General > General Technical Chat
Goodbye Windows, Hello Linux [advice needed for a Linux workstation at home]
<< < (35/46) > >>
radar_macgyver:

--- Quote from: Monkeh on January 24, 2019, 02:35:26 pm ---Some symbols were changed. That's not the same thing as dropping support for, uh, something never supported to begin with.

--- End quote ---
True, but gkh and a couple of other kernel developers are actively trying to make life hard for ZFS-on-Linux devs. There's a workaround, but it comes with a significant performance hit.
Monkeh:

--- Quote from: radar_macgyver on January 24, 2019, 02:44:14 pm ---
--- Quote from: Monkeh on January 24, 2019, 02:35:26 pm ---Some symbols were changed. That's not the same thing as dropping support for, uh, something never supported to begin with.

--- End quote ---
True, but gkh and a couple of other kernel developers are actively trying to make life hard for ZFS-on-Linux devs. There's a workaround, but it comes with a significant performance hit.

--- End quote ---

Again, reading too much into it - they just don't care if they cause problems.
2N3055:

--- Quote from: Monkeh on January 24, 2019, 02:45:43 pm ---
--- Quote from: radar_macgyver on January 24, 2019, 02:44:14 pm ---
--- Quote from: Monkeh on January 24, 2019, 02:35:26 pm ---Some symbols were changed. That's not the same thing as dropping support for, uh, something never supported to begin with.

--- End quote ---
True, but gkh and a couple of other kernel developers are actively trying to make life hard for ZFS-on-Linux devs. There's a workaround, but it comes with a significant performance hit.

--- End quote ---

Again, reading too much into it - they just don't care if they cause problems.

--- End quote ---

No, they don't give a damn about ZFS and will not care to support it... It's a licensing thing with Oracle/Sun.

begin quote :
"
This ZOL + Linux 5.0 issue stems from a set of functions used by this ZFS Linux port for vectorized file-system checksums no longer being exported. The kernel developers don't want to re-export the functionality since as Greg Kroah-Hartman put it, "my tolerance for ZFS is pretty non-existant."

Since that Phoronix article last week, Greg KH followed up on the mailing list with, "Sorry, no, we do not keep symbols exported for no in-kernel users." Longtime Linux kernel developer Christoph Hellwig also suggested users switch instead to FreeBSD if caring about ZFS.

Greg KH also commented on Tuesday that Oracle/Sun are the ones to fundamentally blame for the license by which they put out ZFS as for upstream kernel developers not caring about the support. "ZFS could be the best filesystem ever to grace this planet, that's fantastic, but given that the creators of that code placed it under a license that was specifically designed to not be compatible with Linux to prevent it from ever being used on Linux, well, you can see why I really don't care about it. Why would I?"

"
end quote

from : https://www.phoronix.com/scan.php?page=news_item&px=ZFS-On-Linux-5.0-Workaround

ediit: clarified quoting
Nominal Animal:

--- Quote from: radar_macgyver on January 24, 2019, 02:44:14 pm ---but gkh and a couple of other kernel developers are actively trying to make life hard for ZFS-on-Linux devs.
--- End quote ---
Is that what

--- Quote from: gregkh ---Sun explicitly did not want their code to work on Linux, so why would we do extra work to get their code to work properly?
--- End quote ---
says to you?  Funny.  I thought it says that Linux kernel developers are not willing to do extra work to get ZFS to work on Linux.

The Linux kernel codebase is one of the most actively modified codebase in the world right now.  Demanding that developers retain support for out-of-tree features is demanding that they tie one hand behind their back; that they freeze part of the development ... so that those who do not want to cooperate can keep up.  It is not "actively trying to make life hard", no matter how often it is repeated; it is just refusing to stop for stragglers.
radar_macgyver:

--- Quote from: Nominal Animal on January 24, 2019, 03:21:24 pm ---
--- Quote from: radar_macgyver on January 24, 2019, 02:44:14 pm ---but gkh and a couple of other kernel developers Is that what

--- Quote from: gregkh ---Sun explicitly did not want their code to work on Linux, so why would we do extra work to get their code to work properly?
--- End quote ---
says to you?  Funny.  I thought it says that Linux kernel developers are not willing to do extra work to get ZFS to work on Linux.

--- End quote ---

--- End quote ---
No,

--- Quote from: gregkh ---My tolerance for ZFS is pretty non-existant.

--- End quote ---
This is.

Read the full thread. Gkh turned an inlined function into a regular one, and while doing so marked it GPL-only. I don't see this as 'doing extra work' as much as giving the finger to modules that used that function. In any event, my plan is to switch to FreeBSD for my ZFS servers and I don't use it on workstations anyway (memory hog). We are going way off-topic again.
Navigation
Message Index
Next page
Previous page
There was an error while thanking
Thanking...

Go to full version
Powered by SMFPacks Advanced Attachments Uploader Mod