[Kernel-packages] [Bug 2043005] Re: mlxbf-pmc: Fix crspace event offsets
** Tags removed: verification-needed-jammy-linux-bluefield ** Tags added: verification-done-jammy-linux-bluefield -- You received this bug notification because you are a member of Kernel Packages, which is subscribed to linux-bluefield in Ubuntu. https://bugs.launchpad.net/bugs/2043005 Title: mlxbf-pmc: Fix crspace event offsets Status in linux-bluefield package in Ubuntu: Invalid Status in linux-bluefield source package in Jammy: Fix Committed Bug description: SRU Justification: [Impact] Currently, the offsets calculated for programming crspace type events on BlueField-3 are incorrect and this leads to a mismatch of counter numbers between the driver and actual HW where, for example, event2 in the driver is actually event4 in HW. [Fix] Fix the offset calculation for programming crspace events. [Test Case] On a BlueField-3 platform, program any valid event number to the event sysfs and check that the corresponding counter value increments. [Regression Potential] This is a bug fix and the regression potential can be considered minimal. To manage notifications about this bug go to: https://bugs.launchpad.net/ubuntu/+source/linux-bluefield/+bug/2043005/+subscriptions -- Mailing list: https://launchpad.net/~kernel-packages Post to : kernel-packages@lists.launchpad.net Unsubscribe : https://launchpad.net/~kernel-packages More help : https://help.launchpad.net/ListHelp
[Kernel-packages] [Bug 2043005] [NEW] mlxbf-pmc: Fix crspace event offsets
Public bug reported: SRU Justification: [Impact] Currently, the offsets calculated for programming crspace type events on BlueField-3 are incorrect and this leads to a mismatch of counter numbers between the driver and actual HW where, for example, event2 in the driver is actually event4 in HW. [Fix] Fix the offset calculation for programming crspace events. [Test Case] On a BlueField-3 platform, program any valid event number to the event sysfs and check that the corresponding counter value increments. [Regression Potential] This is a bug fix and the regression potential can be considered minimal. ** Affects: linux-bluefield (Ubuntu) Importance: Undecided Status: New -- You received this bug notification because you are a member of Kernel Packages, which is subscribed to linux-bluefield in Ubuntu. https://bugs.launchpad.net/bugs/2043005 Title: mlxbf-pmc: Fix crspace event offsets Status in linux-bluefield package in Ubuntu: New Bug description: SRU Justification: [Impact] Currently, the offsets calculated for programming crspace type events on BlueField-3 are incorrect and this leads to a mismatch of counter numbers between the driver and actual HW where, for example, event2 in the driver is actually event4 in HW. [Fix] Fix the offset calculation for programming crspace events. [Test Case] On a BlueField-3 platform, program any valid event number to the event sysfs and check that the corresponding counter value increments. [Regression Potential] This is a bug fix and the regression potential can be considered minimal. To manage notifications about this bug go to: https://bugs.launchpad.net/ubuntu/+source/linux-bluefield/+bug/2043005/+subscriptions -- Mailing list: https://launchpad.net/~kernel-packages Post to : kernel-packages@lists.launchpad.net Unsubscribe : https://launchpad.net/~kernel-packages More help : https://help.launchpad.net/ListHelp
[Kernel-packages] [Bug 2042525] Re: mlxbf-pmc: Support 64-bit counter and counting cycles
** Tags removed: verification-needed-jammy-linux-bluefield ** Tags added: verification-done-jammy-linux-bluefield -- You received this bug notification because you are a member of Kernel Packages, which is subscribed to linux-bluefield in Ubuntu. https://bugs.launchpad.net/bugs/2042525 Title: mlxbf-pmc: Support 64-bit counter and counting cycles Status in linux-bluefield package in Ubuntu: Invalid Status in linux-bluefield source package in Jammy: Fix Committed Bug description: SRU Justification: [Impact] Currently, there is no method to keep track of the cycle count when an event is being monitored. And since cycle count increments quickly, the 32-bit counter values could wrap around and hence support for 64-bit counters is also needed for the same. [Fix] Expose 2 additional sysfs entries: count_clock and use_odd_counter. These fields are supported in BlueField-3 PMC hardware and each bit in count_clock corresponds to each counter, while each bit in use_odd_counter corresponds to an even counter.Exposing these fields will allow the user to program any counter of choice to monitor the cycle count for the required duration. Similarly, use_odd_counter can be set to couple 2 adjacent odd and even counters to form a 64-bit counter. [Test Case] 1. Verify that count_clock and use_odd_counter sysfs entries are created for each BlueField-3 HW block. These are not supported by BlueField-1 or BlueField-2 HW. 2. Set any bit in count_clock and check if the corresponding counter values increment after enabling. 3. Set any bit in use_odd_counter and check if the cycle count increments on 2 counters with the odd counter being the lower 32 bits and the even counter being the upper 32 bits. [Regression Potential] Can be considered minimal. To manage notifications about this bug go to: https://bugs.launchpad.net/ubuntu/+source/linux-bluefield/+bug/2042525/+subscriptions -- Mailing list: https://launchpad.net/~kernel-packages Post to : kernel-packages@lists.launchpad.net Unsubscribe : https://launchpad.net/~kernel-packages More help : https://help.launchpad.net/ListHelp
[Kernel-packages] [Bug 2042527] Re: mlxbf-pmc: add support for clock_measure counters
** Tags removed: verification-needed-jammy-linux-bluefield ** Tags added: verification-done-jammy-linux-bluefield -- You received this bug notification because you are a member of Kernel Packages, which is subscribed to linux-bluefield in Ubuntu. https://bugs.launchpad.net/bugs/2042527 Title: mlxbf-pmc: add support for clock_measure counters Status in linux-bluefield package in Ubuntu: Invalid Status in linux-bluefield source package in Jammy: Fix Committed Bug description: SRU Justification: [Impact] Currently, the counters under the clock_measure HW block are not supported by the driver and is a new requirement for calculating performance metrics. [Fix] Add support for "clock_measure" block by passing the block info via ACPI. The events monitored by these counters is added to the driver and exposed via a new sysfs sub-directory for this block. [Test Case] Read any of the registers listed under the "clock_measure" sub-directory. These fields are read-only and hence all writes will be blocked by the driver. [Regression Potential] Can be considered minimal. To manage notifications about this bug go to: https://bugs.launchpad.net/ubuntu/+source/linux-bluefield/+bug/2042527/+subscriptions -- Mailing list: https://launchpad.net/~kernel-packages Post to : kernel-packages@lists.launchpad.net Unsubscribe : https://launchpad.net/~kernel-packages More help : https://help.launchpad.net/ListHelp
[Kernel-packages] [Bug 2042527] [NEW] mlxbf-pmc: add support for clock_measure counters
