Vladimír Štill

Unreasonably large stack variables in array benchmarks

4

The `c/array-multidimensional/add-3-n-u_true-unreach-call` benchmark contains following variables: ```c int i,j,k; int m=1000,n=1500,p=1800; unsigned int A [m][n][p]; unsigned int B [m][n][p]; unsigned int C [m][n][p]; ``` This should allocate 30 GB of...

Floating point division by zero in float-newlib tasks (undefined behaviour)

3

The following tasks in `float-newlib` contain deliberate floating-point division by zero: ``` double_req_bl_0910a_true-unreach-call.c double_req_bl_0910b_true-unreach-call.c double_req_bl_0920a_true-unreach-call.c double_req_bl_0920b_true-unreach-call.c double_req_bl_0921_true-unreach-call.c double_req_bl_0930_true-unreach-call.c double_req_bl_0931_true-unreach-call.c double_req_bl_0960b_true-unreach-call.c double_req_bl_0970a_true-unreach-call.c double_req_bl_0970b_true-unreach-call.c double_req_bl_0971_true-unreach-call.c double_req_bl_0981_true-unreach-call.c float_req_bl_0910a_true-unreach-call.c float_req_bl_0910b_true-unreach-call.c float_req_bl_0920a_true-unreach-call.c float_req_bl_0920b_true-unreach-call.c float_req_bl_0921_true-unreach-call.c float_req_bl_0930_true-unreach-call.c...

Cgroup

5

Add support for (optionally) running nix-daemon actions in cgroups of calling process. Supported only if daemon can access given cgroups.

Is header stack fixed width or not?

8

Hi, I am wondering if a header stack is a fixed-width type and the spec does not seem to help me here (I have checked the `main` branch too). Could...

get_mac_address returns `None` if there are no routes on the Linux system

8

**Describe the bug** If there are no routes on the system, the `getmac.get_mac_address()` returns `None` even if there are network interfaces. This can happen e.g. in testing environments such as...

This is the reason behind #4626. A pass manager hook that is useful for testing (p4test-like) the compiler output or just for the developers to see what passes emit diagnostic...

[P4Testgen] Add a p4test-like target to testgen.

7

@fruffy, I think you have mentioned this already at some point and I think this would make sense. Currently, all the testgen P4 testing happens on one of the targets...

[P4Testgen] The actions' coverage works weirdly if action is called from multiple tables

4

Currently the `--track-coverage ACTIONS` (introduced for #4305) considers actions cloned for different tables to be equal. While that does make some sense, I believe it would make more sense to...