Linus Torvalds launches Linux kernel 5.10, warns devs not to send 5.11 code too close to Christmas

Linus Torvalds has released version 5.10 of the Linux kernel and given developers working on the project a pre-Christmas deadline to get their desired additions for 5.11 into his inbox. Torvalds’ release announcement for version 5.10 stated: “I pretty much always wish that the last week was even calmer than it was, and that's …

[Anonymous Coward]

bye bye 2038->1901, I'll not miss you at all.

"knocks off year 2038 bug" gets my vote as the best Linux kernel yet. No one wants to see this happen:

2039-01-19 03:14:06 (Tuesday)

2039-01-19 03:14:07 (Tuesday)

1901-12-13 20:45:52 (Friday)

1901-12-13 20:45:53 (Friday)

Although an early weekend, might be nice.

[Tom 7]

Re: bye bye 2038->1901, I'll not miss you at all.

Could call it the brexit deal at last bug.

[Robert Grant]

Re: bye bye 2038->1901, I'll not miss you at all.

1901 was the best year anyway! Everyone knew their place, amirite?

[Graham Cunningham]

Re: bye bye 2038->1901, I'll not miss you at all.

@AC, am I missing a subtlety in your post? Unix epoch is 1970-01-01 00:00:00 - 2038-01-19 03:14:07

[amacater]

Re: bye bye 2038->1901, I'll not miss you at all.

Is it just an off by one error :)

[Anonymous Coward]

Re: bye bye 2038->1901, I'll not miss you at all.

(Sorry I typoed 2038 as 2039 in my post, it should have been)

2038-01-19 03:14:06 (Tuesday)

2038-01-19 03:14:07 (Tuesday)

1901-12-13 20:45:52 (Friday)

1901-12-13 20:45:53 (Friday)

1970-01-01 00:00:00 + 0x7FFFFFFF = 2038-01-19 03:14:07

(0x7FFFFFFF hex = 01111111111111111111111111111111 in binary = 2147483647 decimal)

1970-01-01 00:00:00 + 0xFFFFFFF = 1901-12-13 20:45:52

(0x80000000 hex = 10000000000000000000000000000000 in binary = -2147483647 decimal)

The 2038 bug exists because programmers decided to use a signed 32-bit integer, instead of an unsigned 32-bit integer.

[Anonymous Coward]

Re: bye bye 2038->1901, I'll not miss you at all.

I'm off to get my coffee, still asleep.

0xFFFFFFF should have been 0x80000000

[Graham Cunningham]

Re: bye bye 2038->1901, I'll not miss you at all.

@AC Thanks for the workings. I assumed it was unsigned and rolling back to 0, so I learned something today! :)

[Anonymous Coward]

Re: bye bye 2038->1901, I'll not miss you at all.

"programmers decided to use a signed 32-bit integer, instead of an unsigned 32-bit integer"

It wasn't really a choice; they needed to support recent-past dates as of that era. The only alternative would've been unsigned but with an earlier epoch in which case they quite possibly would have chosen 1900-01-01

[Ian Johnston]

What happens to the Linux kernel when Linus retires, dies from COVID-19 or simply falls under a bus? The whole project seems extraordinarily dependent on one person.

[Anonymous Coward]

Funnily enough, he has considered it.

As reported on this esteemed site

Greg Kroah-Hartman would take over the reins, at least temporarily.

[Ian Johnston]

Thanks. Even so, for such an internationally important project to be so informally dependent on a very small number of people still seems a bit risky.

[matjaggard]

It's not really - far better to have a person deciding what goes in than a corporation. If it all goes titsup the worst that happens is several Linux forks become popular. Bad for invention and unity but not really the end of the world and few people using Linux would even notice.