Public bug reported: SRU Justification: [Impact] Currently, the counters under the clock_measure HW block are not supported by the driver and is a new requirement for calculating performance metrics. [Fix] Add support for "clock_measure" block by passing the block info via ACPI. The events monitored by these counters is added to the driver and exposed via a new sysfs sub-directory for this block. [Test Case] Read any of the registers listed under the "clock_measure" sub-directory. These fields are read-only and hence all writes will be blocked by the driver. [Regression Potential] Can be considered minimal. ** Affects: linux-bluefield (Ubuntu) Importance: Undecided Status: New -- You received this bug notification because you are a member of Kernel Packages, which is subscribed to linux-bluefield in Ubuntu. https://bugs.launchpad.net/bugs/2042527 Title: mlxbf-pmc: add support for clock_measure counters Status in linux-bluefield package in Ubuntu: New Bug description: SRU Justification: [Impact] Currently, the counters under the clock_measure HW block are not supported by the driver and is a new requirement for calculating performance metrics. [Fix] Add support for "clock_measure" block by passing the block info via ACPI. The events monitored by these counters is added to the driver and exposed via a new sysfs sub-directory for this block. [Test Case] Read any of the registers listed under the "clock_measure" sub-directory. These fields are read-only and hence all writes will be blocked by the driver. [Regression Potential] Can be considered minimal. To manage notifications about this bug go to: https://bugs.launchpad.net/ubuntu/+source/linux-bluefield/+bug/2042527/+subscriptions -- Mailing list: https://launchpad.net/~kernel-packages Post to : kernel-packages@lists.launchpad.net Unsubscribe : https://launchpad.net/~kernel-packages More help : https://help.launchpad.net/ListHelp
[Kernel-packages] [Bug 2042525] [NEW] mlxbf-pmc: Support 64-bit counter and counting cycles
Public bug reported: SRU Justification: [Impact] Currently, there is no method to keep track of the cycle count when an event is being monitored. And since cycle count increments quickly, the 32-bit counter values could wrap around and hence support for 64-bit counters is also needed for the same. [Fix] Expose 2 additional sysfs entries: count_clock and use_odd_counter. These fields are supported in BlueField-3 PMC hardware and each bit in count_clock corresponds to each counter, while each bit in use_odd_counter corresponds to an even counter.Exposing these fields will allow the user to program any counter of choice to monitor the cycle count for the required duration. Similarly, use_odd_counter can be set to couple 2 adjacent odd and even counters to form a 64-bit counter. [Test Case] 1. Verify that count_clock and use_odd_counter sysfs entries are created for each BlueField-3 HW block. These are not supported by BlueField-1 or BlueField-2 HW. 2. Set any bit in count_clock and check if the corresponding counter values increment after enabling. 3. Set any bit in use_odd_counter and check if the cycle count increments on 2 counters with the odd counter being the lower 32 bits and the even counter being the upper 32 bits. [Regression Potential] Can be considered minimal. ** Affects: linux-bluefield (Ubuntu) Importance: Undecided Status: New -- You received this bug notification because you are a member of Kernel Packages, which is subscribed to linux-bluefield in Ubuntu. https://bugs.launchpad.net/bugs/2042525 Title: mlxbf-pmc: Support 64-bit counter and counting cycles Status in linux-bluefield package in Ubuntu: New Bug description: SRU Justification: [Impact] Currently, there is no method to keep track of the cycle count when an event is being monitored. And since cycle count increments quickly, the 32-bit counter values could wrap around and hence support for 64-bit counters is also needed for the same. [Fix] Expose 2 additional sysfs entries: count_clock and use_odd_counter. These fields are supported in BlueField-3 PMC hardware and each bit in count_clock corresponds to each counter, while each bit in use_odd_counter corresponds to an even counter.Exposing these fields will allow the user to program any counter of choice to monitor the cycle count for the required duration. Similarly, use_odd_counter can be set to couple 2 adjacent odd and even counters to form a 64-bit counter. [Test Case] 1. Verify that count_clock and use_odd_counter sysfs entries are created for each BlueField-3 HW block. These are not supported by BlueField-1 or BlueField-2 HW. 2. Set any bit in count_clock and check if the corresponding counter values increment after enabling. 3. Set any bit in use_odd_counter and check if the cycle count increments on 2 counters with the odd counter being the lower 32 bits and the even counter being the upper 32 bits. [Regression Potential] Can be considered minimal. To manage notifications about this bug go to: https://bugs.launchpad.net/ubuntu/+source/linux-bluefield/+bug/2042525/+subscriptions -- Mailing list: https://launchpad.net/~kernel-packages Post to : kernel-packages@lists.launchpad.net Unsubscribe : https://launchpad.net/~kernel-packages More help : https://help.launchpad.net/ListHelp
[Kernel-packages] [Bug 2039561] Re: mlxbf_pmc: Replace sauce patches with upstream commits
** Tags removed: verification-needed-jammy-linux-bluefield ** Tags added: verification-done-jammy-linux-bluefield -- You received this bug notification because you are a member of Kernel Packages, which is subscribed to linux-bluefield in Ubuntu. https://bugs.launchpad.net/bugs/2039561 Title: mlxbf_pmc: Replace sauce patches with upstream commits Status in linux-bluefield package in Ubuntu: Invalid Status in linux-bluefield source package in Jammy: Fix Committed Bug description: SRU Justification: [Impact] The mlxbf_pmc BlueField platform driver in the Jammy repo consists of some SAUCE patches. These need to be replaced. [Fix] The fix is to revert the four SAUCE patches, replacing them with upstream commits for the same functionality. 2 patches are from upstream linux, while the 3rd is from linux-next. [Test Case] * Boot BF2/BF3 platform, verify no new errors * Test sysfs interface exposed by driver for programming various counters and events. [Regression Potential] The upstream commits are not exactly the same as the SAUCE patches, so technically there is a chance of regression, but its been well-tested and the functionality is the same. [Other] n/a To manage notifications about this bug go to: https://bugs.launchpad.net/ubuntu/+source/linux-bluefield/+bug/2039561/+subscriptions -- Mailing list: https://launchpad.net/~kernel-packages Post to : kernel-packages@lists.launchpad.net Unsubscribe : https://launchpad.net/~kernel-packages More help : https://help.launchpad.net/ListHelp
[Kernel-packages] [Bug 2039561] [NEW] mlxbf_pmc: Replace sauce patches with upstream commits
Public bug reported: SRU Justification: [Impact] The mlxbf_pmc BlueField platform driver in the Jammy repo consists of some SAUCE patches. These need to be replaced. [Fix] The fix is to revert the four SAUCE patches, replacing them with upstream commits for the same functionality. 2 patches are from upstream linux, while the 3rd is from linux-next. [Test Case] * Boot BF2/BF3 platform, verify no new errors * Test sysfs interface exposed by driver for programming various counters and events. [Regression Potential] The upstream commits are not exactly the same as the SAUCE patches, so technically there is a chance of regression, but its been well-tested and the functionality is the same. [Other] n/a ** Affects: linux-bluefield (Ubuntu) Importance: Undecided Status: New -- You received this bug notification because you are a member of Kernel Packages, which is subscribed to linux-bluefield in Ubuntu. https://bugs.launchpad.net/bugs/2039561 Title: mlxbf_pmc: Replace sauce patches with upstream commits Status in linux-bluefield package in Ubuntu: New Bug description: SRU Justification: [Impact] The mlxbf_pmc BlueField platform driver in the Jammy repo consists of some SAUCE patches. These need to be replaced. [Fix] The fix is to revert the four SAUCE patches, replacing them with upstream commits for the same functionality. 2 patches are from upstream linux, while the 3rd is from linux-next. [Test Case] * Boot BF2/BF3 platform, verify no new errors * Test sysfs interface exposed by driver for programming various counters and events. [Regression Potential] The upstream commits are not exactly the same as the SAUCE patches, so technically there is a chance of regression, but its been well-tested and the functionality is the same. [Other] n/a To manage notifications about this bug go to: https://bugs.launchpad.net/ubuntu/+source/linux-bluefield/+bug/2039561/+subscriptions -- Mailing list: https://launchpad.net/~kernel-packages Post to : kernel-packages@lists.launchpad.net Unsubscribe : https://launchpad.net/~kernel-packages More help : https://help.launchpad.net/ListHelp
