[Mellanox] Skip the leftover hardware reboot cause in case of last boot is warm/fast reboot (sonic-net#13246)

- Why I did it
In case of warm/fast reboot, the hardware reboot cause will NOT be cleared because CPLD will not be touched in this flow. To not confuse the reboot cause determine logic, the leftover hardware reboot cause shall be skipped by the platform API, platform API will return the 'REBOOT_CAUSE_NON_HARDWARE' instead of the "hardware" reboot cause.

- How I did it
Check the proc cmdline to see whether the last reboot is a warm or fast reboot, if yes skip checking the leftover hardware reboot cause.

- How to verify it
a. Manual test:
    - Perform a power loss
    - Perform a warm/fast reboot
    - Check the reboot cause should be "warm-reboot" or "fast-reboot" instead of "power loss"
b. Run reboot cause related regression test.

Signed-off-by: Kebo Liu <[email protected]>

Author: keboliu

platform/mellanox/mlnx-platform-api/sonic_platform/chassis.py

@@ -30,6 +30,7 @@
     from .utils import extract_RJ45_ports_index
     from . import utils
     from .device_data import DeviceDataManager
+    import re
 except ImportError as e:
     raise ImportError (str(e) + "- required module not found")
 
@@ -61,6 +62,10 @@
 
 REBOOT_CAUSE_FILE_LENGTH = 1
 
+REBOOT_TYPE_KEXEC_FILE = "/proc/cmdline"
+REBOOT_TYPE_KEXEC_PATTERN_WARM = ".*SONIC_BOOT_TYPE=(warm|fastfast).*"
+REBOOT_TYPE_KEXEC_PATTERN_FAST = ".*SONIC_BOOT_TYPE=(fast|fast-reboot).*"
+
 # Global logger class instance
 logger = Logger()
 
@@ -736,6 +741,18 @@ def initialize_reboot_cause(self):
         self.reboot_by_software = 'reset_sw_reset'
         self.reboot_cause_initialized = True
 
+    def _parse_warmfast_reboot_from_proc_cmdline(self):
+        if os.path.isfile(REBOOT_TYPE_KEXEC_FILE):
+            with open(REBOOT_TYPE_KEXEC_FILE) as cause_file:
+                cause_file_kexec = cause_file.readline()
+            m = re.search(REBOOT_TYPE_KEXEC_PATTERN_WARM, cause_file_kexec)
+            if m and m.group(1):
+                return 'warm-reboot'
+            m = re.search(REBOOT_TYPE_KEXEC_PATTERN_FAST, cause_file_kexec)
+            if m and m.group(1):
+                return 'fast-reboot'
+        return None
+
     def get_reboot_cause(self):
         """
         Retrieves the cause of the previous reboot
@@ -748,6 +765,14 @@ def get_reboot_cause(self):
             to pass a description of the reboot cause.
         """
         #read reboot causes files in the following order
+
+        # To avoid the leftover hardware reboot cause confusing the reboot cause determine service
+        # Skip the hardware reboot cause check if warm/fast reboot cause found from cmdline
+        if utils.is_host():
+            reboot_cause = self._parse_warmfast_reboot_from_proc_cmdline()
+            if reboot_cause:
+                return self.REBOOT_CAUSE_NON_HARDWARE, ''
+
         if not self.reboot_cause_initialized:
             self.initialize_reboot_cause()
 

platform/mellanox/mlnx-platform-api/tests/test_chassis.py

@@ -224,6 +224,35 @@ def read_int_from_file(file_path, *args, **kwargs):
             assert minor == value
             mock_file_content[file_path] = 0
 
+        utils.is_host = mock.MagicMock(return_value=True)
+        chassis._parse_warmfast_reboot_from_proc_cmdline = mock.MagicMock(return_value='warm-reboot')
+        for key, value in chassis.reboot_major_cause_dict.items():
+            file_path = os.path.join(REBOOT_CAUSE_ROOT, key)
+            mock_file_content[file_path] = 1
+            major, minor = chassis.get_reboot_cause()
+            assert major == chassis.REBOOT_CAUSE_NON_HARDWARE
+            assert minor == ''
+            mock_file_content[file_path] = 0
+
+        for key, value in chassis.reboot_minor_cause_dict.items():
+            file_path = os.path.join(REBOOT_CAUSE_ROOT, key)
+            mock_file_content[file_path] = 1
+            major, minor = chassis.get_reboot_cause()
+            assert major == chassis.REBOOT_CAUSE_NON_HARDWARE
+            assert minor == value
+            mock_file_content[file_path] = 0
+
+    def test_parse_warmfast_reboot_from_proc_cmdline(self):
+        chassis = Chassis()
+        with mock.patch("builtins.open", mock.mock_open(read_data="SONIC_BOOT_TYPE=warm")):
+            assert chassis._parse_warmfast_reboot_from_proc_cmdline() == "warm-reboot"
+
+        with mock.patch("builtins.open", mock.mock_open(read_data="SONIC_BOOT_TYPE=fast")):
+            assert chassis._parse_warmfast_reboot_from_proc_cmdline() == "fast-reboot"
+
+        with mock.patch("builtins.open", mock.mock_open(read_data="SONIC_BOOT_TYPE=None")):
+            assert chassis._parse_warmfast_reboot_from_proc_cmdline() == None
+
     def test_module(self):
         from sonic_platform.chassis import ModularChassis
         # Test get_num_modules, it should not create any SFP objects