Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Cleanups #913

Draft
wants to merge 59 commits into
base: master_devel
Choose a base branch
from
Draft

Conversation

tzafrir-mellanox
Copy link
Contributor

No description provided.

tomer540 and others added 30 commits December 7, 2023 16:48
When cross-compiling, the test binary used to verify basic std::regex functionality
will always fallback on GNU regex which may not necessarily be the expected behavior.
This patch assumes that the cross-compiler has complete std::regex support.

It is the developer's responsibility to verify the functionality of std::regex
of the cross-compiler in use.

Signed-off-by: Markus Theil <[email protected]>
The musl toolchain doesn't have <sys/pci.h>, only <linux/pci.h>,
resulting in a compilation error. This patch checks if <sys/pci.h>
is present and alternatively uses <linux/pci.h> if not.

Signed-off-by: Til Kaiser <[email protected]>
Description:
When Hotplug is activated, resetting the device causes the bdf to change.
As a result, running lspci on the previously known bdf will not be successful due to the altered bdf.
Running mst restart is necessary to re-enumerate the mst devices.

MSTFlint port needed: Yes
Tested OS: Linux, pcie-hpg11-01
Tested devices: mt4125
Tested flows: mlxfwreset -d /dev/mst/mt4125_pciconf0 --sync 1 -y r --log debug

Known gaps (with RM ticket):

Issue: 3720202
Change-Id: Ie3f03115351921086717e3b94670bf9219f88333
Signed-off-by: Itay Avraham <[email protected]>
… inform the user about it

Description:
The purpose of this feature is to give users an indication of what will happen when they execute the reset flow using mlxfwreset while the BF is in controller mode.

When a full chip reset is triggered, and there are no additional hosts except for the ARM itself (especially in controller mode), the ARM will reset the entire DPU.
This includes an ARM reset, and it immediately initiates an ARM OS shutdown.
During this process, the setup will be unavailable for some time, which the user may not expect.

The solution is to check the reset type:
The message will not printed to the user when the reset type is bit 1 ("Keep network port active during reset") or bit 2("NIC only reset for SoC devices").

MSTFlint port needed: Yes.
Tested OS: BF Controller (10.7.4.188) and regular BF (apps-173)
Tested devices: BF3
Tested flows: mlxfwreset -d /dev/mst/mt41692_pciconf0 r

Known gaps (with RM ticket):

Issue: 3725353
Change-Id: If7af544b380b1e275f15383d0803bfd6943972f0
Signed-off-by: Itay Avraham <[email protected]>
Description:
The supported switches for MRSR-1 that are already supported in mlxfwreset: Spectrum, Spectrum-2, Spectrum-3, Switch-IB, Switch-IB-2, Quantum,  Quantum-2.
The supported switches for MRSR-6 that have not been implemented yet and need to be addressed with a new feature request: Quantum-3, Spectrum-4 and above

MSTFlint port needed: No.
Tested OS: Linux (Mellanox OS)
Tested devices: All switches.
Tested flows: mlxfwreset -d <switch> r

Known gaps (with RM ticket):

Issue: 3738146
Change-Id: Ic3e3b555dd811098675c3fbd70996a75aa06204b
Signed-off-by: Itay Avraham <[email protected]>
Description:
[mlxlink] - update amber collect to version 2.8

MSTFlint port needed: yes
Tested OS: None
Tested devices: None
Tested flows: None

Known gaps (with RM ticket):

Issue: 3698551
Description:
updating the supported compliance speed

MSTFlint port needed: None
Tested OS: None
Tested devices: None
Tested flows: None

Known gaps (with RM ticket):

Issue: 3741924
… of MGIR

Description: requested by FR, since MGIR returned the wrong value for NDR SP4
We changed it first to SLRG - but this register has a FW bug when running in 25G per lane or less.
So switched to SLTP.

MSTFlint port needed:
Tested OS:
Tested devices:
Tested flows:

Known gaps (with RM ticket):

Issue: 3706030
Issue: 3747125
…o sp2 & sp3 adb files.

Description:
Auto generated code

MSTFlint port needed: yes
Tested OS:
Tested devices:
Tested flows:

Known gaps (with RM ticket):

Issue: None
Description:

MSTFlint port needed: no
Tested OS:
Tested devices:
Tested flows:

Known gaps (with RM ticket):

Issue: None
Description:

MSTFlint port needed: no
Tested OS:
Tested devices:
Tested flows:

Known gaps (with RM ticket):

Issue: None
Description:
Fixing NDR modulation config in prbs testmode

MSTFlint port needed: Yes
Tested OS: Linux64
Tested devices: QTM-2
Tested flows: mlxlink prbs test_mode

Known gaps (with RM ticket): NA

Issue: 3612477
Description:

-Cleaning mlxlink_commander from any user input parameters
-Moving all user input parameters to UserInput class

MSTFlint port needed: No
Tested OS: Linux64
Tested devices: CX-6
Tested flows: mlxlink full test

Known gaps (with RM ticket): NA

Issue: 3554777
Description:

Adding new flag (--plane) to the internal tool for plane view
The default view will be aggregate view
If the plane is provided then the tool will set plane_ind field
for all access reg requests
MSTFlint port needed: Yes
Tested OS: Linux64
Tested devices: NA
Tested flows: mlxlink --plane X

Known gaps (with RM ticket): NA

Issue: 3566234
…g a text format public key

Description: plain text public key (without authentication_configuration struct) will be applicable for rsa_sign command if the same key was already inserted prior to signing.

