From mboxrd@z Thu Jan 1 00:00:00 1970 From: =?iso-8859-1?Q?M=E5ns_Rullg=E5rd?= Subject: Re: [PATCH 1/3] ata: sata_dwc_460ex: use "dmas" DT property to find dma channel Date: Fri, 18 Dec 2015 01:38:01 +0000 Message-ID: References: <1450221935-6034-1-git-send-email-mans@mansr.com> <1450364395.30729.136.camel@linux.intel.com> <1450367702.30729.146.camel@linux.intel.com> <1450371184.30729.175.camel@linux.intel.com> <5672F77F.4010306@candw.ms> <567302E8.5050303@candw.ms> <5673061A.4070700@candw.ms> <56732C04.9040100@candw.ms> <56735A69.1000006@candw.ms> Mime-Version: 1.0 Content-Type: text/plain; charset=utf-8 Content-Transfer-Encoding: QUOTED-PRINTABLE Return-path: In-Reply-To: <56735A69.1000006@candw.ms> (Julian Margetson's message of "Thu, 17 Dec 2015 20:59:21 -0400") Sender: linux-kernel-owner@vger.kernel.org To: Julian Margetson Cc: Andy Shevchenko , Tejun Heo , linux-ide@vger.kernel.org, linux-kernel@vger.kernel.org List-Id: linux-ide@vger.kernel.org Julian Margetson writes: > On 12/17/2015 8:06 PM, M=C3=A5ns Rullg=C3=A5rd wrote: >> Julian Margetson writes: >> >>> On 12/17/2015 3:53 PM, M=C3=A5ns Rullg=C3=A5rd wrote: >>>> Julian Margetson writes: >>>> >>>>> On 12/17/2015 2:51 PM, M=C3=A5ns Rullg=C3=A5rd wrote: >>>>>> Julian Margetson writes: >>>>>> >>>>>>> On 12/17/2015 1:59 PM, M=C3=A5ns Rullg=C3=A5rd wrote: >>>>>>>> Julian Margetson writes: >>>>>>>> >>>>>>>>> I have been running my machine mostly configured for pciex1 = thus with >>>>>>>>> the sata_dwc disabled. >>>>>>>>> The changes to sata_dwc-460ex do cause an oops. >>>>>>>>> I will try to give more detailed info over this weekend . >>>>>>>> The driver as is upstream would do that since it unconditional= ly >>>>>>>> dereferences a null pointer in the probe function. My patch f= ixes that >>>>>>>> as a side-effect. >>>>>>>> >>>>>>> patching file drivers/ata/Kconfig >>>>>>> >>>>>>> Hunk #1 FAILED at 296. >>>>>> [...] >>>>>> >>>>>>> root@julian-VirtualBox:/usr/src/linux-3.18.25# >>>>>> The patch is against 4.4-rc5. >>>>>> >>>>> CC drivers/ata/sata_dwc_460ex.o >>>>> >>>>> drivers/ata/sata_dwc_460ex.c:198:15: error: variable =C3=A2=E2=82= =AC=CB=9Csata_dwc_dma_dws=C3=A2=E2=82=AC=E2=84=A2 has initializer but i= ncomplete type >>>>> >>>>> static struct dw_dma_slave sata_dwc_dma_dws =3D { >>>>> ^ >>>> It builds, albeit with an unrelated warning, using the attached co= nfig. >>>> Maybe there's a missing config dependency somewhere. >>>> >>> I am attempting to cross compile under Ubuntu 14.04 X86 in Virtualb= ox >>> with your .config. >>> 4.4.0-rc5 builds ok with no patches applied . >>> Once your patch is applied it fails to build . >>> >>> CC drivers/ata/sata_dwc_460ex.o >>> drivers/ata/sata_dwc_460ex.c:198:15: error: variable =E2=80=98sata_= dwc_dma_dws=E2=80=99 has initializer but incomplete type >>> static struct dw_dma_slave sata_dwc_dma_dws =3D { >>> ^ >> Bizarre. This is what it looks like here: >> >> mru@unicorn:/tmp/linux-sata$ git status >> On branch sata-dwc >> nothing to commit, working directory clean >> mru@unicorn:/tmp/linux-sata$ git describe >> v4.4-rc5 >> mru@unicorn:/tmp/linux-sata$ sha1sum /tmp/0001-ata-sata_dwc_460ex-us= e-dmas-DT-property-to-find-dma-.patch >> e300971aa483390f82de2e9120dc16e460e74feb /tmp/0001-ata-sata_dwc_460= ex-use-dmas-DT-property-to-find-dma-.patch >> mru@unicorn:/tmp/linux-sata$ git am /tmp/0001-ata-sata_dwc_460ex-use= -dmas-DT-property-to-find-dma-.patch >> Applying: ata: sata_dwc_460ex: use "dmas" DT property to find dma ch= annel >> mru@unicorn:/tmp/linux-sata$ sha1sum .config >> 4e7615b8d2fa9a1c4b4ae9ffc363aefcaf3789ca .config >> mru@unicorn:/tmp/linux-sata$ make ARCH=3Dpowerpc CROSS_COMPILE=3Dpow= erpc64-none-linux-gnu- oldconfig >> HOSTCC scripts/basic/fixdep >> HOSTCC scripts/kconfig/conf.o >> SHIPPED scripts/kconfig/zconf.tab.c >> SHIPPED scripts/kconfig/zconf.lex.c >> SHIPPED scripts/kconfig/zconf.hash.c >> HOSTCC scripts/kconfig/zconf.tab.o >> HOSTLD scripts/kconfig/conf >> scripts/kconfig/conf --oldconfig Kconfig >> # >> # configuration written to .config >> # >> mru@unicorn:/tmp/linux-sata$ make ARCH=3Dpowerpc CROSS_COMPILE=3Dpow= erpc64-none-linux-gnu- drivers/ata/sata_dwc_460ex.o >> scripts/kconfig/conf --silentoldconfig Kconfig [...] >> CC drivers/ata/sata_dwc_460ex.o >> drivers/ata/sata_dwc_460ex.c: In function 'dma_dwc_xfer_setup': >> drivers/ata/sata_dwc_460ex.c:383:20: warning: cast from pointer to i= nteger of different size [-Wpointer-to-int-cast] >> dma_addr_t addr =3D (dma_addr_t)&hsdev->sata_dwc_regs->dmadr; >> ^ >> mru@unicorn:/tmp/linux-sata$ >> >> Patch file and .config attached. >> >> Looking into that warning, I doubt it works as is, but that's not ca= used >> by my patch. I can try to come up with a fix, but again, I can't te= st it. >> > I am using > > make ARCH=3Dpowerpc CROSS_COMPILE=3Dpowerpc-linux-gnu- Shouldn't matter since the compiler flags include -m32, and I only had = a powerpc64 toolchain built. Anyhow, I built a 32-bit toolchain and it still builds. Just to make sure you applied the patch correctly: mru@unicorn:/tmp/linux-sata$ sha1sum drivers/ata/sata_dwc_460ex.c=20 c8a7927840aade75ac62b04a2c9acc8335a34d6f drivers/ata/sata_dwc_460ex.c Digging deeper into that warning, it is clearly a bug which has always been there. The reason it ever worked appears to be that the 460EX has a dedicated DMA unit hard-wired to the SATA controller ignoring that address. The situation is similar on my hardware. --=20 M=C3=A5ns Rullg=C3=A5rd