the outer IP
source address by 1 each time, thus it will give
different flow each time, and all other items will
have open masks.
The current design have single core insertion rate.
Signed-off-by: Wisam Jaddo <wisamm@mellanox.com> Acked-by: Xiaoyu Min <jackmin@mellanox.com>
Xiaolong Ye [Sat, 13 Jun 2020 15:49:19 +0000 (23:49 +0800)]
mbuf: fix free space update for dynamic field
The value free_space[i] is used to save the size of biggest aligned
element that can fit in the zone, current implementation has one flaw,
for example, if user registers dynfield1 (size = 4, align = 4, req = 124)
first, the free_space would be as below after registration:
Further request dynfield3 (size = 8, align = 8) would fail to register
due to alignment requirement can't be satisfied, though there is enough
space remained in mbuf.
This patch fixes above issue by saving alignment only in aligned zone,
after the fix, above registrations order can be satisfied, free_space
would be like:
Xiaolong Ye [Sat, 13 Jun 2020 15:49:17 +0000 (23:49 +0800)]
mbuf: fix boundary check at dynamic field registration
We should make sure off + size < sizeof(struct rte_mbuf) to avoid
possible out-of-bounds access of free_space array, there is no issue
currently due to the low bits of free_flags (which is adjacent to
free_space) are always set to 0. But we shouldn't rely on it since it's
fragile and layout of struct mbuf_dyn_shm may be changed in the future.
This patch adds boundary check explicitly to avoid potential risk of
out-of-bounds access.
Fixes: 4958ca3a443a ("mbuf: support dynamic fields and flags") Cc: stable@dpdk.org Signed-off-by: Xiaolong Ye <xiaolong.ye@intel.com> Acked-by: Olivier Matz <olivier.matz@6wind.com>
Haiyue Wang [Thu, 25 Jun 2020 03:50:46 +0000 (11:50 +0800)]
bus/pci: fix VF memory access
To fix CVE-2020-12888, the linux vfio-pci module will invalidate mmaps
and block MMIO access on disabled memory, it will send a SIGBUS to the
application:
https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?id=abafbc551fdd
When the application opens the vfio PCI device, the vfio-pci module will
enable the bus memory space through PCI read/write access. According to
the PCIe specification, the 'Memory Space Enable' is always zero for VF:
Table 9-13 Command Register Changes
Bit Location | PF and VF Register Differences | PF | VF
| From Base | Attributes | Attributes
-------------+--------------------------------+------------+-----------
| Memory Space Enable - Does not | |
| apply to VFs. Must be hardwired| Base | 0b
1 | to 0b for VFs. VF Memory Space | |
| is controlled by the VF MSE bit| |
| in the VF Control register. | |
-------------+--------------------------------+------------+-----------
Afterwards the vfio-pci will initialize its own virtual PCI config space
data ('vconfig') by reading the VF's physical PCI config space, then the
'Memory Space Enable' bit in vconfig will always be 0b value. This will
make the vfio-pci treat the BAR memory space as disabled, and the SIGBUS
will be triggered if access these BARs.
By investigation, the VF PCI device *passthrough* into the Guest OS by
QEMU has the 'Memory Space Enable' with 1b value. That's because every
PCI driver will start to enable the memory space, and this action will
be hooked by vfio-pci virtual PCI read/write to set the 'Memory Space
Enable' in vconfig space to 1b. So VF runs in guest OS has 'Mem+', but
VF runs in host OS has 'Mem-'.
Align with PCI working mode in Guest/QEMU/Host, in DPDK, enable the PCI
bus memory space explicitly to avoid access on disabled memory.
Fixes: 33604c31354a ("vfio: refactor PCI BAR mapping") Cc: stable@dpdk.org Signed-off-by: Haiyue Wang <haiyue.wang@intel.com> Acked-by: Anatoly Burakov <anatoly.burakov@intel.com> Tested-by: Harman Kalra <hkalra@marvell.com> Tested-by: David Marchand <david.marchand@redhat.com> Tested-by: Thierry Martin <thierry.martin.public@gmail.com>
David Marchand [Fri, 19 Jun 2020 09:58:28 +0000 (11:58 +0200)]
test/bitops: fix command name
Caught by code review, bitops test name is incorrect.
Fixes: 7660614c11e2 ("test/bitops: add bit operations test case") Signed-off-by: David Marchand <david.marchand@redhat.com> Acked-by: Phil Yang <phil.yang@arm.com>
Long Li [Fri, 12 Jun 2020 00:48:25 +0000 (17:48 -0700)]
bus/vmbus: fix ring buffer mapping
vmbus_map_addr is used as the next start virtual address for mapping ring
buffer. However it's updated based on ring_buf, which is a pointer to an
address on the stack. The next ring buffer may be mapped to an unexpected
address.
Fix this by calculating vmbus_map_addr based on returned virtual address.
