Date: Wed, 2 Mar 2016 23:10:02 -0800 From: Russell Haley <russ.haley@gmail.com> To: Ilya Bakulin <ilya@bakulin.de> Cc: "freebsd-arm@freebsd.org" <freebsd-arm@freebsd.org> Subject: Re: MMC/SDIO stack under CAM Message-ID: <CABx9NuRPA6XLb4MON=-3C96DbfshRrM5BRBQxvuTL-gdWRcaQg@mail.gmail.com> In-Reply-To: <1456154162.1294.51.camel@freebsd.org> References: <20140216111153.GA74858@olymp.kibab.com> <5C2CF572-360D-4CA0-81C7-18A5C455AED5@bsdimp.com> <20140224142642.GA32538@olymp.kibab.com> <CAJ-VmomNzCMc1T=0jAnyd_uGXbvgeTzZTtmhUPSvZ0DKUEjtKg@mail.gmail.com> <53120EE8.1080600@bakulin.de> <CAJ-VmonPkdVVq7nC3FdopcgzmSTsj3gTO=Cghx-62XS5K25YQg@mail.gmail.com> <5688F015.4090002@bakulin.de> <CAASDrVkHKzt4LKnLEFbwZ0gWRpwEOcpjaCFaOTg%2Bs2DACebopA@mail.gmail.com> <E0EFD24D-B888-4025-AC41-C3D21E5536E9@bakulin.de> <6942A46B-110B-4E1F-9DA1-F965009E8E92@FreeBSD.org> <38dd08fc2a5930d58b09e9bd3cb6d3e7@bakulin.de> <BFA9D9D2-493F-489E-96D2-BB33F6034AD3@gmail.com> <CANCZdfrcR=MQ0KfYN0sgTWPe_evH8p9Z466DU04=uohiuDkqSA@mail.gmail.com> <CABx9NuSVD0JBORkCR2CeNfGqHrbax7nqA6pWrtqDYcDTb0%2BAnw@mail.gmail.com> <1456154162.1294.51.camel@freebsd.org>
next in thread | previous in thread | raw e-mail | index | archive | help
Hello Ilya,
I have tried running patch D4761 against r296336 and ran into a make
error. As follows:
cd /usr/src
svn patch /usr/patches/D4761.diff
U etc/mtree/BSD.include.dist
> applied hunk @@ -92,6 +92,8 @@ with fuzz 2
> applied hunk @@ -224,7 +226,7 @@ with offset 2
U include/Makefile
U lib/libcam/Makefile
U sbin/camcontrol/camcontrol.c
U sys/Makefile
A sys/arm/conf/BEAGLEBONE-CAM
U sys/arm/ti/ti_sdhci.c
U sys/cam/cam_ccb.h
U sys/cam/cam_xpt.c
U sys/cam/cam_xpt_internal.h
A sys/cam/mmc
A sys/cam/mmc/mmc.h
A sys/cam/mmc/mmc_all.h
A sys/cam/mmc/mmc_bus.h
A sys/cam/mmc/mmc_da.c
A sys/cam/mmc/mmc_sdio.h
A sys/cam/mmc/mmc_sdio.c
A sys/cam/mmc/mmc_xpt.c
A sys/cam/mmc/mmcreg.h
U sys/conf/files
> applied hunk @@ -1554,10 +1557,10 @@ with offset 6
> applied hunk @@ -1960,6 +1963,7 @@ with offset 10
A sys/dev/mmcnull
A sys/dev/mmcnull/Makefile
A sys/dev/mmcnull/mmcnull.c
U sys/dev/sdhci/sdhci.h
U sys/dev/sdhci/sdhci.c
A sys/modules/mmcnull
A sys/modules/mmcnull/Makefile
root@a5e5b565-d3ae-11e5-a511-0c54a518c15b:/usr/src # ls
.arcconfig bin rescue
.arclint cddl sbin
.svn contrib secure
COPYRIGHT crypto share
LOCKS etc sys
MAINTAINERS gnu targets
Makefile include tests
Makefile.inc1 kerberos5 tools
ObsoleteFiles.inc lib usr.bin
README libexec usr.sbin
UPDATING release
#make TARGET=arm TARGET_ARCH=armv6 KERNCONF=IMX6 -j20 buildkernel
...