[Kernel-packages] [Bug 2004235] Re: mlxbf-pmc counters not functional for llt_miss block
** Tags removed: verification-needed-jammy ** Tags added: verification-done-jammy -- You received this bug notification because you are a member of Kernel Packages, which is subscribed to linux-bluefield in Ubuntu. https://bugs.launchpad.net/bugs/2004235 Title: mlxbf-pmc counters not functional for llt_miss block Status in linux-bluefield package in Ubuntu: Invalid Status in linux-bluefield source package in Focal: Fix Committed Status in linux-bluefield source package in Jammy: Fix Committed Bug description: SRU Justification: [Impact] Currently, the driver uses fixed offsets for the control and data registers in the performance counter module. However, this is not correct since the module is not uniform in all HW blocks, and in this case, the llt_miss block is different from the rest. [Fix] The number of counters present in each block is not fixed, and hence the offsets for the control and data registers should also be calculated accordingly. [Test Case] Without the fix, writing 1 to the llt_miss "enable" file will not result in any of the counters incrementing. With the fix, the counters should increment even with the default event settings. [Regression Potential] Can be considered minimal. To manage notifications about this bug go to: https://bugs.launchpad.net/ubuntu/+source/linux-bluefield/+bug/2004235/+subscriptions -- Mailing list: https://launchpad.net/~kernel-packages Post to : kernel-packages@lists.launchpad.net Unsubscribe : https://launchpad.net/~kernel-packages More help : https://help.launchpad.net/ListHelp
[Kernel-packages] [Bug 2002501] Re: mlxbf-pmc support for BlueField-3
** Tags removed: verification-needed-focal ** Tags added: verification-done-focal -- You received this bug notification because you are a member of Kernel Packages, which is subscribed to linux-bluefield in Ubuntu. https://bugs.launchpad.net/bugs/2002501 Title: mlxbf-pmc support for BlueField-3 Status in linux-bluefield package in Ubuntu: New Status in linux-bluefield source package in Focal: Fix Committed Status in linux-bluefield source package in Jammy: Fix Committed Bug description: SRU Justification: [Impact] Currently, the mlxbf-pmc driver supports only BlueField-1 and BlueField-2 devices. BlueField-3 is not supported. [Fix] Add support for BlueField-3 devices by adding new access mechanism for programming and reading counter values as in the new design, along with the list of supported events in each HW block in BlueField-3. [Test Case] Previously, the driver would not be loaded when the OS is loaded on a BlueField-3 card. With the fix, the driver should get loaded and create sysfs entries as a hwmon device under /sys/class/hwmon/hwmon0. [Regression Potential] The ACPI table needs to be in sync with this change since it depends on newly introduces parameters such as mss_enable and llt_enable, which were not present earlier. To manage notifications about this bug go to: https://bugs.launchpad.net/ubuntu/+source/linux-bluefield/+bug/2002501/+subscriptions -- Mailing list: https://launchpad.net/~kernel-packages Post to : kernel-packages@lists.launchpad.net Unsubscribe : https://launchpad.net/~kernel-packages More help : https://help.launchpad.net/ListHelp
[Kernel-packages] [Bug 2004235] Re: mlxbf-pmc counters not functional for llt_miss block
** Tags removed: verification-needed-focal ** Tags added: verification-done-focal -- You received this bug notification because you are a member of Kernel Packages, which is subscribed to linux-bluefield in Ubuntu. https://bugs.launchpad.net/bugs/2004235 Title: mlxbf-pmc counters not functional for llt_miss block Status in linux-bluefield package in Ubuntu: Invalid Status in linux-bluefield source package in Focal: Fix Committed Status in linux-bluefield source package in Jammy: Fix Committed Bug description: SRU Justification: [Impact] Currently, the driver uses fixed offsets for the control and data registers in the performance counter module. However, this is not correct since the module is not uniform in all HW blocks, and in this case, the llt_miss block is different from the rest. [Fix] The number of counters present in each block is not fixed, and hence the offsets for the control and data registers should also be calculated accordingly. [Test Case] Without the fix, writing 1 to the llt_miss "enable" file will not result in any of the counters incrementing. With the fix, the counters should increment even with the default event settings. [Regression Potential] Can be considered minimal. To manage notifications about this bug go to: https://bugs.launchpad.net/ubuntu/+source/linux-bluefield/+bug/2004235/+subscriptions -- Mailing list: https://launchpad.net/~kernel-packages Post to : kernel-packages@lists.launchpad.net Unsubscribe : https://launchpad.net/~kernel-packages More help : https://help.launchpad.net/ListHelp
[Kernel-packages] [Bug 2004234] [NEW] mlxbf-pmc counters not functional for llt_miss block
Public bug reported: SRU Justification: [Impact] Currently, the driver uses fixed offsets for the control and data registers in the performance counter module. However, this is not correct since the module is not uniform in all HW blocks, and in this case, the llt_miss block is different from the rest. [Fix] The number of counters present in each block is not fixed, and hence the offsets for the control and data registers should also be calculated accordingly. [Test Case] Without the fix, writing 1 to the llt_miss "enable" file will not result in any of the counters incrementing. With the fix, the counters should increment even with the default event settings. [Regression Potential] Can be considered minimal. ** Affects: linux-bluefield (Ubuntu) Importance: Undecided Status: New -- You received this bug notification because you are a member of Kernel Packages, which is subscribed to linux-bluefield in Ubuntu. https://bugs.launchpad.net/bugs/2004234 Title: mlxbf-pmc counters not functional for llt_miss block Status in linux-bluefield package in Ubuntu: New Bug description: SRU Justification: [Impact] Currently, the driver uses fixed offsets for the control and data registers in the performance counter module. However, this is not correct since the module is not uniform in all HW blocks, and in this case, the llt_miss block is different from the rest. [Fix] The number of counters present in each block is not fixed, and hence the offsets for the control and data registers should also be calculated accordingly. [Test Case] Without the fix, writing 1 to the llt_miss "enable" file will not result in any of the counters incrementing. With the fix, the counters should increment even with the default event settings. [Regression Potential] Can be considered minimal. To manage notifications about this bug go to: https://bugs.launchpad.net/ubuntu/+source/linux-bluefield/+bug/2004234/+subscriptions -- Mailing list: https://launchpad.net/~kernel-packages Post to : kernel-packages@lists.launchpad.net Unsubscribe : https://launchpad.net/~kernel-packages More help : https://help.launchpad.net/ListHelp
