Re: [PATCH RFC] block - ataflop.c: fix breakage introduced at blk-mq refactoring

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



On 10/18/21 4:51 PM, Finn Thain wrote:

On Mon, 18 Oct 2021, Jens Axboe wrote:

Was going to ask if this driver was used by anyone, since it's taken 3
years for the breakage to be spotted... 

A lack of bug reports never implied a lack of users (or potential users). 
That falacy is really getting tiresome.

If it's utterly broken, I'd say yes, it very much does imply that really
nobody is using it. 

It is much more difficult to report regressions than it is to use a 
workaround (i.e. boot a known good kernel). And I have plenty of sympathy 
for end-users who may assume that the people and corporations who create 
the breakage will take responsibility for fixing it.

We're talking about a floppy driver here, and one for ATARI no less.
It's not much of a leap of faith to assume that

a) those users are more savvy than the average computer user, as they
   have to compile their own kernels anyway.

b) that there are essentially zero of them left. The number is clearly
   different from zero, but I doubt by much.

Hence it would stand to reason that if someone was indeed in the group
of ATARI floppy users that they would know how to report a bug. Not to
mention that it was pretty broken to begin with, so can't have been used
much before either.

The reason I ask is always to have an eye out for what drivers can be
eventually removed. The older drivers, and particurly the floppy ones,
are quite horrible and would need a lot of work to bring up to modern
standards in terms of how they are written. And if nobody is really
using them, then we'd all be better off cutting some of that dead
baggage.

Do maintainers really expect innocent end users to report and bisect 
regressions, and also test a series of potential fixes until one of them 
is finally found to both work and pass code review?

If someone reports a bug to me, the most basic is usually "It worked in
this version, it's broken in this one". Then you take it from there,
depending on the abilities of the user. I'd only ask someone to bisect
an issue if it's really puzzling and the user is capable and willing.
But it doesn't take much to send a simple email saying that something
used to work and now it's broken.

-- 
Jens Axboe




[Index of Archives]     [Video for Linux]     [Yosemite News]     [Linux S/390]     [Linux Kernel]     [Linux SCSI]

  Powered by Linux