Re: [patch] Wrong assertion in kern_umtx.c

On Thu, Nov 13, 2014 at 12:39:40PM -0500, Eric van Gyzen wrote:
> There is a [practically] tautological assertion in kern_umtx.c.  I have
> not even compile-tested the following patch.  I'll test it when I have
> time.  I'd be grateful if someone beats me to it.
> 
> Eric
> 
> 
> diff --git a/sys/kern/kern_umtx.c b/sys/kern/kern_umtx.c
> index 33fdf71..c6b42c0 100644
> --- a/sys/kern/kern_umtx.c
> +++ b/sys/kern/kern_umtx.c
> _at__at_ -169,7 +169,7 _at__at_ struct umtxq_chain {
>  };
>  
>  #define        UMTXQ_LOCKED_ASSERT(uc)        
> mtx_assert(&(uc)->uc_lock, MA_OWNED)
> -#define        UMTXQ_BUSY_ASSERT(uc)   KASSERT(&(uc)->uc_busy, ("umtx
> chain is not busy"))
> +#define        UMTXQ_BUSY_ASSERT(uc)   KASSERT((uc)->uc_busy, ("umtx
> chain is not busy"))
>  
>  /*
>   * Don't propagate time-sharing priority, there is a security reason,
> 
Yes, I tested it, thanks for the submission.

I committed r274478, and I decided to remove macro used in single place,
at all.  There is one more place, which I added several weeks ago, but
I really do not see much point in using the macro, it obfuscates the code.