--- mmc_da.o ---
/usr/src/sys/cam/mmc/mmc_da.c:855:20: error: format specifies type
'unsigned long' but the argument has type 'uint64_t' (aka 'unsigned
long long') [-Werror,-Wformat]
softc->csd.capacity,
^~~~~~~~~~~~~~~~~~~
/usr/src/sys/cam/cam_debug.h:97:10: note: expanded from macro 'CAM_DEBUG'
printf printfargs; \
^
--- modules-all ---
--- all_subdir_accf_dns ---
--- accf_dns.o ---
--- mmc_da.o ---
1 error generated.
*** [mmc_da.o] Error code 1
make[2]: stopped in /usr/obj/arm.armv6/usr/src/sys/IMX6
--- modules-all ---
Thanks,
Russ
On Mon, Feb 22, 2016 at 7:16 AM, Ian Lepore <ian@freebsd.org> wrote:
> On Sun, 2016-02-21 at 23:28 -0800, Russell Haley wrote:
>> On Mon, Feb 15, 2016 at 9:33 PM, Warner Losh <imp@bsdimp.com> wrote:
>> > On Mon, Feb 15, 2016 at 7:22 AM, NGie Cooper <yaneurabeya@gmail.com
>> > > wrote:
>> >
>> > >
>> > > > On Feb 15, 2016, at 02:13, Ilya Bakulin <ilya@bakulin.de>
>> > > > wrote:
>> > > >
>> > > > On 2016-02-11 19:54, Stanislav Sedov wrote:
>> > > > > > On Feb 11, 2016, at 10:47 AM, Ilya Bakulin <ilya@bakulin.de
>> > > > > > > wrote:
>> > > > > > I'll use an excellent opportunity to post a small status
>> > > > > > update about
>> > > my work :-)
>> > > > > > * SDHC controller on Wandboard now works with the new
>> > > > > > stack;
>> > > > > > * SDIO block read now works!
>> > > > > > * camcontrol userland app is extended to support "mmcsdcmd"
>> > > > > > command
>> > > that allows to send MMC commands from userland apps directly to
>> > > the card
>> > > via pass(4) device -- now we can write WLAN driver in userland :
>> > > -D
>> > > > > Great news, userspace drivers are the best!:)
>> > > > > So what are the remaining pieces that prevent this work from
>> > > > > hitting
>> > > the HEAD?
>> > > > > --
>> > > > > Stanislav Sedov
>> > > > > ST4096-RIPE
>> > > >
>> > > > Hi Stas,
>> > > >
>> > > > As I'm not a committer, someone needs to review my code and
>> > > > assist in
>> > > intergration into -HEAD :-)
>> > > > Currently nobody was able to do a review because of -ENOTIME.
>> > > > The only feature that is missing in the new stack (from my PoV)
>> > > > is
>> > > working with high-speed cards -- I just haven't implemented
>> > > switching to
>> > > high-speed mode yet. Although now it's possible to send required
>> > > commands
>> > > to the card and then switch controller speed -- all using
>> > > camcontrol
>> > > mmcsdcmd :-).
>> > > >
>> > > > Do you know anyone not on CC line who is able to help me with
>> > > > this? Or
>> > > maybe you could even find some time yourself?
>> > >
>> > > Hi Ilya,
>> > > Could you please post the patch to phrabricator and CC the
>> > > interested
>> > > parties/me?
>> > >
>> >
>> > It's been up on phab for a while. There's been some comments on it.
>> > There's
>> > some
>> > things wrong still that I've been meaning to get bcak to Ilya on.
>> >
>> > When it is ready, I plan on committing this. It goes hand in hand
>> > with the
>> > nvme CAM stuff
>> > I've been working on.
>> >
>> > Anybody can take a look at it: https://reviews.freebsd.org/D4761
>> >
>>
>> I tried to apply the SDIO patch like this:
>>
>> mv ~/Downloads/D4761.diff <jailpath>/usr/patches
>>
>> Then from within my build jail:
>> cd /usr/src
>> patch < /usr/patches/D4761.diff
>>
>> and I wind up with the mmc source files in my root. What should my
>> working directory be when applying this patch?
>
> That's an annoying historical quirk of patch... for files in the
> patches where the path doesn't already exist in your filesystem it just
> creates the files in the current dir, unless you add -p0, which will
> make it create missing directories. (-p normally tells it how many
> leading dir nodes to strip from the paths found in the patches, but -p0
> is magical in that it strips nothing but creates dirs).
>
> -- Ian
>
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?CABx9NuRPA6XLb4MON=-3C96DbfshRrM5BRBQxvuTL-gdWRcaQg>