clock: the TSC keeps CPUID 15H's rate or one the HPET refines; a step the RTC confirms is slewed, a second demotes it by Japabu · Pull Request #1104 · ToyOSOrg/ToyOS

clock: the TSC keeps CPUID 15H's rate or one the HPET refines, and every CPU holds it to the HPET

Japabu

and

claude

committed

80edb7f

clock: the HPET's drift is the spec's two figures, and a runtime sample is one read

Japabu

and

claude

committed

d1f10ef

clock: one CPU samples the HPET, each CPU checks itself, and a span the package may have slept in is left out

Japabu

and

claude

committed

7d5b011

clock: the ACPI PM timer tells which of the TSC and the HPET moved

Japabu

and

claude

committed

0fc019a

clock: the PM timer test's literals in even groups

Japabu

and

claude

committed

471e0d8

clock: the RTC tells a TSC step from a chipset stall; the first is slewed away, a second demotes the TSC

Japabu

and

claude

committed

1a65ddc

review r2: a reading takes the counter inside the line's section, a demotion hands over monotonically, and the sampler goes on after it

Japabu

and

claude

committed

fec4998

tests: the clock rows print a span's left-out record, not every report's count of them

Japabu

and

claude

committed

d739b07

review r3: a demotion hands over at every CPU's clock, a drift base never reaches across a step, and an AP's proven offset demotes

Japabu

and

claude

committed

f5a102e

review r4: the handover's wait answers every round, a demotion before the release floors alone, and floors take the replaced line too

Japabu

and

claude

committed

0650c9d

review r5: the handover's comments and record say what it covers

Japabu

and

claude

committed

6cc3e1c

Navigation Menu

Repository navigation

clock: the TSC keeps CPUID 15H's rate or one the HPET refines; a step the RTC confirms is slewed, a second demotes it - #1104

clock: the TSC keeps CPUID 15H's rate or one the HPET refines; a step the RTC confirms is slewed, a second demotes it#1104
Japabu merged 11 commits into
mainfrom
system-clock

Commits on Nov 1, 2026

clock: the TSC keeps CPUID 15H's rate or one the HPET refines, and every CPU holds it to the HPET

clock: the HPET's drift is the spec's two figures, and a runtime sample is one read

clock: one CPU samples the HPET, each CPU checks itself, and a span the package may have slept in is left out

clock: the ACPI PM timer tells which of the TSC and the HPET moved

clock: the PM timer test's literals in even groups

clock: the RTC tells a TSC step from a chipset stall; the first is slewed away, a second demotes the TSC

review r2: a reading takes the counter inside the line's section, a demotion hands over monotonically, and the sampler goes on after it

tests: the clock rows print a span's left-out record, not every report's count of them

review r3: a demotion hands over at every CPU's clock, a drift base never reaches across a step, and an AP's proven offset demotes

review r4: the handover's wait answers every round, a demotion before the release floors alone, and floors take the replaced line too

review r5: the handover's comments and record say what it covers