[Kernel-packages] [Bug 2004235] [NEW] mlxbf-pmc counters not functional for llt_miss block
Public bug reported: SRU Justification: [Impact] Currently, the driver uses fixed offsets for the control and data registers in the performance counter module. However, this is not correct since the module is not uniform in all HW blocks, and in this case, the llt_miss block is different from the rest. [Fix] The number of counters present in each block is not fixed, and hence the offsets for the control and data registers should also be calculated accordingly. [Test Case] Without the fix, writing 1 to the llt_miss "enable" file will not result in any of the counters incrementing. With the fix, the counters should increment even with the default event settings. [Regression Potential] Can be considered minimal. ** Affects: linux-bluefield (Ubuntu) Importance: Undecided Status: New -- You received this bug notification because you are a member of Kernel Packages, which is subscribed to linux-bluefield in Ubuntu. https://bugs.launchpad.net/bugs/2004235 Title: mlxbf-pmc counters not functional for llt_miss block Status in linux-bluefield package in Ubuntu: New Bug description: SRU Justification: [Impact] Currently, the driver uses fixed offsets for the control and data registers in the performance counter module. However, this is not correct since the module is not uniform in all HW blocks, and in this case, the llt_miss block is different from the rest. [Fix] The number of counters present in each block is not fixed, and hence the offsets for the control and data registers should also be calculated accordingly. [Test Case] Without the fix, writing 1 to the llt_miss "enable" file will not result in any of the counters incrementing. With the fix, the counters should increment even with the default event settings. [Regression Potential] Can be considered minimal. To manage notifications about this bug go to: https://bugs.launchpad.net/ubuntu/+source/linux-bluefield/+bug/2004235/+subscriptions -- Mailing list: https://launchpad.net/~kernel-packages Post to : kernel-packages@lists.launchpad.net Unsubscribe : https://launchpad.net/~kernel-packages More help : https://help.launchpad.net/ListHelp
[Kernel-packages] [Bug 2002500] [NEW] mlxbf-pmc support for BlueField-3
Public bug reported: SRU Justification: [Impact] Currently, the mlxbf-pmc driver supports only BlueField-1 and BlueField-2 devices. BlueField-3 is not supported. [Fix] Add support for BlueField-3 devices by adding new access mechanism for programming and reading counter values as in the new design, along with the list of supported events in each HW block in BlueField-3. [Test Case] Previously, the driver would not be loaded when the OS is loaded on a BlueField-3 card. With the fix, the driver should get loaded and create sysfs entries as a hwmon device under /sys/class/hwmon/hwmon0. [Regression Potential] The ACPI table needs to be in sync with this change since it depends on newly introduces parameters such as mss_enable and llt_enable, which were not present earlier. ** Affects: linux-bluefield (Ubuntu) Importance: Undecided Status: New -- You received this bug notification because you are a member of Kernel Packages, which is subscribed to linux-bluefield in Ubuntu. https://bugs.launchpad.net/bugs/2002500 Title: mlxbf-pmc support for BlueField-3 Status in linux-bluefield package in Ubuntu: New Bug description: SRU Justification: [Impact] Currently, the mlxbf-pmc driver supports only BlueField-1 and BlueField-2 devices. BlueField-3 is not supported. [Fix] Add support for BlueField-3 devices by adding new access mechanism for programming and reading counter values as in the new design, along with the list of supported events in each HW block in BlueField-3. [Test Case] Previously, the driver would not be loaded when the OS is loaded on a BlueField-3 card. With the fix, the driver should get loaded and create sysfs entries as a hwmon device under /sys/class/hwmon/hwmon0. [Regression Potential] The ACPI table needs to be in sync with this change since it depends on newly introduces parameters such as mss_enable and llt_enable, which were not present earlier. To manage notifications about this bug go to: https://bugs.launchpad.net/ubuntu/+source/linux-bluefield/+bug/2002500/+subscriptions -- Mailing list: https://launchpad.net/~kernel-packages Post to : kernel-packages@lists.launchpad.net Unsubscribe : https://launchpad.net/~kernel-packages More help : https://help.launchpad.net/ListHelp
[Kernel-packages] [Bug 2002501] [NEW] mlxbf-pmc support for BlueField-3
Public bug reported: SRU Justification: [Impact] Currently, the mlxbf-pmc driver supports only BlueField-1 and BlueField-2 devices. BlueField-3 is not supported. [Fix] Add support for BlueField-3 devices by adding new access mechanism for programming and reading counter values as in the new design, along with the list of supported events in each HW block in BlueField-3. [Test Case] Previously, the driver would not be loaded when the OS is loaded on a BlueField-3 card. With the fix, the driver should get loaded and create sysfs entries as a hwmon device under /sys/class/hwmon/hwmon0. [Regression Potential] The ACPI table needs to be in sync with this change since it depends on newly introduces parameters such as mss_enable and llt_enable, which were not present earlier. ** Affects: linux-bluefield (Ubuntu) Importance: Undecided Status: New -- You received this bug notification because you are a member of Kernel Packages, which is subscribed to linux-bluefield in Ubuntu. https://bugs.launchpad.net/bugs/2002501 Title: mlxbf-pmc support for BlueField-3 Status in linux-bluefield package in Ubuntu: New Bug description: SRU Justification: [Impact] Currently, the mlxbf-pmc driver supports only BlueField-1 and BlueField-2 devices. BlueField-3 is not supported. [Fix] Add support for BlueField-3 devices by adding new access mechanism for programming and reading counter values as in the new design, along with the list of supported events in each HW block in BlueField-3. [Test Case] Previously, the driver would not be loaded when the OS is loaded on a BlueField-3 card. With the fix, the driver should get loaded and create sysfs entries as a hwmon device under /sys/class/hwmon/hwmon0. [Regression Potential] The ACPI table needs to be in sync with this change since it depends on newly introduces parameters such as mss_enable and llt_enable, which were not present earlier. To manage notifications about this bug go to: https://bugs.launchpad.net/ubuntu/+source/linux-bluefield/+bug/2002501/+subscriptions -- Mailing list: https://launchpad.net/~kernel-packages Post to : kernel-packages@lists.launchpad.net Unsubscribe : https://launchpad.net/~kernel-packages More help : https://help.launchpad.net/ListHelp
[Kernel-packages] [Bug 1989172] Re: mlxbf-pmc: error when reading unprogrammed events
