Comment 296 for bug 441835

yes Martin, I duly understand your point.

I'm not familiar at all how ubuntu is internally organised when it comes to who does what; just reading your reply it appears that there is a different group of devs that 'look' after the kernel development, functionality and its well being (for sure you know them v well - after all, you're colleagues).

maybe I'm 'barking at wrong tree' but having in mind the first para, I'm kindly asking you, and the kernel devs, to take a look at this issue and get it solved for the benefit of ubuntu os.

just a couple of points that crossed my mind about

.....'I'm pretty sure that it already has a timeout, since the detection calls do fail eventually
- it just takes excruciatingly long (some 20 seconds).'.....

- the BIOS which is written in assembler (the fastest) detects the floppy presence in fractions of a second. for sure, no doubt about it, the higher programming languages used in the kernel introduce delays, to what degree ? v hard for me to say. just a question pls. - would it be possible to have at least this part of kernel that deals with this issue be written in assembler/machine language ? and then read the cpu register/s and configure/not configure the device ? from what I remember the assembler instructions/functions are virtually the same when it comes to floppy detection since the ancient 8088/8086 cpu.

- the second thought was - how comes that the proprietary os manages to get this type of device recognised if present and render it fully operational without this 20 sec delays ? I saw it myself on a friend's pc that uses win - the floppy device detection was almost as fast as the BIOS detection.

I absolutely refuse to accept that the open source is NOT THE BEST when it comes to writing code be it the os kernel or anything else. this IS after all its v strength.