Support zstd compression (port of Allan Judes patch from FreeBSD) #8941
Conversation
behlendorf
added
the
Status: Design Review Needed
|
oh man... seems like a never-ending story... |
|
There is newer code on my side as well, that cleans up a number of issues and sorts out some of the API changes. I am looking at proposing it as a number of separate pull requests to make it easier to review as well. |
|
can you point me to your current patch / code. so i can review it and
merge your changes? this is the main problem right now. we dont have
shared codebase since this patch is delayed for a long time so its hard
for me to track your changes. the bsd variant of ZoL which has been
established several months ago doesnt contain your code, so i need to
know where your current code is located to review the changes you made
for merging it into my tree
Am 29.06.2019 um 00:39 schrieb Allan Jude:
…
There is newer code on my side as well, that cleans up a number of
issues and sorts out some of the API changes. I am looking at
proposing it as a number of separate pull requests to make it easier
to review as well.
—
You are receiving this because you authored the thread.
Reply to this email directly, view it on GitHub
<#8941>,
or mute the thread
<https://github.com/notifications/unsubscribe-auth/AB2WNE3K4DLKFRHBC46HH2TP42HJ5ANCNFSM4HZTF3NQ>.
|
|
in addition. i run this code on my main filesystem on a server for compiling sourcecodes on my router firmware project which proofed that even the current zol variant is bullet proof stable. (at least for me). i would appreciate if it gets merged to the development tree finally. so we can focus to enhance the api for bsd code sharing as next step. but we finally need a shared codebase to track changes from your side |
Same on my side. I run 2 systems (1x Proxmox Hypervisor with Kernel 4.15) and a Kernel 5.0 Ubuntu machine with mentioned ZSTD patches for several months without problems aswell. |
me too |
|
@allanjude everyone is waiting on you :P |
|
I am aware, I'll try to get things posted as soon as I can |
|
My updated version has been posted finally: #9024 |
BrainSlayer
force-pushed
the
master
branch
10 times, most recently
from
421255d
to
fe68cd2
Compare
i will review your version soon and merge it with my changes. i already have seen that your version has some issues which have been resolved in my tree. mainly the memory allocation issues arent handled by your code and will fail |
BrainSlayer
force-pushed
the
master
branch
7 times, most recently
from
e489c0f
to
568593c
Compare
|
is a very bad reason to keep code that's outside test coverage anyway. |
|
yes and as i said. its simple enough to bring it back if required one time in future. |
|
@Ornias1993 @c0d3z3r0 finally i managed how to use git merge + rebase to get clean history :-) |
|
Someone deserves a cookie! |
|
yay for random crashes....... Amazon 2 unrelated |
|
@c0d3z3r0 i dont expect any error here related to my last commits. these are cosmetic |
yeah, seems unrelated at all to me |
oh.. oops. just realized I commented on the wrong PR... sorry |
|
Just FYI: |
|
@dan-and cool. thanks for update. any performance issues or everything running smooth? |
|
All running smooth. I used compressible and non-compressible data. Performance just great. It's all fine |
|
@dan-and Thats what we wanna hear! :) |
Signed-off-by: Sebastian Gottschall <s.gottschall@dd-wrt.com>
in case a background thread released already a object at that time, we are able to reuse it. so we dont need to skip the object if the size fits for our needs Signed-off-by: Sebastian Gottschall <s.gottschall@dd-wrt.com>
right now kmem_alloc using KM_SLEEP will never fail since its dead lock routing, so the double mutex_exit will not happen. but implementation wise its still wrong. there are future plans to enhance the memory allocation to allow it to be failed under OOM conditions to prevent deadlocks. so this bad code here must be fixed Signed-off-by: Sebastian Gottschall <s.gottschall@dd-wrt.com>
Signed-off-by: Sebastian Gottschall <s.gottschall@dd-wrt.com>
Signed-off-by: Sebastian Gottschall <s.gottschall@dd-wrt.com>
previously this macro was unavailable in zfs, but since its now introduced we dont need to use our own solution anymore Signed-off-by: Sebastian Gottschall <s.gottschall@dd-wrt.com>
|
@BrainSlayer Why the force push? |
|
@Ornias1993 just rebase to be in sync with upstream |
|
i will close this PR here in favor of #9735 which is the successor of this PR based on the same code. |
|
LGTM |
This Patch adds zstd compression support zo ZFS
Note:
this is a rework of the original pull request to fullfill the requirement of only offering a single patch. unfortunatly it seems that i have to spend alot of work every day to maintain this pull requests since it gets in conflict with the upstream master quickly so it will evolve quickly and will change usually every day.
Signed-off-by: Sebastian Gottschall s.gottschall@dd-wrt.com
Motivation and Context
Description
How Has This Been Tested?
Types of changes
Checklist:
Signed-off-by.