Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

Notes on understanding how kernel can use BTF #36

Merged
merged 25 commits into from
Jun 11, 2020
Merged

Notes on understanding how kernel can use BTF #36

merged 25 commits into from
Jun 11, 2020

Conversation

netoptimizer
Copy link
Member

Keeping notes while trying to understand how we as kernel developers can use BTF from the kernel-side to validate and dynamically extend data provided from userspace.

Trying to share as a GitHub pull-request, not sure it makes sense, but I'm trying to find a way to share it with @LorenzoBianconi and @dsahern

@netoptimizer
areas/arm64: Adjust order of U-Boot workaround for board_nxp_ls1088
52ee641 
Signed-off-by: Jesper Dangaard Brouer <brouer@redhat.com>
@netoptimizer
areas/core: Dig deeper into using BTF
78f9744 
Signed-off-by: Jesper Dangaard Brouer <brouer@redhat.com>
@netoptimizer
areas/core: BTF motivation
814f6d6 
Signed-off-by: Jesper Dangaard Brouer <brouer@redhat.com>
@netoptimizer
areas/core: BTF investigate other code
a434b6d 
Signed-off-by: Jesper Dangaard Brouer <brouer@redhat.com>
@netoptimizer
areas/core: BTF idea/reason for making is more dynamic
fb5b220 
Signed-off-by: Jesper Dangaard Brouer <brouer@redhat.com>
@netoptimizer
areas/core: BTF name of struct is important
d707ad4 
Signed-off-by: Jesper Dangaard Brouer <brouer@redhat.com>
@netoptimizer
areas/core: BTF target audience and link to cool article by FB
6b06b37 
Signed-off-by: Jesper Dangaard Brouer <brouer@redhat.com>
@netoptimizer netoptimizer self-assigned this May 22, 2020
@netoptimizer
areas/core: BTF reason for dynamic map-value
2b821ce 
Signed-off-by: Jesper Dangaard Brouer <brouer@redhat.com>
@netoptimizer
areas/core: BTF follow map_check_btf code
aaface0 
Signed-off-by: Jesper Dangaard Brouer <brouer@redhat.com>
@netoptimizer
BTF experiments + debugging map with BTF
76f674b 
Signed-off-by: Jesper Dangaard Brouer <brouer@redhat.com>
@netoptimizer
BTF failure debugging via ftrace and trace-cmd
a1fe895 
Record these trace-cmd commands, even-though it was not super
helpfully to debug the issue.   As I found it manually via
looking at the code,

The ftrace output was still useful to understand the kernel
call-graph flow.

Signed-off-by: Jesper Dangaard Brouer <brouer@redhat.com>
@netoptimizer
BTF: Propose changing return code for map_check_no_btf
f209a53 
Signed-off-by: Jesper Dangaard Brouer <brouer@redhat.com>
@netoptimizer
V2: bpf: Fix map_check_no_btf return code
9a7de10 
Signed-off-by: Jesper Dangaard Brouer <brouer@redhat.com>
@netoptimizer
BTF some code notes
6d4279f 
Signed-off-by: Jesper Dangaard Brouer <brouer@redhat.com>
@netoptimizer
BTF concrete proposal notes
55b2572 
Signed-off-by: Jesper Dangaard Brouer <brouer@redhat.com>
@netoptimizer
areas/core: BTF finding existing BTF to BTF compare funcs
a2bf432 
Signed-off-by: Jesper Dangaard Brouer <brouer@redhat.com>
@netoptimizer
BTF attempt-1 rejected, start on proposal 2
f63e6b1 
Signed-off-by: Jesper Dangaard Brouer <brouer@redhat.com>
@netoptimizer
BTF proposal and attempt-2, digging into code details
38818bf 
Signed-off-by: Jesper Dangaard Brouer <brouer@redhat.com>
@netoptimizer
BTF forgot to commit notes for attempt-2
a21b7ca 
Signed-off-by: Jesper Dangaard Brouer <brouer@redhat.com>
@netoptimizer
Abandon BTF: avoid FD zero is used by BPF
fa97977 
This attempt-3 bascally abandon using BTF to solve the kABI issue.

Signed-off-by: Jesper Dangaard Brouer <brouer@redhat.com>
@netoptimizer
Patchset: attempt-3 - no FD zero
3d2e153 
Add cover letter

Signed-off-by: Jesper Dangaard Brouer <brouer@redhat.com>
@netoptimizer
areas/core: BTF_02 add links to Patches attempt-1
0f604db 
Signed-off-by: Jesper Dangaard Brouer <brouer@redhat.com>
@netoptimizer
areas/core: BTF_02 record that proposal was rejected upstream
ce73688 
Keeping this document around, as it shows how kernel can interact
with the BTF data-structures inside the kernel.

Signed-off-by: Jesper Dangaard Brouer <brouer@redhat.com>
@netoptimizer
areas/core: BTF_02 attempt-4
3e754d8 
Patchset without BTF V2.

Signed-off-by: Jesper Dangaard Brouer <brouer@redhat.com>
@netoptimizer
areas/core/BTF: cleanup notes
55fcdd8 
Signed-off-by: Jesper Dangaard Brouer <brouer@redhat.com>
@netoptimizer
Copy link
Member Author

Going to merge these BTF notes, even-though my idea was not merged upstream.
These notes are still valuable to get an understanding of BTF internals as a kernel developer.

@netoptimizer netoptimizer merged commit cd600fa into master Jun 11, 2020
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Reviewers

@LorenzoBianconi LorenzoBianconi Awaiting requested review from LorenzoBianconi

@dsahern dsahern Awaiting requested review from dsahern

Assignees

@netoptimizer netoptimizer

Labels
None yet
Projects
None yet
Milestone
No milestone
Development

Successfully merging this pull request may close these issues.
1 participant
@netoptimizer