** Tags removed: verification-needed-jammy ** Tags added: verification-done-jammy -- You received this bug notification because you are a member of Kernel Packages, which is subscribed to linux-bluefield in Ubuntu. https://bugs.launchpad.net/bugs/1989172 Title: mlxbf-pmc: error when reading unprogrammed events Status in linux-bluefield package in Ubuntu: Invalid Status in linux-bluefield source package in Focal: Fix Released Status in linux-bluefield source package in Jammy: Fix Committed Bug description: SRU Justification: [Impact] When the counters are not programmed to monitor any events, reading the "event" file results in the following error: cat: /sys/class/hwmon/hwmon0/smmu0/event0: Invalid argument This is misleading and instead needs to indicate that the counter is currently not enabled. [Fix] * First, skip the check for whether the counter is enabled when reading the event info. If the counter is not enabled, it should return 0, which is the default value when coming out of reset. * 0 is not a valid event in most blocks as per the respective event lists, so add a "Disable" event to all the event lists. [Test Case] Without the fix, reading any "event" file right after reset when the counters are not programmed should result in the following error: cat: /sys/class/hwmon/hwmon0/smmu0/event0: Invalid argumen With the fix, it should instead print: 0x0: DISABLE [Regression Potential] Can be considered minimal To manage notifications about this bug go to: https://bugs.launchpad.net/ubuntu/+source/linux-bluefield/+bug/1989172/+subscriptions -- Mailing list: https://launchpad.net/~kernel-packages Post to : kernel-packages@lists.launchpad.net Unsubscribe : https://launchpad.net/~kernel-packages More help : https://help.launchpad.net/ListHelp
[Kernel-packages] [Bug 1995109] Re: bluefield-edac: Potentially overflowing expression
** Tags removed: verification-needed-focal verification-needed-jammy ** Tags added: verification-done-focal verification-done-jammy -- You received this bug notification because you are a member of Kernel Packages, which is subscribed to linux-bluefield in Ubuntu. https://bugs.launchpad.net/bugs/1995109 Title: bluefield-edac: Potentially overflowing expression Status in linux-bluefield package in Ubuntu: Invalid Status in linux-bluefield source package in Focal: Fix Committed Status in linux-bluefield source package in Jammy: Fix Committed Bug description: SRU Justification: [Impact] Potential integer overflow in men_ctrl_idx detected by static tool analyser. [Fix] Declare as type u64 instead of type int [Test Case] NA since the variable is read from the ACPI table [Regression Potential] Can be considered minimal To manage notifications about this bug go to: https://bugs.launchpad.net/ubuntu/+source/linux-bluefield/+bug/1995109/+subscriptions -- Mailing list: https://launchpad.net/~kernel-packages Post to : kernel-packages@lists.launchpad.net Unsubscribe : https://launchpad.net/~kernel-packages More help : https://help.launchpad.net/ListHelp
[Kernel-packages] [Bug 1995109] [NEW] bluefield-edac: Potentially overflowing expression
Public bug reported: SRU Justification: [Impact] Potential integer overflow in men_ctrl_idx detected by static tool analyser. [Fix] Declare as type u64 instead of type int [Test Case] NA since the variable is read from the ACPI table [Regression Potential] Can be considered minimal ** Affects: linux-bluefield (Ubuntu) Importance: Undecided Status: New -- You received this bug notification because you are a member of Kernel Packages, which is subscribed to linux-bluefield in Ubuntu. https://bugs.launchpad.net/bugs/1995109 Title: bluefield-edac: Potentially overflowing expression Status in linux-bluefield package in Ubuntu: New Bug description: SRU Justification: [Impact] Potential integer overflow in men_ctrl_idx detected by static tool analyser. [Fix] Declare as type u64 instead of type int [Test Case] NA since the variable is read from the ACPI table [Regression Potential] Can be considered minimal To manage notifications about this bug go to: https://bugs.launchpad.net/ubuntu/+source/linux-bluefield/+bug/1995109/+subscriptions -- Mailing list: https://launchpad.net/~kernel-packages Post to : kernel-packages@lists.launchpad.net Unsubscribe : https://launchpad.net/~kernel-packages More help : https://help.launchpad.net/ListHelp
[Kernel-packages] [Bug 1989172] Re: mlxbf-pmc: error when reading unprogrammed events
** Tags added: verification-done-jammy -- You received this bug notification because you are a member of Kernel Packages, which is subscribed to linux-bluefield in Ubuntu. https://bugs.launchpad.net/bugs/1989172 Title: mlxbf-pmc: error when reading unprogrammed events Status in linux-bluefield package in Ubuntu: Invalid Status in linux-bluefield source package in Focal: Fix Committed Status in linux-bluefield source package in Jammy: Fix Committed Bug description: SRU Justification: [Impact] When the counters are not programmed to monitor any events, reading the "event" file results in the following error: cat: /sys/class/hwmon/hwmon0/smmu0/event0: Invalid argument This is misleading and instead needs to indicate that the counter is currently not enabled. [Fix] * First, skip the check for whether the counter is enabled when reading the event info. If the counter is not enabled, it should return 0, which is the default value when coming out of reset. * 0 is not a valid event in most blocks as per the respective event lists, so add a "Disable" event to all the event lists. [Test Case] Without the fix, reading any "event" file right after reset when the counters are not programmed should result in the following error: cat: /sys/class/hwmon/hwmon0/smmu0/event0: Invalid argumen With the fix, it should instead print: 0x0: DISABLE [Regression Potential] Can be considered minimal To manage notifications about this bug go to: https://bugs.launchpad.net/ubuntu/+source/linux-bluefield/+bug/1989172/+subscriptions -- Mailing list: https://launchpad.net/~kernel-packages Post to : kernel-packages@lists.launchpad.net Unsubscribe : https://launchpad.net/~kernel-packages More help : https://help.launchpad.net/ListHelp
[Kernel-packages] [Bug 1989172] Re: mlxbf-pmc: error when reading unprogrammed events
** Tags added: verification-done-focal -- You received this bug notification because you are a member of Kernel Packages, which is subscribed to linux-bluefield in Ubuntu. https://bugs.launchpad.net/bugs/1989172 Title: mlxbf-pmc: error when reading unprogrammed events Status in linux-bluefield package in Ubuntu: Invalid Status in linux-bluefield source package in Focal: Fix Committed Bug description: SRU Justification: [Impact] When the counters are not programmed to monitor any events, reading the "event" file results in the following error: cat: /sys/class/hwmon/hwmon0/smmu0/event0: Invalid argument This is misleading and instead needs to indicate that the counter is currently not enabled. [Fix] * First, skip the check for whether the counter is enabled when reading the event info. If the counter is not enabled, it should return 0, which is the default value when coming out of reset. * 0 is not a valid event in most blocks as per the respective event lists, so add a "Disable" event to all the event lists. [Test Case] Without the fix, reading any "event" file right after reset when the counters are not programmed should result in the following error: cat: /sys/class/hwmon/hwmon0/smmu0/event0: Invalid argumen With the fix, it should instead print: 0x0: DISABLE [Regression Potential] Can be considered minimal To manage notifications about this bug go to: https://bugs.launchpad.net/ubuntu/+source/linux-bluefield/+bug/1989172/+subscriptions -- Mailing list: https://launchpad.net/~kernel-packages Post to : kernel-packages@lists.launchpad.net Unsubscribe : https://launchpad.net/~kernel-packages More help : https://help.launchpad.net/ListHelp
[Kernel-packages] [Bug 1989172] [NEW] mlxbf-pmc: error when reading unprogrammed events
