Date: Thu, 4 Jun 2009 07:26:17 +0400 From: Eygene Ryabinkin <rea-fbsd@codelabs.ru> To: Marius Strobl <marius@alchemy.franken.de> Cc: kmacy@freebsd.org, sparc64@freebsd.org, rwatson@freebsd.org, FreeBSD Tinderbox <tinderbox@freebsd.org>, khb@freebsd.org, current@freebsd.org Subject: Re: [head tinderbox] failure on sparc64/sun4v Message-ID: <k2BOEYfMcn9Q2wvaWZK0dLZVLhc@BpFm1zkZmHABxHH1eUOcQSRoWTc> In-Reply-To: <20090603194453.GA43137@alchemy.franken.de> References: <20090602222445.2F6017302F@freebsd-current.sentex.ca> <mqQn8SFFPOY77oNsI7n1tk5O7LE@10Ilc7MfiXA2JVIRVQpZfk7cTQ4> <rPuoszoUcXlrNmrZFDbbbNJuZzs@XX1fo6zQUfC4h0jjRC6IBz3oNH4> <20090603194453.GA43137@alchemy.franken.de>
next in thread | previous in thread | raw e-mail | index | archive | help
Marius, good day.
First, thanks for committing the fix.
Wed, Jun 03, 2009 at 09:44:53PM +0200, Marius Strobl wrote:
> On Wed, Jun 03, 2009 at 01:54:30PM +0400, Eygene Ryabinkin wrote:
> > KTR's case seems to be wrong for PCPU_NAME_LEN larger than 24 bytes.
> > Just now we won't be able to reach this with the current definition
> > for PCPU_NAME_LEN, but some day (N - (PCPU_NAME_LEN + 7)/8) can
> > become negative and that's bad.
>
> If it actually becomes negative the build should break again,
> which IMO is sufficient protection.
Yes, "protection" is here. But it will break the build again
and that's a bit uncomfortable.
> > The attached patch should fix this (although I have no sun4v to test
> > on, so take it with a grain of salt).
>
> I think this is overengineered, especially if not also
> adjusting the padding for other macros which may change the
> size of both MD and MI parts of struct pcpu.
Hmm, don't fully understand about "other macros". Could you,
please, provide an example?
> > By the way, having looked at sys/sys/pcpu.h, I see that there are parts
> > of 'struct pcpu' that depend on the KTR_PERCPU being defined and they
> > are never compensated with padding in PCPU_MD_FIELDS for sun4v. Is
> > KTR_PERCPU constant for sun4v (inexisting or defined everytime) or I am
> > missing something?
>
> It's just not taken into account but AFAICT also dead code.
Yes, seems like so. John, may be we can eliminate the only reference to
KTR_PERCPU from sys/sys/pcpu.h? Both 'struct pcpu' fields seem to be
unused (grep'ped -CURRENT sources).
--
Eygene
_ ___ _.--. #
\`.|\..----...-'` `-._.-'_.-'` # Remember that it is hard
/ ' ` , __.--' # to read the on-line manual
)/' _/ \ `-_, / # while single-stepping the kernel.
`-'" `"\_ ,_.-;_.-\_ ', fsc/as #
_.-'_./ {_.' ; / # -- FreeBSD Developers handbook
{_.-``-' {_/ #
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?k2BOEYfMcn9Q2wvaWZK0dLZVLhc>