From 59f6ebbc49fc8914f4315c8341c603b2e3226046 Mon Sep 17 00:00:00 2001 From: intrigeri <intrigeri@riseup.net> Date: Thu, 16 Jan 2025 15:49:06 +0000 Subject: [PATCH] Test suite: Bump the time we wait before checking memory poisoning, again Waiting 3 seconds instead of 1, as we did in !1630, helped a lot: tails/tails#20354 used to happen "often", now it's quite rare. But it still happens: I've seen "3.766% of the free memory still has the pattern, but less than 0.800% was expected". In https://gitlab.tails.boum.org/tails/tails/-/merge_requests/1630#note_240453 @anonym wrote: > 10 sounds a bit much from a "what latency we call good enough for this > feature" perspective, not from a "minimizing test suite failures" one. I think > your original 3 is worth trying. So let's try to find a middle ground. Note that as opposed to !1630, which bumped the sleep time from 1s to 3s for all the memory erasure scenarios, here I'm bumping it only for the specific scenario where I've seen a recent failure. Interestingly, while tails/tails#20354 does not say which exact scenarios were found to be affected back then, in the attached debug.log the only affected one is the scenario that I saw failing here. So it might be than 1 second worked just fine for all the other scenarios, and only this one needed a bump. I'm not going to dive into this deeper as 3 seconds seems to be a reasonable baseline. For more context, see the full discussion around the aforementioned comment. --- features/erase_memory.feature | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/features/erase_memory.feature b/features/erase_memory.feature index 61eb2d6a2a0..5c889256a88 100644 --- a/features/erase_memory.feature +++ b/features/erase_memory.feature @@ -16,7 +16,7 @@ Feature: System memory erasure on shutdown When I start a process allocating 128 MiB of memory with a known pattern Then patterns cover at least 128 MiB in the guest's memory When I kill the allocating process - Then I find very few patterns in the guest's memory after 3 seconds + Then I find very few patterns in the guest's memory after 5 seconds Scenario: Erasure of tmpfs data on unmount Given I have started Tails from DVD without network and logged in -- GitLab