Hi Zenghui,
On Tue, 3 Jun 2025, Zenghui Yu wrote:
On 2025/5/27 22:24, Sebastian Ott wrote:
Some small fixes for arch_timer_edge_cases that I stumbled upon while debugging failures for this selftest on ampere-one.
Changes since v1: modified patch 3 based on suggestions from Marc.
I've done some tests with this on various machines - seems to be all good, however on ampere-one I now hit this in 10% of the runs: ==== Test Assertion Failure ==== arm64/arch_timer_edge_cases.c:481: timer_get_cntct(timer) >= DEF_CNT + (timer_get_cntfrq() * (uint64_t)(delta_2_ms) / 1000) pid=166657 tid=166657 errno=4 - Interrupted system call 1 0x0000000000404db3: test_run at arch_timer_edge_cases.c:933 2 0x0000000000401f9f: main at arch_timer_edge_cases.c:1062 3 0x0000ffffaedd625b: ?? ??:0 4 0x0000ffffaedd633b: ?? ??:0 5 0x00000000004020af: _start at ??:? timer_get_cntct(timer) >= DEF_CNT + msec_to_cycles(delta_2_ms)
This is not new, it was just hidden behind the other failure. I'll try to figure out what this is about (seems to be independent of the wait time)..
Not sure if you have figured it out. I can easily reproduce it on my box and I *guess* it is that we have some random XVAL values when we enable the timer..
Yes, I think so, too.
test_reprogramming_timer() { local_irq_disable(); reset_timer_state(timer, DEF_CNT);
My first attempt was to also initialize cval here
/* Program the timer to DEF_CNT + delta_1_ms. */ set_tval_irq(timer, msec_to_cycles(delta_1_ms), CTL_ENABLE);
[...] }
set_tval_irq() { timer_set_ctl(timer, ctl);
// There is a window that we enable the timer with *random* XVAL // values and we may get the unexpected interrupt.. And it's // unlikely that KVM can be aware of TVAL's change (and // re-evaluate the interrupt's pending state) before hitting the // GUEST_ASSERT().
timer_set_tval(timer, tval_cycles);
Yes, I stumbled over this as well. I've always assumed that this order is becauase of this from the architecture "If CNTV_CTL_EL0.ENABLE is 0, the value returned is UNKNOWN." However re-reading that part today I realized that this only concerns register reads.
Maybe somone on cc knows why it's in that order?
I'm currently testing this with the above swapped and it's looking good, so far.
}
I'm not familiar with the test so I'm not 100% sure that this is the root cause. But I hope this helps with your analysis ;-) .
It did, thanks!
Sebastian