feat: more link inspection support #1896
Conversation
secDre4mer
force-pushed
the
feat/add-link-info-types
branch
2 times, most recently
from
26a871d to
9b0dcdd
Compare
mmat11
left a comment
mmat11
left a comment
There was a problem hiding this comment.
looks mostly good to me, left a few comments; can this be split in multiple commits?
| if umi.offsets == nil { | ||
| return nil, fmt.Errorf("no offsets available") | ||
| } | ||
| ex, err := OpenExecutable(umi.path) |
There was a problem hiding this comment.
hmm I wonder if this is safe to do; should we just return addresses like for kprobeMulti?
There was a problem hiding this comment.
I'm not sure, that's why there are the different Offsets and Symbols functions. Offsets returns the raw offsets, and only Symbols tries to resolve them.
Sure, I'll try. Any ideas on how the commits should be structured? E.g. one focused on uprobe, one focused on tracepoint, ...? |
I guess you could split code generation and the different link types in different commits |
secDre4mer
force-pushed
the
feat/add-link-info-types
branch
2 times, most recently
from
d0b8794 to
5f0a645
Compare
|
Split is done. |
|
@secDre4mer sorry for the delay, I was on vacation; I ran CI and some tests are failing, could you please take a look? |
Don't worry about the delay, I hope your vacation was nice! The CI complains that some generated files are out of date, see here; however, running |
hmm I think you would need to update the script has been added after last time I tried to update kernel types so I am not sure this is the correct thing to do -- /cc @lmb |
secDre4mer
force-pushed
the
feat/add-link-info-types
branch
from
5f0a645 to
ab1e2f4
Compare
Signed-off-by: Max Altgelt <max.altgelt@nextron-systems.com>
secDre4mer
force-pushed
the
feat/add-link-info-types
branch
from
ab1e2f4 to
fe34173
Compare
|
@mmat11 Thanks for the hints, the solution turned out to be somewhat easier than expected; main was already on 6.16, so I only had to rebase the branch and re-run Now there are some issues with connecting to git.kernel.org, but those look unrelated to the changes I've made. |
|
Re-triggered the tests, seems like we got some legit failures now |
|
I'll take a look. |
Signed-off-by: Max Altgelt <max.altgelt@nextron-systems.com>
Signed-off-by: Max Altgelt <max.altgelt@nextron-systems.com>
Signed-off-by: Max Altgelt <max.altgelt@nextron-systems.com>
secDre4mer
force-pushed
the
feat/add-link-info-types
branch
from
fe34173 to
fc75121
Compare
The netfilter options / info required / contained protocol family and hook, with little information for the user what values were expected. Add explicit types with constants to clarify what values are supported. Signed-off-by: Max Altgelt <max.altgelt@nextron-systems.com>
Signed-off-by: Max Altgelt <max.altgelt@nextron-systems.com>
Signed-off-by: Max Altgelt <max.altgelt@nextron-systems.com>
Signed-off-by: Max Altgelt <max.altgelt@nextron-systems.com>
secDre4mer
force-pushed
the
feat/add-link-info-types
branch
from
fc75121 to
cb8c71f
Compare
3 participants
Add support for more link inspection types, e.g. to query a uprobe's symbol.