Linux 4.19.8 Released With BLK-MQ Fix To The Recent Data Corruption Bug

[TemplarGR]

As far as i know this bug was only happening if you used blk-mq without selecting an I/O scheduler... On Archlinux i use the Zen kernel and it appears they had it enabled but they also currently use bfq i/o scheduler. I suppose that is why i wasn't affected.

[jtoaspern]

I am using a nvme SSD with bfq low_latency and F2FS. Would I benefit from using blk-mq? I may be misunderstanding something.

[duby229]

Again proves that distros in general don't give a sh*t about performance.

I wouldn't say that at all. CFQ "just works" and that is more important to Ubuntu and Redhat than relatively untested recent code. Honestly there are hundreds of distros so it seems likely you picked the wrong one for you. If you picked Ubuntu and you want a performance optimized experience then you picked the wrong distro. You're not the only one tho, most people do it.

The real truth is Ubuntu is -exactly- the -most- wrong distro for most people.

I think it's time for you to distro hop till you find one you like most.

[puleglot]

I am using a nvme SSD with bfq low_latency and F2FS. Would I benefit from using blk-mq? I may be misunderstanding something.

You are already using blk-mq. There is no other alternative for nvme.

[Weasel]

lol this bug affected NVMe mostly? Sounds pretty minor to me, such drama is unwarranted.

Maybe don't use NVMe without expecting it to be a niche thing.

[puleglot]

lol this bug affected NVMe mostly?

No, NVMe is not affected

https://bugzilla.kernel.org/show_bug.cgi?id=201685#c276

[Mel Spektor]

4.19 is a mess, 4.20 is not going anywhere stay away.

[caligula]

4.19 is a mess, 4.20 is not going anywhere stay away.

Cool story bro.