Documentation ¶
Overview ¶
Package kernel contains local Tast tests that exercise the Linux kernel's configuration and behavior.
Index ¶
- func Bloat(ctx context.Context, s *testing.State)
- func CPUCgroup(ctx context.Context, s *testing.State)
- func ConfigVerify(ctx context.Context, s *testing.State)
- func CryptoAPI(ctx context.Context, s *testing.State)
- func CryptoDigest(ctx context.Context, s *testing.State)
- func ECDeviceNode(ctx context.Context, s *testing.State)
- func HighResTimers(ctx context.Context, s *testing.State)
- func Limits(ctx context.Context, s *testing.State)
- func PerfCallgraph(ctx context.Context, s *testing.State)
Constants ¶
This section is empty.
Variables ¶
This section is empty.
Functions ¶
func ConfigVerify ¶
ConfigVerify reads the Linux kernel version and arch to verify validity of the information returned depending on version.
func ECDeviceNode ¶
ECDeviceNode confirms that /dev/cros_ec (the device node for the primary embedded controller) exists. If this test fails in the CQ, it's likely to be related to a kernel change and not a change in the primary embedded controller's firmware, as EC firmware for the quota pool (which backs the CQ) is changed only after a candidate build goes through FAFT (which would mean the EC is likely to be behaving as a responder to I2C or SPI messages). Note that this test is similar to firmware.ECVersion, but doesn't depend on ectool or the EC reporting its version in any particular format, so a failure is likely to be associated with a kernel change. See https://chromium.googlesource.com/chromiumos/third_party/kernel/+/bd0447f72de0963129612bf18202204d5b25e133/ for an example where a change to a SPI driver prevented /dev/cros_ec from being created (ultimately discovered when flashrom invocations involving the EC failed).
func HighResTimers ¶
HighResTimers reads from /proc/timer_list to verify that any resolution listed in nsecs has a value of 1.
Types ¶
This section is empty.