Re: [PATCH blktests] zbd: do not handle write pointer values as numeric for full zones

[Shinichiro Kawasaki <shinichiro.kawasaki@wdc.com>]

On Feb 20, 2026 / 06:35, Damien Le Moal wrote:
> On 2/19/26 21:22, Shin'ichiro Kawasaki wrote:
> >  	local _IFS=$IFS
> > -	local -i loop=0
> > +	local -i loop=0 cond
> >  	IFS=$' ,:'
> >  	while read -r -a _tokens
> >  	do
> >  		ZONE_STARTS+=($((_tokens[1])))
> >  		ZONE_LENGTHS+=($((_tokens[3])))
> >  		ZONE_CAPS+=($((_tokens[cap_idx])))
> > -		ZONE_WPTRS+=($((_tokens[wptr_idx])))
> > -		ZONE_CONDS+=($((${_tokens[conds_idx]%\(*})))
> > +		# The latest blkzone reports 'N/A' as write pointers for full
> > +		# zones. In that case, do not handle it as numeric.
> > +		cond=$((${_tokens[conds_idx]%\(*}))
> > +		if ((cond == ZONE_COND_FULL)); then
> 
> Missing handling of readonly and offline too. These also have invalid WP.

That's right. And I noticed that the latest blkzone reports "N/A" as write
pointers for conventional zones also. Then the above part will need to check
zone type also.

> 
> Anyway, why even look at the output from blkzone for this case ? The WP is
> invalid, so its value is "garbage" !
> Why not simply something like:
> 
> 		ZONE_WPTRS+=($(( $((_tokens[1])) + $((_tokens[3])))
> 
> or just use ULLONG_MAX/-1 for the value. Any value will do. Doing so, you are
> not dependent on what blkzone displays for full, readonly and offline zones.

That sounds good. I will revise this patch to -1 in v2.