MSTFlint port needed: no
Tested OS: linux
Tested devices: image
Tested flows:
rsa_sign CX7 bin with a plaintext pub key that's already in public_keys_2/3 => success
rsa_sign CX7 bin with a new plaintext pub key => failed
rsa_sign CX7 bin with a new public_key_2/3 pub key => success

Known gaps (with RM ticket): N/A

Issue: 3611237
…W update from ARM side

Description: MCC update handle initializtion removed from FwCompsMgr init, will be performed lazily right before each call to read/write.

MSTFlint port needed: yes
Tested OS: linux
Tested devices: BF3, CX7
Tested flows: flint query on BF3 with disabled FW update, flint query on CX7, flint burn on CX7

Known gaps (with RM ticket): N/A

Issue: 3580180
Description: added support for arcuse 4mb flash

MSTFlint port needed: yes
Tested OS: linux
Tested devices: non secure arcuse PLD
Tested flows: flint -ocr q on virtual mtusb

Known gaps (with RM ticket): N/A

Issue:3635328
Description: added support for new winbond and micron flashes

MSTFlint port needed: yes
Tested OS: linux
Tested devices: CX7 with new flashes
Tested flows: flint hw query, flint -ocr burn

Known gaps (with RM ticket): N/A

Issue: 3664099
Description: added support for FS5 image burn start

MSTFlint port needed: yes
Tested OS: linux
Tested devices: QM3 emu
Tested flows: flint -d -i burn

Known gaps (with RM ticket):  N/A

Issue: 3666597
Change-Id: I1a9d2a8c1ff70e7e717cf49af824a5492961b334

Signed-off-by: Alex Shargorodsky <[email protected]>
Description: added FS5 support for -ocr burn.

MSTFlint port needed: yes
Tested OS: linux
Tested devices: QM3
Tested flows: flint -ocr burn

Known gaps (with RM ticket): N/A

Issue: 3672348
Change-Id: I5d3d5d9275e08c579c85e03818547ba795de3ed5

Signed-off-by: Alex Shargorodsky <[email protected]>
Description: added CX8 sw_dev_id

MSTFlint port needed: yes
Tested OS: linux
Tested devices: N/A
Tested flows:
mlxburn -fw mlx -wrimage /tmp/tmp_file.bin -c ini -base_guid 0x0002c903002ef500 -base_mac 0x0002c92ef500 -dev_type 4131

Known gaps (with RM ticket): N/A

Issue: 3676569
…w DTOC offset

Description: support for configuration of root node selection for multi root mlx files.
The ini file will have a "root" field under a "[CONF_FILE_INFO]" section, this field will determine the root node.
In addition, "image_info" itoc section will have a new field "dtoc_offset", indicating where to look for the dTOC header in the image.

MSTFlint port needed: yes
Tested OS: linux
Tested devices: N/A
Tested flows:
Generate images from multi root cx7 mlx file.
Generate image from non multi root cx7 mlx file.
Generate image from cx7 mlx file.

Known gaps (with RM ticket): N/A

Issue: 3673629
Description: added new MCC error and error string

MSTFlint port needed: yes
Tested OS: N/A
Tested devices: N/A
Tested flows: N/A

Known gaps (with RM ticket): N/A

Issue: 3675014
Change-Id: Ica189d09f336d8af1ddefbc558e1f4c1ba94758d

Signed-off-by: Alex Shargorodsky <[email protected]>
lgofman and others added 19 commits January 24, 2024 10:32
Description:
- Supporting the new serdes_tx params for XDR generation
- Supporting eye info

MSTFlint port needed: No
Tested OS: Linux64
Tested devices: NA
Tested flows:
mlxlink eye_info
mlxlink serdes_tx

Known gaps (with RM ticket): NA

Issue: 3482368
Issue:  3700963
Description: fixed uninitialized variables coverity issues.

Tested OS: linux
Tested devices: n/a
Tested flows: compilation

Known gaps (with RM ticket): n/a

Issue: 3657795
Description:
The port flag will take the IB port number for BW port access

MSTFlint port needed: Yes
Tested OS: Linux64
Tested devices: BW00
Tested flows: mlxlink full commands

Known gaps (with RM ticket): N/A

Issue: 3509849
…r "[u]?int[0-9]+_t"-family of types.

This commit addresses compilation failure below, which is caused
by changes in standard headers on modern versions of GCC:

    In file included from elfio/elfio.hpp:34,
                     from cryptodata.cpp:15:
    ./elfio/elf_types.hpp:30:20: error: 'uint16_t' does not name a type
       30 | using Elf_Half   = uint16_t;
          |                    ^~~~~~~~
    ./elfio/elf_types.hpp:1:1: note: 'uint16_t' is defined in header '<cstdint>'; did you forget to '#include <cstdint>'?
      +++ |+#include <cstdint>
        1 | /*

See https://gcc.gnu.org/gcc-13/porting_to.html for reference.

(cherry picked from commit bc33aa0)
Signed-off-by: AlexBlago <[email protected]>
Signed-off-by: Tzafrir Cohen <[email protected]>
Signed-off-by: Tzafrir Cohen <[email protected]>
Signed-off-by: Tzafrir Cohen <[email protected]>
Signed-off-by: Tzafrir Cohen <[email protected]>
FIXME: why is that OK?

Also: removed ignoring a warning that already worked.

Signed-off-by: Tzafrir Cohen <[email protected]>
FIXME: How can we be sure those copies are safe?

Signed-off-by: Tzafrir Cohen <[email protected]>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.