Fixes: 3f9277031a2e ("bus/vmbus: fix check for mmap failure") Cc: stable@dpdk.org Signed-off-by: Long Li <longli@microsoft.com> Acked-by: Stephen Hemminger <stephen@networkplumber.org>
Hrvoje Habjanic [Tue, 26 May 2020 17:24:55 +0000 (19:24 +0200)]
sched: fix subport freeing
In function rte_sched_subport_free, there is code to free all allocated
stuff related to scheduler subport.
First there are some checks, and in the end, rte_bitmap_free is called.
Now, rte_bitmap_free is a dummy function, and it just checks if
provided pointer to bitmap is valid or not. So, actual memory for
subport is not freed.
This patch fixes this by removing call to rte_bitmap_free, and
instead calling rte_free.
Fixes: d9213b829a31 ("sched: remove pipe params config from port level") Cc: stable@dpdk.org Signed-off-by: Hrvoje Habjanic <hrvoje.habjanic@zg.ht.hr> Acked-by: Stephen Hemminger <stephen@networkplumber.org> Acked-by: Jasvinder Singh <jasvinder.singh@intel.com>
When printf()'s stdout is line-buffered for terminal, it is fully
buffered for pipes. So, stdout listener can only get the output
when it is flushed (on program termination, when buffer is filled or
manual flush).
stdout buffer might fill slowly since every stats report could be small.
Also when it is fully filled it might contain a part of the last stats
report which makes it very inconvenient for any automation which reads
and parses the output.
Fixes: af75078fece3 ("first public release") Cc: stable@dpdk.org Signed-off-by: Georgiy Levashov <georgiy.levashov@oktetlabs.ru> Signed-off-by: Andrew Rybchenko <arybchenko@solarflare.com>
Jerin Jacob [Wed, 24 Jun 2020 11:46:17 +0000 (17:16 +0530)]
bus/pci: optimize bus scan
In order to optimize the PCI management, RTE_KDRV_NONE based
device driver probing removed by not adding them to list in
the scan phase.
The legacy virtio is the only consumer of RTE_KDRV_NONE based device
driver probe scheme. The legacy virtio support will be available
through the existing VFIO/UIO based kernel driver scheme.
This patch also removes the deprecation notice for the same.
To fill the gap with linux kernel eBPF implementation,
add support for two non-generic instructions:
(BPF_ABS | <size> | BPF_LD) and (BPF_IND | <size> | BPF_LD)
which are used to access packet data.
These instructions can only be used when BPF context is a pointer
to 'struct rte_mbuf' (i.e: RTE_BPF_ARG_PTR_MBUF type).
Signed-off-by: Konstantin Ananyev <konstantin.ananyev@intel.com>
Fady Bader [Wed, 17 Jun 2020 08:24:31 +0000 (11:24 +0300)]
meter: remove inline functions from export list
The code didn't compile when using exported meter functions under Windows.
error LNK2001: unresolved external symbol
rte_meter_srtcm_color_aware_check
error LNK2001: unresolved external symbol
rte_meter_srtcm_color_blind_check
error LNK2001: unresolved external symbol
rte_meter_trtcm_color_aware_check
error LNK2001: unresolved external symbol
rte_meter_trtcm_color_blind_check
error LNK2001: unresolved external symbol
rte_meter_trtcm_rfc4115_color_aware_check
error LNK2001: unresolved external symbol
rte_meter_trtcm_rfc4115_color_blind_check
The cause was that there were some inline functions that were included in
the export list.
To solve this the functions were removed from rte_meter_version.map export
list which are implemented in the header and shouldn't be exported.
Fixes: 655796d2b5fb ("meter: support RFC4115 trTCM") Fixes: 9d41beed24b0 ("lib: provide initial versioning") Cc: stable@dpdk.org Signed-off-by: Fady Bader <fady@mellanox.com>
Fady Bader [Thu, 18 Jun 2020 06:55:21 +0000 (09:55 +0300)]
timer: move from common to Unix directory
EAL common timer doesn't compile under Windows.
Compilation log:
error LNK2019:
unresolved external symbol nanosleep referenced in function
rte_delay_us_sleep
error LNK2019:
unresolved external symbol get_tsc_freq referenced in function set_tsc_freq
error LNK2019:
unresolved external symbol sleep referenced in function set_tsc_freq
The reason was that some functions called POSIX functions.
The solution was to move POSIX dependent functions from common to Unix.
Dmitry Kozlyuk [Sat, 20 Jun 2020 22:35:44 +0000 (01:35 +0300)]
doc: clarify compilation with MinGW-w64
Provide a more direct link for installer download and clarify thread
model choice during installation. As pthread is not a requirement,
remove notice about its possible runtime dependency.
Dmitry Kozlyuk [Sat, 20 Jun 2020 22:35:42 +0000 (01:35 +0300)]
config: never link with pthread on Windows
Even if pthread is provided by the toolchain, it is not needed for DPDK
on Windows, because internal shim is used. As a side-effect, this
enables cross-build with MinGW configured with non-POSIX thread library,
e.g. mcfgthread, which is the default on some distributions.