Public bug reported: SRU Justification: [Impact] When the counters are not programmed to monitor any events, reading the "event" file results in the following error: cat: /sys/class/hwmon/hwmon0/smmu0/event0: Invalid argument This is misleading and instead needs to indicate that the counter is currently not enabled. [Fix] * First, skip the check for whether the counter is enabled when reading the event info. If the counter is not enabled, it should return 0, which is the default value when coming out of reset. * 0 is not a valid event in most blocks as per the respective event lists, so add a "Disable" event to all the event lists. [Test Case] Without the fix, reading any "event" file right after reset when the counters are not programmed should result in the following error: cat: /sys/class/hwmon/hwmon0/smmu0/event0: Invalid argumen With the fix, it should instead print: 0x0: DISABLE [Regression Potential] Can be considered minimal ** Affects: linux-bluefield (Ubuntu) Importance: Undecided Status: New -- You received this bug notification because you are a member of Kernel Packages, which is subscribed to linux-bluefield in Ubuntu. https://bugs.launchpad.net/bugs/1989172 Title: mlxbf-pmc: error when reading unprogrammed events Status in linux-bluefield package in Ubuntu: New Bug description: SRU Justification: [Impact] When the counters are not programmed to monitor any events, reading the "event" file results in the following error: cat: /sys/class/hwmon/hwmon0/smmu0/event0: Invalid argument This is misleading and instead needs to indicate that the counter is currently not enabled. [Fix] * First, skip the check for whether the counter is enabled when reading the event info. If the counter is not enabled, it should return 0, which is the default value when coming out of reset. * 0 is not a valid event in most blocks as per the respective event lists, so add a "Disable" event to all the event lists. [Test Case] Without the fix, reading any "event" file right after reset when the counters are not programmed should result in the following error: cat: /sys/class/hwmon/hwmon0/smmu0/event0: Invalid argumen With the fix, it should instead print: 0x0: DISABLE [Regression Potential] Can be considered minimal To manage notifications about this bug go to: https://bugs.launchpad.net/ubuntu/+source/linux-bluefield/+bug/1989172/+subscriptions -- Mailing list: https://launchpad.net/~kernel-packages Post to : kernel-packages@lists.launchpad.net Unsubscribe : https://launchpad.net/~kernel-packages More help : https://help.launchpad.net/ListHelp
[Kernel-packages] [Bug 1979045] Re: misleading error prints from mlx-trio
** Tags removed: verification-needed-focal ** Tags added: verification-done-focal -- You received this bug notification because you are a member of Kernel Packages, which is subscribed to linux-bluefield in Ubuntu. https://bugs.launchpad.net/bugs/1979045 Title: misleading error prints from mlx-trio Status in linux-bluefield package in Ubuntu: Invalid Status in linux-bluefield source package in Focal: Fix Committed Bug description: SRU Justification: [Impact] TRIO IRQs are triggered during normal operation and non-error events such as mlxfwreset and the prints from the handler are misleading to users in such instances. [Fix] Replace the dev_err calls with pr_debug. This will ensure that users do not see extraneous error messages from the driver. [Test Case] Running mlxfwreset with the trio driver loaded will see a lot of error prints and this patch will fix the same. [Regression Potential] Can be considered minimal. To manage notifications about this bug go to: https://bugs.launchpad.net/ubuntu/+source/linux-bluefield/+bug/1979045/+subscriptions -- Mailing list: https://launchpad.net/~kernel-packages Post to : kernel-packages@lists.launchpad.net Unsubscribe : https://launchpad.net/~kernel-packages More help : https://help.launchpad.net/ListHelp
[Kernel-packages] [Bug 1979264] Re: mlx-trio module license update
** Tags removed: verification-needed-focal ** Tags added: verification-done-focal -- You received this bug notification because you are a member of Kernel Packages, which is subscribed to linux-bluefield in Ubuntu. https://bugs.launchpad.net/bugs/1979264 Title: mlx-trio module license update Status in linux-bluefield package in Ubuntu: Fix Committed Bug description: SRU Justification: [Impact] mlx-trio module license info is currently only GPL and needs to be updated similar to other Mellanox platform drivers [Fix] Change module license from GPL to Dual BSD/GPL [Test Case] modinfo prints the updated license info [Regression Potential] Can be considered minimal To manage notifications about this bug go to: https://bugs.launchpad.net/ubuntu/+source/linux-bluefield/+bug/1979264/+subscriptions -- Mailing list: https://launchpad.net/~kernel-packages Post to : kernel-packages@lists.launchpad.net Unsubscribe : https://launchpad.net/~kernel-packages More help : https://help.launchpad.net/ListHelp
[Kernel-packages] [Bug 1979045] [NEW] misleading error prints from mlx-trio
Public bug reported: SRU Justification: [Impact] TRIO IRQs are triggered during normal operation and non-error events such as mlxfwreset and the prints from the handler are misleading to users in such instances. [Fix] Replace the dev_err calls with pr_debug. This will ensure that users do not see extraneous error messages from the driver. [Test Case] Running mlxfwreset with the trio driver loaded will see a lot of error prints and this patch will fix the same. [Regression Potential] Can be considered minimal. ** Affects: linux-bluefield (Ubuntu) Importance: Undecided Status: New -- You received this bug notification because you are a member of Kernel Packages, which is subscribed to linux-bluefield in Ubuntu. https://bugs.launchpad.net/bugs/1979045 Title: misleading error prints from mlx-trio Status in linux-bluefield package in Ubuntu: New Bug description: SRU Justification: [Impact] TRIO IRQs are triggered during normal operation and non-error events such as mlxfwreset and the prints from the handler are misleading to users in such instances. [Fix] Replace the dev_err calls with pr_debug. This will ensure that users do not see extraneous error messages from the driver. [Test Case] Running mlxfwreset with the trio driver loaded will see a lot of error prints and this patch will fix the same. [Regression Potential] Can be considered minimal. To manage notifications about this bug go to: https://bugs.launchpad.net/ubuntu/+source/linux-bluefield/+bug/1979045/+subscriptions -- Mailing list: https://launchpad.net/~kernel-packages Post to : kernel-packages@lists.launchpad.net Unsubscribe : https://launchpad.net/~kernel-packages More help : https://help.launchpad.net/ListHelp
[Kernel-packages] [Bug 1948436] Re: Add RevID field to VPD info in EEPROM
