Ah, then that probably explains why it hasn't come up as a bug here before and I do apologise for the interruption, and greatly appreciate how quickly you responded.
I am running a mostly vanilla upstream kernel (for a class of special-purpose devices that have some unusual quirks) that does include blank lines. $ uname -rv 4.4.163 #1 SMP Thu Nov 15 21:07:22 UTC 2018 (4.4 rather than 4.19 and not running Debian patches, where it looks like an Ubuntu patch was adopted to remove the newline) $ sudo sh -c 'grep -l "^ *$" /proc/*/status' /proc/1/status /proc/10/status /proc/1006/status ... 99 in total, which is the same as the number of running processes + 1 (for thread-self) $ cat /sys/devices/system/cpu/vulnerabilities/spectre_v2 Mitigation: Full generic retpoline So it looks like my problem here is due to my kernel not being Debian-stock. Because of this, I think the severity of this issue can be dropped, likely to minor: this problem is not likely to affect most users, and it was pretty straightforward for me to sync upstream source and rebuild using the Debian packaging you had put together to make a usable internal build. I don't think it should be closed outright, though, because what I've done to reach this point is documented: https://kernel-team.pages.debian.net/kernel-handbook/ch-common-tasks.html#s-common-official (though I could also just rebuild my kernel with the blank-line suppressing patch); it just probably doesn't make sense to prioritise addressing an atypical usecase. -Neil On Mon, Apr 1, 2019 at 10:44 PM Paul Wise <p...@debian.org> wrote: > Control: tags -1 + unreproducible fixed-upstream > Control: forwarded -1 > https://repo.or.cz/iotop.git/commit/7c51ce0e29bd135c216f18e18f0c4ab769af0d6f > https://repo.or.cz/iotop.git/commit/0392b205b5c3973a326721c2e9f97f0fa2eefa82 > > On Mon, 2019-04-01 at 16:43 -0600, Neil Tallim wrote: > > > I'm afraid there's a flaw in the version of iotop available in Debian > (even in > > sid): retpoline support adds a blank line to /proc/<pid>/status, which > iotop was > > not able to deal with before May of 2018. The effect of this is a crash > with a > > Python stack-trace. It affects all usage of iotop on any kernels with > spectre > > mitigation. > > As far as I can tell the commit that added blank lines to the > /proc/*/status files was never added to Debian or mainline kernels and > indeed iotop works for me with Linux 4.19.28-2 from Debian unstable. > > Could you provide some info I can use to reproduce this? > > $ uname -rv > 4.19.0-4-amd64 #1 SMP Debian 4.19.28-2 (2019-03-15) > > $ sudo sh -c 'grep -l "^ *$" /proc/*/status' > > $ cat /sys/devices/system/cpu/vulnerabilities/spectre_v2 > Mitigation: Full generic retpoline, IBPB: conditional, IBRS_FW, STIBP: > disabled, RSB filling > > -- > bye, > pabs > > https://wiki.debian.org/PaulWise > >