Karra Satwik [Fri, 12 Jun 2020 22:10:20 +0000 (03:40 +0530)]
net/cxgbe: always enable HASH filter support
Disable all unused firmware resources during init time to give
more resources for HASH (exact-match) filter region and always
request firmware to enable HASH filter support when resources
are available.
Signed-off-by: Karra Satwik <kaara.satwik@chelsio.com> Signed-off-by: Rahul Lakkireddy <rahul.lakkireddy@chelsio.com>
Ophir Munk [Tue, 16 Jun 2020 09:44:45 +0000 (09:44 +0000)]
net/mlx5: add memory region callbacks in per-device cache
Prior to this commit MR operations were verbs based and hard coded under
common/mlx5/linux directory. This commit enables upper layers (e.g.
net/mlx5) to determine which MR operations to use. For example the net
layer could set devx based MR operations in non-Linux environments. The
reg_mr and dereg_mr callbacks are added to the global per-device MR
cache 'struct mlx5_mr_share_cache'.
Ophir Munk [Tue, 16 Jun 2020 09:44:44 +0000 (09:44 +0000)]
common/mlx5: export memory region Verbs operations
The glue verbs operations reg_mr and dereg_mr are wrapped and exported
in functions mlx5_common_verbs_reg_mr and mlx5_common_verbs_dereg_mr
respectively. The exported functions are added to a new file
linux/mlx5_common_verbs.c.
Ophir Munk [Tue, 16 Jun 2020 09:44:43 +0000 (09:44 +0000)]
common/mlx5: remove memory region dependency on Verbs
Replace 'struct ibv_mr *' (in 'struct mlx5_mr') with a new 'struct
mlx5_pmd_mr'. The new struct contains the required MR field: lkey,
addr, len and is independent of ibv.
Rahul Lakkireddy [Fri, 12 Jun 2020 22:07:27 +0000 (03:37 +0530)]
net/cxgbe: ignore flow default masks for unrequested fields
commit 536db938a444 ("net/cxgbe: add devargs to control filtermode and
filtermask") allows configuring hardware to select specific combination
of header fields to match in the incoming packets. However, the default
mask is set for all fields in the requested pattern items, even if the
field is not explicitly set in the combination and results in
validation errors. To prevent this, ignore setting the default masks
for the unrequested fields and the hardware will also ignore them in
validation, accordingly. Also, tweak the filter spec before finalizing
the masks.
Fixes: 536db938a444 ("net/cxgbe: add devargs to control filtermode and filtermask") Cc: stable@dpdk.org Signed-off-by: Rahul Lakkireddy <rahul.lakkireddy@chelsio.com>
Rahul Lakkireddy [Fri, 12 Jun 2020 22:07:26 +0000 (03:37 +0530)]
net/cxgbe: fix SMT leak in filter error and free path
Free up Source MAC Table (SMT) entry properly during filter create
failure and filter delete.
Fixes: 993541b2fa4f ("net/cxgbe: support flow API for source MAC rewrite") Cc: stable@dpdk.org Signed-off-by: Rahul Lakkireddy <rahul.lakkireddy@chelsio.com>
Rahul Lakkireddy [Fri, 12 Jun 2020 22:07:25 +0000 (03:37 +0530)]
net/cxgbe: fix double MPS alloc by flow validate and create
The Multi Port Switch (MPS) entry is allocated twice when both
flow validate and create are invoked, but only freed once during
flow destroy. Avoid double alloc by moving MPS entry allocation
closer to when the filter create request is sent to hardware and
will be ignored for filter validate request.
Fixes: fefee7a619a4 ("net/cxgbe: add flow ops to match based on dest MAC") Cc: stable@dpdk.org Signed-off-by: Rahul Lakkireddy <rahul.lakkireddy@chelsio.com>
Rahul Lakkireddy [Fri, 12 Jun 2020 22:07:23 +0000 (03:37 +0530)]
net/cxgbe: fix CLIP leak in filter error path
Free up Compressed Local IP (CLIP) entry properly during filter
creation failure path. Also consolidate all various tables
cleanup to a common function and invoke it from both wild-card
and exact-match filter paths.
Fixes: af44a577988b ("net/cxgbe: support to offload flows to HASH region") Cc: stable@dpdk.org Signed-off-by: Rahul Lakkireddy <rahul.lakkireddy@chelsio.com>
Qi Zhang [Mon, 15 Jun 2020 02:05:15 +0000 (10:05 +0800)]
net/ice/base: remove PPPoD from PPPoE bitmap
Remove PPPoD's packet type from PPPoE's ptype bitmap.
Signed-off-by: Dan Nowlin <dan.nowlin@intel.com> Signed-off-by: Paul M Stillwell Jr <paul.m.stillwell.jr@intel.com> Signed-off-by: Qi Zhang <qi.z.zhang@intel.com> Acked-by: Qiming Yang <qiming.yang@intel.com>
git.droids-corp.org / dpdk.git / log