|
From: | David Hildenbrand |
Subject: | Re: [PATCH v1 3/3] util/oslib-posix: Support concurrent os_mem_prealloc() invocation |
Date: | Tue, 20 Jul 2021 16:35:59 +0200 |
User-agent: | Mozilla/5.0 (X11; Linux x86_64; rv:78.0) Gecko/20100101 Thunderbird/78.11.0 |
Of course with the madvise impl we're bypassing the SIGBUS dance entirely. This is good for people with new kernels, but is this SIGBUS stuff safe for older kernels ?It remains broken with old kernels I guess. There isn't too much that we can do: disabling prealloc=on once the VM is running breaks existing use cases.Ok, while refactoring this, could you add a scary warning next to the sigaction calls mentioning that this code is not likely to play well with qemu's other handling of sigbus, as a reminder to future reviewers.
Sure thing!
Fortunately, running into that race seems to be rare, at least I never hear reports.The failure mode is likely to be silent or easily mis-interpreted Is there any value in emitting a one-time per process warning message on stderr if we take the old codepath post-startup ?
Will look into emitting a warning when running this code while the VM is already running. (hope it won't be too ugly to have vm state checks in util/oslib-posix).
-- Thanks, David / dhildenb
[Prev in Thread] | Current Thread | [Next in Thread] |