FreeBSD Is Trying To Figure Out If Anyone Uses Its VGL Graphics Library

[Guest]

Well FreeBSD has much better networking performance, an excellent file system in ZFS(Linux doesn't have an alternative where the data is critical), a better compiler, much better security, Bhyve and jails are nice(although LXC and KVM are nice too), moving between releases is much easier.
FreeNAS, Pfsense are the only good open source NAS/router OSs.

Too bad it's nothing more, but yours wishful thinking. If it had better network performance it would rule the server market, but it's nearly non existent in comparison to Linux. Furthermore, it's Linux used in most critical workloads, facebook, youtube etc. Linux has much better networking performance. ZFS is also available on Linux (and it's not BSD file system, but Solaris one) and Linux also has btrfs. GCC beats this crap, but llvm is also available for Linux. Linux has much better security (SELinux, Apparmour, TOMOYO, cgroups, aslr etc. etc.), better containers and virtualization, upgrades are easier.

http://news.netcraft.com/archives/20...-may-2016.html

[Guest]

https://www.phoronix.com/scan.php?page=news_item&px=MTc1NjY

Granted that marginal performance differences do not matter for general use, but when you want to push high pps/throughput on 10G cards, BSD is the way to go. The network stack is much more mature. An eg. would be http://dpdk.org/ already being integrated in Pfsense, whereas there is absolutely no evidence for this on Linux. You can already push nearly 10 Gbps of IPsec through an Atom processor and like 40 Gbps on other Intel Quickassist CPUs.

Another piece of unproved bullshit. The freebsd network stack is more mature? You just made my day:

http://yarchive.net/comp/linux/sendfile.html

The fact I dislike about the HP-UX implementation is that it is so _obviously_ stupid. And I have to say that I absolutely despise the BSD people. They did sendfile() after both Linux and HP-UX had done it, and they must have known about both implementations. And they chose the HP-UX braindamage, and even brag about the fact that they were stupid and didn't understand TCP_CORK (they don't say so in those exact words, of course - they just show that they were stupid and clueless by the things they brag about). Oh, well. Not everybody can be as goodlooking as me. It's a curse. Linus

[starshipeleven]

Another piece of unproved bullshit. The freebsd network stack is more mature? You just made my day:

http://yarchive.net/comp/linux/sendfile.html

Lol, someone needs to turn all these mailing lists into a book about development. I find Torvalds way of explaining things very entertaining.

[aht0]

Too bad it's nothing more, but yours wishful thinking. If it had better network performance it would rule the server market, but it's nearly non existent in comparison to Linux.

Results of more aggressive marketing. Most CEO's have heard "something about Linux". Never about BSD

Furthermore, it's Linux used in most critical workloads, facebook, youtube etc. Linux has much better networking performance.

Debatable and depends on setups used. At least one satellite operator is using FreeBSD (tell me where reliability counts more than out in space?) and Netflix is using FreeBSD successfully.

ZFS is also available on Linux (and it's not BSD file system, but Solaris one) and Linux also has btrfs.

ZFS on Linux and ZFS on BSD are not even really same species of animals. Zebra may look somewhat like purebreed charger but it's not.You can even cross it with a horse but the child will look like striped bastard. ZFS on FreeBSD is more mature and out of stage of child-deceases..
BTRFS is not even fully stable yet. Typical Linux development model. Start something with big "hurrah" and when 80-90% is done, development stops or slows down to crawl because everyone involved have by then new interests..
Linux usable for Enterprise purposes is generally very conservative and without all the bleeding edge stuff you so much holler around. Doesn't make it principally that different from FreeBSD itself.

GCC beats this crap, but llvm is also available for Linux.

LLVM compiles faster, GCC has better optimizations. "beats this crap" is utterly how you choose to take it.

Linux has much better security (SELinux, Apparmour, TOMOYO, cgroups, aslr etc. etc.)

plug capital design holes in security with another layer of security is not "much better security". BSDs do not even have need for most of this.. void argument. Wanna recall all the shellshocks, ghosts and critical sudden Linux kernel vulnerabilities over past few years?

better containers and virtualization, upgrades are easier.

