From mboxrd@z Thu Jan 1 00:00:00 1970 From: marex@denx.de (Marek Vasut) Date: Thu, 19 Jul 2012 11:13:29 +0200 Subject: Linux ARM with XZ compression issue In-Reply-To: References: <201207181634.58089.marex@denx.de> <201207190024.50846.marex@denx.de> Message-ID: <201207191113.29529.marex@denx.de> To: linux-arm-kernel@lists.infradead.org List-Id: linux-arm-kernel.lists.infradead.org Dear Jim Cromie, Note, try keeping the CC list please > hi again, > > On Wed, Jul 18, 2012 at 4:24 PM, Marek Vasut wrote: > > Dear Jim Cromie, > > > >> hi Marek, > >> > >> Surprising problem - is strstr the only one missing ? > > > > I think so. > > > >> Its only used in the stub fn, so you could cheat and > >> just return 0, or warn less discriminately and return 0. > >> Obviously, that wouldnt be acceptable for mainline. > > > > Correct, so I'm looking for proper solution. > > Ok. I think youre on the right track - ie adding the fn. > > just one more thought - > add a strcmp call into the func, see if it compiles or breaks > the same way. If it breaks, theres a larger build issue, > and Im not sure how LKML would react. Hm, looks like strcmp is also missing, even though it's declared in arch/arm/boot/compressed/string.c > >> What happens if you build with compression other than XZ ? > > > > LZMA is good, GZ as well. > > ok, thats good - > It suggests the problem is quite localized. > > I suppose strcmp could still be missing, > and possibly more likely, since the issue is so localized, > and not exposed by other builds in other arches. > > Do you already know where strstr should be added ? arch/arm/boot/compressed/string.c I believe. > Im asking cuz Im completely unfamiliar with arm, > but did recently buy a raspberry-PI, so I'll be starting > to hack on that eventually :-) > > >> is it specific to XZ as you said, or a more general problem ? > > > > Specific to XZ. > > [...] Best regards, Marek Vasut