Public bug reported:
On ARM, when reading from a PCI mapped address via /proc/mem/self, the
kernel crashes with the following:
[ 146.827032] [ T7089] Unable to handle kernel read from unreadable memory
at virtual address 800086828000
[ 146.836077] [ T7089] Mem abort info:
[ 146.838864
I have verified that the issue doesn't get reproduced with kernel
4.15.0-38-generic from bionic-proposed.
** Tags removed: verification-needed-bionic
** Tags added: verification-done-bionic
--
You received this bug notification because you are a member of Kernel
Packages, which is subscribed to
Some extra information: a tool to reproduce the problem is available in
the linux-block mailing list: https://www.spinics.net/lists/linux-
block/msg28507.html
--
You received this bug notification because you are a member of Kernel
Packages, which is subscribed to linux-signed in Ubuntu.
https://
Thank you for looking at this bug report.
We have some additional information.
qemu/kvm with cache=none (common) and aio=threads (default) is required
to trigger the bug. aio=native (another common option) does not trigger
it in our testing.
We applied the following patches on top of Ubuntu-4.15
Public bug reported:
TLDR: commit 72ecad22d9f198aafee64218512e02ffa7818671 (in v4.10)
introduced silent data corruption for O_DIRECT uses, it's fixed in
17d51b10d7773e4618bcac64648f30f12d4078fb (in v4.18)
A silent data corruption was introduced in v4.10-rc1 with commit
72ecad22d9f198aafee64218512
5 matches
Mail list logo
