Re: [dmidecode] [PATCH] dmidecode: additions from smbios 2.6.1 spec update

[Jean Delvare]

On Fri, 28 Aug 2009 11:23:35 -0400, Jarod Wilson wrote:
> On Friday 28 August 2009 11:05:06 Jean Delvare wrote:
> > Hi Jarod,
> > 
> > On Thu, 27 Aug 2009 09:25:40 -0400, Jarod Wilson wrote:
> > > dmidecode: additions from smbios 2.6.1 spec update
> > > 
> > > Tested on an HP DL360G6 with LGA1366 socket cpus, PCI-E Gen 2 slots
> > > and DDR3 memory, no longer returns any <OUT OF SPEC> info.
> > > 
> > > Signed-off-by: Jarod Wilson <address@hidden>
> > > 
> > 
> > Review:
> ...
> > > @@ -1383,10 +1402,15 @@ static const char *dmi_cache_associativi
> > >           "4-way Set-associative",
> > >           "Fully Associative",
> > >           "8-way Set-associative",
> > > -         "16-way Set-associative" /* 0x08 */
> > > +         "16-way Set-associative"
> > > +         "12-way Set-associative"
> > > +         "24-way Set-associative"
> > > +         "32-way Set-associative"
> > > +         "48-way Set-associative"
> > > +         "64-way Set-associative" /* 0x0D */
> > 
> > There's a bug here: you forgot the trailing comma at the end of each
> > line, so the compiler will see a single string. This string-splitting
> > feature of the C language is also a call for bugs, it seems.
> 
> Gah. Copy-n-paste fail. Coupled with the compiler not complaining, and
> me not having a test system w/any of the newer associativities to make
> dmidecode go boom. Good catch, fixing that up locally...

I didn't check, did you add full support for SMBIOS 2.6.1 already? If
you did, then you probably want to update the specification reference
at the top of the source file. Also please make it visible in the
CHANGELOG file. You may search for the SMBIOS 2.6.0 support addition
for how this could be formatted.

> > Other than that, it looks good to me (although I admit I did not
> > cross-check with the SMBIOS specifications). Thanks for your
> > contribution!
> 
> No problem. Reasonably sure I got everything else right, but its also
> possible there are things I neglected to add...

I have 3 SMBIOS 2.6 implementations at hand and things look OK after
adding the missing commas.

-- 
Jean Delvare