BSDs do not have containers at all. Jails. Upgrades. Depends. On conservative LTS distro, upgrades generally work. On rolling ones, you have pretty big probability of having to reinstall whole OS. FreeBSD updates are generally "freebsd-update fetch" and "freebsd-update install", if one more command is "much harder", so be it. You can do building from source but YOU DONT HAVE TO.
Virtualization support in general is overall better in Linux, ok, agree.

-------
BSD network stack maturity. AFAIK Linux "borrowed" it's initial network stack from BSD because it could, thanks to it's license. Simply because of it's age and parentage, it's logically more mature.

[Guest]

Results of more aggressive marketing. Most CEO's have heard "something about Linux". Never about BSD

And let's blame Linux for this. There's no logic here. Unix was very popular and I heard BSD is spiritual successor to Unix. If CEO's didn't hear about it, it sucks.

Debatable and depends on setups used. At least one satellite operator is using FreeBSD (tell me where reliability counts more than out in space?) and Netflix is using FreeBSD successfully.

They admitted they have experience with BSD, so it could be an only reason to use it.

ZFS on Linux and ZFS on BSD are not even really same species of animals. Zebra may look somewhat like purebreed charger but it's not.You can even cross it with a horse but the child will look like striped bastard. ZFS on FreeBSD is more mature and out of stage of child-deceases..
BTRFS is not even fully stable yet. Typical Linux development model. Start something with big "hurrah" and when 80-90% is done, development stops or slows down to crawl because everyone involved have by then new interests..
Linux usable for Enterprise purposes is generally very conservative and without all the bleeding edge stuff you so much holler around. Doesn't make it principally that different from FreeBSD itself.

You're probably right.

LLVM compiles faster, GCC has better optimizations. "beats this crap" is utterly how you choose to take it.

Right, but GCC is more mature.

plug capital design holes in security with another layer of security is not "much better security". BSDs do not even have need for most of this.. void argument. Wanna recall all the shellshocks, ghosts and critical sudden Linux kernel vulnerabilities over past few years?

Are you calling single bugs as bad design decisions? The same can be said about dozens of BSD critical bugs. Without security layers you're not able to defend against many attacks, so thanks for confirming BSD doesn't care about security. However, you're mistaken, because there are some security layers in BSD as well. However, they're far behind Linux.

BSD network stack maturity. AFAIK Linux "borrowed" it's initial network stack from BSD because it could, thanks to it's license. Simply because of it's age and parentage, it's logically more mature.

I don't know if it's true or not, but if it's true it could proves my point as well, but.. The age isn't so meaningful here. There's so many new technologies available and Linux pioneered many of them. Good example was sendfile mentioned earlier. Linux had it before BSD and it seems BSD messed it up.

[aht0]

And let's blame Linux for this. There's no logic here. Unix was very popular and I heard BSD is spiritual successor to Unix. If CEO's didn't hear about it, it sucks.

By my understanding, BSD crowd just does not care. Especially OpenBSD devs.

They admitted they have experience with BSD, so it could be an only reason to use it.

Yeah, and using it worked, so what and where would imply FreeBSD's "inherent inferiority" you keep implying?

Right, but GCC is more mature.

And clunky, full of extraneous code etc etc. One likes mother, other guy likes daughter.

Are you calling single bugs as bad design decisions? The same can be said about dozens of BSD critical bugs. Without security layers you're not able to defend against many attacks, so thanks for confirming BSD doesn't care about security. However, you're mistaken, because there are some security layers in BSD as well. However, they're far behind Linux.

If you have to introduce whole new security layers in order to nullify bad design on systemd's part, then yes - I would call it. And I am not wholly sure what's the point of all those "layers" when Linux code itself is not so secure at all because of programming errors that are so common in Linux code. BSD code, especially OpenBSD's gets a lot more vetting. In fact, OpenBSD devs have separate group of people who's only job is to search security vulnerabilities in code.

I don't know if it's true or not, but if it's true it could proves my point as well, but.. The age isn't so meaningful here. There's so many new technologies available and Linux pioneered many of them. Good example was sendfile mentioned earlier. Linux had it before BSD and it seems BSD messed it up.

http://www.openbsd.org/innovations.html have fun reading trough it.

