Comments (5)
In user-space we limit length of origin chains to some small number (e.g. 5). Perhaps we need the same in kernel.
Also, if it's updated in a loop, we could ignore addition of the same stack to the chain. I.e. if the old chain is (A+B) and we add B again, then we return the current (A+B) stack id. Does it make any sense? Or we already do this?
from kmsan.
from kmsan.
In kernel stackdepot mapping from id to stack is super fast:
https://github.com/torvalds/linux/blob/master/lib/stackdepot.c#L197-L207
Should do the same in userspace.
from kmsan.
We do have a limit on chaining (7 stacks), and actually in this particular case the problem doesn't lead to stack depot explosion, because the stack traces are always the same (they start from the exit() syscall).
There's a warning in KMSAN that's printed on every 10000 dropped chains, but the case in question is on a hot path, and it renders this warning useless (and also leads to tons of unnecessary accesses to the stack depot).
from kmsan.
This shouldn't be a problem anymore.
When an origin chain hits the maximum length, it stops growing, and __msan_chain_origin() will continue assigning the same origin id to every origin created from that chain.
There will be cases when this happens too often - I am planning to introduce a tracepoint to detect such cases, but they won't be hurting us noticeably anyway.
from kmsan.
Related Issues (20)
- duplicate stack origin PC HOT 2
- The return value of 'sscanf()' should be checked HOT 1
- Don't zero out origin of uninitialized bytes in kmsan_internal_set_shadow_origin() HOT 1
- KMSAN must use contiguous shadow for physical pages HOT 1
- Should we override gfp_flags when allocating metadata pages? HOT 1
- recipe for target 'vmlinux' failed? HOT 2
- trim aux frames from stacks HOT 1
- duplicate frame for local var origin HOT 1
- BUG: KMSAN: uninit-value in kmsan_handle_dma HOT 8
- Potential deadlock in stack_depot_save() HOT 1
- Some crypto configs that use asm are not disabled HOT 1
- BUG: NULL pointer dereference in wp_page_copy (kmsan specific) HOT 3
- Compile kmsan with clang-9 with -O2 ERROR HOT 2
- compile KMSAN error HOT 5
- https://www.google.com/webhp?client=ms-android-americamovil-us-revc&source=android-home&gws_rd=ssl
- Security Policy violation Binary Artifacts HOT 15
- can't compile if select KMSAN in ubuntu 2204 HOT 3
- arch/x86/kernel/module.c:230:49: error: call to undeclared library function 'memcpy' HOT 13
- KMSAN doesn't support paravirtualization
- Why do we need to sync metadata in x86_32 page faults? HOT 3
Recommend Projects
-
React
A declarative, efficient, and flexible JavaScript library for building user interfaces.
-
Vue.js
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
-
Typescript
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
-
TensorFlow
An Open Source Machine Learning Framework for Everyone
-
Django
The Web framework for perfectionists with deadlines.
-
Laravel
A PHP framework for web artisans
-
D3
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
-
Recommend Topics
-
javascript
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
-
web
Some thing interesting about web. New door for the world.
-
server
A server is a program made to process requests and deliver data to clients.
-
Machine learning
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
-
Visualization
Some thing interesting about visualization, use data art
-
Game
Some thing interesting about game, make everyone happy.
Recommend Org
-
Facebook
We are working to build community through open source technology. NB: members must have two-factor auth.
-
Microsoft
Open source projects and samples from Microsoft.
-
Google
Google ❤️ Open Source for everyone.
-
Alibaba
Alibaba Open Source for everyone
-
D3
Data-Driven Documents codes.
-
Tencent
China tencent open source team.
from kmsan.