Skip site navigation (1)Skip section navigation (2)
Date:      Mon, 9 Mar 1998 14:37:47 +1100
From:      Bruce Evans <bde@zeta.org.au>
To:        bde@zeta.org.au, toor@dyson.iquest.net
Cc:        cvs-all@FreeBSD.ORG, cvs-committers@FreeBSD.ORG, cvs-sys@FreeBSD.ORG, dyson@FreeBSD.ORG, gibbs@plutotech.com
Subject:   Re: cvs commit: src/sys/miscfs/specfs spec_vnops.c
Message-ID:  <199803090337.OAA06533@godzilla.zeta.org.au>

next in thread | raw e-mail | index | archive | help
>> >>  Modified files:
>> >>    sys/miscfs/specfs    spec_vnops.c 
>> >>  Log:
>> >>  Initialize b_resid, and also print out better diagnostics on I/O
>> >>  errors.  This will allow for better tracking of user error reports.
>> >
>> >What is b_resid initialized to?  CAM only touches b_resid when an error
>> >occurs or there is a residual, so I'm hoping it's initialized to 0.
>> >I think that other portions of the tree also expect it to be zero.
>> 
>> It is initialized to 0.  This is bogus since it is already initialized to
>> 0 (in initpbuf()).  It might be useful to initialize it to a bad value to
>> punish portions of the tree expect it to be zero.  It used to be abused
>> as the cylinder number for disksort(), so old code won't expect it to be
>> initialized.
>> 
>We don't need to punish right now.  Let's just try to stabilize, and where
>we have weaknesses, let's put appropriate diagnostics in the code.

0 is not very useful for diagnosis, since it can't be distinguished from
the value for a successful completion.  b_resid is only valid while B_DONE
is set.

Bruce



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199803090337.OAA06533>