[Guest]

If you have to introduce whole new security layers in order to nullify bad design on systemd's part, then yes - I would call it. And I am not wholly sure what's the point of all those "layers" when Linux code itself is not so secure at all because of programming errors that are so common in Linux code. BSD code, especially OpenBSD's gets a lot more vetting. In fact, OpenBSD devs have separate group of people who's only job is to search security vulnerabilities in code.

Bad design like what? It's a big step forward when comes to other init systems. Those so called 'layers' are nothing, but extended Mandatory Access Control. The programming errors are present in BSD as well, but they also lack Linux's security features. FreeBSD didn't even have aslr for many years and OpenBSD is/was far from being secure:

https://allthatiswrong.wordpress.com/2010/01/20/the-insecurity-of-openbsd/#default

http://www.openbsd.org/innovations.html have fun reading trough it.

Many great tools, but it doesn't tell a lot about networking code.

[aht0]

Bad design like what? It's a big step forward when comes to other init systems. Those so called 'layers' are nothing, but extended Mandatory Access Control. The programming errors are present in BSD as well, but they also lack Linux's security features. FreeBSD didn't even have aslr for many years and OpenBSD is/was far from being secure:

https://allthatiswrong.wordpress.com/2010/01/20/the-insecurity-of-openbsd/#default

race conditions systemd would have introduced without. So, bunch of additional layers had to be introduced.. Want example?
PID1 and process supervisor.. You DO need namespaces and cgroups to cover up possibility of compromise.. It is pluggin' a hole that did not needed to be plugged in the first place.

Programming errors are certainly present in BSD's but in a degree magnitude smaller. And compromising almost any BSD component part could not have impact remotely comparable to impact of successful 0-day exploit hitting systemd.

I wonder why systemd fans are not simply moving over to MS-Windows. Even better hardware support, you seem to want one big file that controls everything in OS and seem to have immense liking for monopolistic organisations/companies that are intent on controlling whole ecosystem :P

You do not like BSD init. Not enough functionality for admin's? Fine, there are alternatives like OpenRC that allows more fine grained control. There exist FreeBSD derivatives that run OpenRC init (Arch/PacBSD)

a)that blog is from a known guy with a grudge towards OpenBSD, this not his first rant. Some of his claims were valid, some were just incompetence.
b)it's from 2010. Lot's have changed in 6,5 years in OpenBSD.

[Guest]

race conditions systemd would have introduced without. So, bunch of additional layers had to be introduced..

systemd actually fixed some of the race conditions when comes to services.

PID1 and process supervisor.. You DO need namespaces and cgroups to cover up possibility of compromise.. It is pluggin' a hole that did not needed to be plugged in the first place.

When comes to systemd and PID1 you may be right, but it's probably an only significant downside. Namespaces and cgroups are security features used in many other places.

Programming errors are certainly present in BSD's but in a degree magnitude smaller. And compromising almost any BSD component part could not have impact remotely comparable to impact of successful 0-day exploit hitting systemd.

I didn't think about systemd in such way, so I won't argue here. You made a point that needs to be checked.

I wonder why systemd fans are not simply moving over to MS-Windows. Even better hardware support, you seem to want one big file that controls everything in OS and seem to have immense liking for monopolistic organisations/companies that are intent on controlling whole ecosystem :P

Windows sucks in so many ways it's unusable for me. It even sucks when comes to hardware support. I had BSODs with GeForce 740M, logitech gamepad didn't want to work and Sound Blaster X-FI doesn't work in USB 3 port. All of this works out of the box without single problem under Linux.

You do not like BSD init. Not enough functionality for admin's? Fine, there are alternatives like OpenRC that allows more fine grained control. There exist FreeBSD derivatives that run OpenRC init (Arch/PacBSD)

Ok.

[starshipeleven]

Programming errors are certainly present in BSD's but in a degree magnitude smaller.

[citation needed]

And compromising almost any BSD component part could not have impact remotely comparable to impact of successful 0-day exploit hitting systemd.

Systemd (the init system binary itself) has a much much smaller attack surface than whatever shell is running initscripts.
Now what about something "compromising" the shell, punk?
A systemd system is still (potentially) more protected if its services are set up using kernel security features, on BSD... nope.