** Tags removed: verification-needed-focal ** Tags added: verification-done-focal -- You received this bug notification because you are a member of Kernel Packages, which is subscribed to linux-bluefield in Ubuntu. https://bugs.launchpad.net/bugs/1948436 Title: Add RevID field to VPD info in EEPROM Status in linux-bluefield package in Ubuntu: Fix Committed Bug description: SRU Justification: [Impact] Currently, the VPD info stored in the EEPROM does not include a Rev ID. This field is incremented when a PN goes through minor HW changes. Sometimes, SW needs to differentiate the two versions of the card. Hence, this field should be supported in the EEPROM MFG partition. [Fix] EC (Engineering Changes) field is used to denote the HW Revision of a PN. Add it to the EEPROM MFG and provide a sysfs entry to access the same via mlx-bootctl. [Test Case] This field is typically programmed at the manufacturing step using the bfcfg tool which uses the write mechanism in the driver and then the MFG partition is locked. Reading the "rev" sysfs will show the current value. [Regression Potential] Can be considered minimum. To manage notifications about this bug go to: https://bugs.launchpad.net/ubuntu/+source/linux-bluefield/+bug/1948436/+subscriptions -- Mailing list: https://launchpad.net/~kernel-packages Post to : kernel-packages@lists.launchpad.net Unsubscribe : https://launchpad.net/~kernel-packages More help : https://help.launchpad.net/ListHelp
[Kernel-packages] [Bug 1948434] Re: Check if secure boot is enabled with development keys
** Tags removed: verification-needed-focal ** Tags added: verification-done-focal -- You received this bug notification because you are a member of Kernel Packages, which is subscribed to linux-bluefield in Ubuntu. https://bugs.launchpad.net/bugs/1948434 Title: Check if secure boot is enabled with development keys Status in linux-bluefield package in Ubuntu: Invalid Status in linux-bluefield source package in Focal: In Progress Bug description: SRU Justification: [Impact] Currently, there is no indication from mlx-bootctl when the user reads the lifecycle_state sysfs as to whether secure boot is enabled with development keys or production keys. In order to make this clear to the user, add a check in the driver. [Fix] Check the secure boot development mode status bit. If secure boot is enabled with the development key, then print it to the output buffer when lifecycle_state_show() is invoked. [Test Case] On a system in secure state, if it has been programmed with development keys, then reading the lifecycle_state sysfs entry in the mlx-bootctl driver should print a message that indicates the same. Similarly, a secure system which has been programmed with production keys must print the appropriate message when the lifecycle_state sysfs is read. [Regression Potential] Can be considered minimum. To manage notifications about this bug go to: https://bugs.launchpad.net/ubuntu/+source/linux-bluefield/+bug/1948434/+subscriptions -- Mailing list: https://launchpad.net/~kernel-packages Post to : kernel-packages@lists.launchpad.net Unsubscribe : https://launchpad.net/~kernel-packages More help : https://help.launchpad.net/ListHelp
[Kernel-packages] [Bug 1948436] [NEW] Add RevID field to VPD info in EEPROM
Public bug reported: SRU Justification: [Impact] Currently, the VPD info stored in the EEPROM does not include a Rev ID. This field is incremented when a PN goes through minor HW changes. Sometimes, SW needs to differentiate the two versions of the card. Hence, this field should be supported in the EEPROM MFG partition. [Fix] EC (Engineering Changes) field is used to denote the HW Revision of a PN. Add it to the EEPROM MFG and provide a sysfs entry to access the same via mlx-bootctl. [Test Case] This field is typically programmed at the manufacturing step using the bfcfg tool which uses the write mechanism in the driver and then the MFG partition is locked. Reading the "rev" sysfs will show the current value. [Regression Potential] Can be considered minimum. ** Affects: linux-bluefield (Ubuntu) Importance: Undecided Status: New -- You received this bug notification because you are a member of Kernel Packages, which is subscribed to linux-bluefield in Ubuntu. https://bugs.launchpad.net/bugs/1948436 Title: Add RevID field to VPD info in EEPROM Status in linux-bluefield package in Ubuntu: New Bug description: SRU Justification: [Impact] Currently, the VPD info stored in the EEPROM does not include a Rev ID. This field is incremented when a PN goes through minor HW changes. Sometimes, SW needs to differentiate the two versions of the card. Hence, this field should be supported in the EEPROM MFG partition. [Fix] EC (Engineering Changes) field is used to denote the HW Revision of a PN. Add it to the EEPROM MFG and provide a sysfs entry to access the same via mlx-bootctl. [Test Case] This field is typically programmed at the manufacturing step using the bfcfg tool which uses the write mechanism in the driver and then the MFG partition is locked. Reading the "rev" sysfs will show the current value. [Regression Potential] Can be considered minimum. To manage notifications about this bug go to: https://bugs.launchpad.net/ubuntu/+source/linux-bluefield/+bug/1948436/+subscriptions -- Mailing list: https://launchpad.net/~kernel-packages Post to : kernel-packages@lists.launchpad.net Unsubscribe : https://launchpad.net/~kernel-packages More help : https://help.launchpad.net/ListHelp
[Kernel-packages] [Bug 1948434] [NEW] Check if secure boot is enabled with development keys
Public bug reported: SRU Justification: [Impact] Currently, there is no indication from mlx-bootctl when the user reads the lifecycle_state sysfs as to whether secure boot is enabled with development keys or production keys. In order to make this clear to the user, add a check in the driver. [Fix] Check the secure boot development mode status bit. If secure boot is enabled with the development key, then print it to the output buffer when lifecycle_state_show() is invoked. [Test Case] On a system in secure state, if it has been programmed with development keys, then reading the lifecycle_state sysfs entry in the mlx-bootctl driver should print a message that indicates the same. Similarly, a secure system which has been programmed with production keys must print the appropriate message when the lifecycle_state sysfs is read. [Regression Potential] Can be considered minimum. ** Affects: linux-bluefield (Ubuntu) Importance: Undecided Status: New -- You received this bug notification because you are a member of Kernel Packages, which is subscribed to linux-bluefield in Ubuntu. https://bugs.launchpad.net/bugs/1948434 Title: Check if secure boot is enabled with development keys Status in linux-bluefield package in Ubuntu: New Bug description: SRU Justification: [Impact] Currently, there is no indication from mlx-bootctl when the user reads the lifecycle_state sysfs as to whether secure boot is enabled with development keys or production keys. In order to make this clear to the user, add a check in the driver. [Fix] Check the secure boot development mode status bit. If secure boot is enabled with the development key, then print it to the output buffer when lifecycle_state_show() is invoked. [Test Case] On a system in secure state, if it has been programmed with development keys, then reading the lifecycle_state sysfs entry in the mlx-bootctl driver should print a message that indicates the same. Similarly, a secure system which has been programmed with production keys must print the appropriate message when the lifecycle_state sysfs is read. [Regression Potential] Can be considered minimum. To manage notifications about this bug go to: https://bugs.launchpad.net/ubuntu/+source/linux-bluefield/+bug/1948434/+subscriptions -- Mailing list: https://launchpad.net/~kernel-packages Post to : kernel-packages@lists.launchpad.net Unsubscribe : https://launchpad.net/~kernel-packages More help : https://help.launchpad.net/ListHelp
[Kernel-packages] [Bug 1931843] Re: Update mlx-bootctl to access new fields in EEPROM MFG
** Tags removed: verification-needed-focal ** Tags added: verification-done-focal -- You received this bug notification because you are a member of Kernel Packages, which is subscribed to linux-bluefield in Ubuntu. https://bugs.launchpad.net/bugs/1931843 Title: Update mlx-bootctl to access new fields in EEPROM MFG Status in linux-bluefield package in Ubuntu: Invalid Status in linux-bluefield source package in Focal: Fix Committed Bug description: SRU Justification: [Impact] The EEPROM MFG partition on BlueField-2 has been updated to include the VPD information for each card. In order to access these newly added fields, the mlx-bootctl driver needs to be updated to provide an access mechanism. [Fix] Add support for VPD fields in the EEPROM MFG and provide access to these via sysfs entries. The newly added sysfs entries are: sku (SKU ID), modl (Model Number), sn (Serial Number) and uuid (UUID). And the previously added opn_str sysfs has been renamed to opn. [Test Case] Though the driver provides read and write access through sysfs, the contents of the MFG partition are written during Manufacturing and then locked in order to protect the info. Writing to this partition will therefore require resetting the MFG info from the UEFI Device Manager, which will unlock the partition and allow for it to be reprogrammed. Reading the sysfs entries will print the contents of each field. It could also be empty if the field was not programmed earlier. [Regression Potential] Can be considered minimum, since the new fields have been added without interfering with the existing fields which might already be present in the EEPROM. To manage notifications about this bug go to: https://bugs.launchpad.net/ubuntu/+source/linux-bluefield/+bug/1931843/+subscriptions -- Mailing list: https://launchpad.net/~kernel-packages Post to : kernel-packages@lists.launchpad.net Unsubscribe : https://launchpad.net/~kernel-packages More help : https://help.launchpad.net/ListHelp
[Kernel-packages] [Bug 1931981] Re: mlx-bootctl: Fix potential buffer overflow
*** This bug is a duplicate of bug 1931843 *** https://bugs.launchpad.net/bugs/1931843 ** This bug has been marked a duplicate of bug 1931843 Update mlx-bootctl to access new fields in EEPROM MFG -- You received this bug notification because you are a member of Kernel Packages, which is subscribed to linux-bluefield in Ubuntu. https://bugs.launchpad.net/bugs/1931981 Title: mlx-bootctl: Fix potential buffer overflow Status in linux-bluefield package in Ubuntu: New Bug description: SRU Justification: [Impact] The sysfs store/show functions use sprintf without specifying a size which could lead to potential buffer overflow. [Fix] Replace sprintf with snprintf to avoid buffer overflow. Also, remove the redundant strlen usage since count is already available in the _store functions. [Test Plan] Read/write access to the EEPROM MFG fields can be tested via the sysfs entries that are exposed by the driver. Please note that the MFG partition is locked in order to protect the data and this could block all writes to it. In order to enable writes to the EEPROM, the MFG Info needs to be reset via the UEFI Device Manager. [Regression Potential] Can be considered minimum. To manage notifications about this bug go to: https://bugs.launchpad.net/ubuntu/+source/linux-bluefield/+bug/1931981/+subscriptions -- Mailing list: https://launchpad.net/~kernel-packages Post to : kernel-packages@lists.launchpad.net Unsubscribe : https://launchpad.net/~kernel-packages More help : https://help.launchpad.net/ListHelp
[Kernel-packages] [Bug 1931984] Re: mlx-bootctl: Update license and version info
*** This bug is a duplicate of bug 1931843 *** https://bugs.launchpad.net/bugs/1931843 ** This bug has been marked a duplicate of bug 1931843 Update mlx-bootctl to access new fields in EEPROM MFG -- You received this bug notification because you are a member of Kernel Packages, which is subscribed to linux-bluefield in Ubuntu. https://bugs.launchpad.net/bugs/1931984 Title: mlx-bootctl: Update license and version info Status in linux-bluefield package in Ubuntu: New Bug description: SRU Justification: [Impact] License info needs to be updated since the current info is no longer accurate. Driver version needs to be incremented since new features have been added. [Fix] Update license info to "Dual BSD/GPL". Increment version to 1.4 [Test Plan] Verify version change [Regression Potential] None To manage notifications about this bug go to: https://bugs.launchpad.net/ubuntu/+source/linux-bluefield/+bug/1931984/+subscriptions -- Mailing list: https://launchpad.net/~kernel-packages Post to : kernel-packages@lists.launchpad.net Unsubscribe : https://launchpad.net/~kernel-packages More help : https://help.launchpad.net/ListHelp
[Kernel-packages] [Bug 1933642] Re: mlx-bootctl: Fix exclusion issues around arm_smccc_smc
*** This bug is a duplicate of bug 1931843 *** https://bugs.launchpad.net/bugs/1931843 ** This bug has been marked a duplicate of bug 1931843 Update mlx-bootctl to access new fields in EEPROM MFG -- You received this bug notification because you are a member of Kernel Packages, which is subscribed to linux-bluefield in Ubuntu. https://bugs.launchpad.net/bugs/1933642 Title: mlx-bootctl: Fix exclusion issues around arm_smccc_smc Status in linux-bluefield package in Ubuntu: New Bug description: SRU Justification: [Impact] Exclusion is being handled around arm_smccc_smc() only in the store functions. It should be implemented in the show functions also as this call might not be thread-safe. [Fix] Add mutex_lock/unlock around arm_smccc_smc calls in the show functions. [Test Case] Multiple simultaneous accesses to a sysfs entry exposed by the driver are successful. [Regression Potential] Can be considered minimum To manage notifications about this bug go to: https://bugs.launchpad.net/ubuntu/+source/linux-bluefield/+bug/1933642/+subscriptions -- Mailing list: https://launchpad.net/~kernel-packages Post to : kernel-packages@lists.launchpad.net Unsubscribe : https://launchpad.net/~kernel-packages More help : https://help.launchpad.net/ListHelp
[Kernel-packages] [Bug 1933642] [NEW] mlx-bootctl: Fix exclusion issues around arm_smccc_smc
Public bug reported: SRU Justification: [Impact] Exclusion is being handled around arm_smccc_smc() only in the store functions. It should be implemented in the show functions also as this call might not be thread-safe. [Fix] Add mutex_lock/unlock around arm_smccc_smc calls in the show functions. [Test Case] Multiple simultaneous accesses to a sysfs entry exposed by the driver are successful. [Regression Potential] Can be considered minimum ** Affects: linux-bluefield (Ubuntu) Importance: Undecided Status: New -- You received this bug notification because you are a member of Kernel Packages, which is subscribed to linux-bluefield in Ubuntu. https://bugs.launchpad.net/bugs/1933642 Title: mlx-bootctl: Fix exclusion issues around arm_smccc_smc Status in linux-bluefield package in Ubuntu: New Bug description: SRU Justification: [Impact] Exclusion is being handled around arm_smccc_smc() only in the store functions. It should be implemented in the show functions also as this call might not be thread-safe. [Fix] Add mutex_lock/unlock around arm_smccc_smc calls in the show functions. [Test Case] Multiple simultaneous accesses to a sysfs entry exposed by the driver are successful. [Regression Potential] Can be considered minimum To manage notifications about this bug go to: https://bugs.launchpad.net/ubuntu/+source/linux-bluefield/+bug/1933642/+subscriptions -- Mailing list: https://launchpad.net/~kernel-packages Post to : kernel-packages@lists.launchpad.net Unsubscribe : https://launchpad.net/~kernel-packages More help : https://help.launchpad.net/ListHelp