From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: by yocto-www.yoctoproject.org (Postfix, from userid 118) id A2246E00A08; Fri, 10 Mar 2017 18:54:14 -0800 (PST) X-Spam-Checker-Version: SpamAssassin 3.3.1 (2010-03-16) on yocto-www.yoctoproject.org X-Spam-Level: X-Spam-Status: No, score=-2.0 required=5.0 tests=BAYES_00,DKIM_SIGNED, DKIM_VALID,DKIM_VALID_AU autolearn=ham version=3.3.1 X-Spam-HAM-Report: * -1.9 BAYES_00 BODY: Bayes spam probability is 0 to 1% * [score: 0.0000] * -0.1 DKIM_VALID_AU Message has a valid DKIM or DK signature from author's * domain * 0.1 DKIM_SIGNED Message has a DKIM or DK signature, not necessarily * valid * -0.1 DKIM_VALID Message has at least one valid DKIM or DK signature Received: from fllnx209.ext.ti.com (fllnx209.ext.ti.com [198.47.19.16]) by yocto-www.yoctoproject.org (Postfix) with ESMTP id 85E57E009B9 for ; Fri, 10 Mar 2017 18:54:11 -0800 (PST) Received: from dlelxv90.itg.ti.com ([172.17.2.17]) by fllnx209.ext.ti.com (8.15.1/8.15.1) with ESMTP id v2B2rTLL001150; Fri, 10 Mar 2017 20:53:30 -0600 DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/simple; d=ti.com; s=ti-com-17Q1; t=1489200810; bh=3VWekxpLuk/UmeXyQGm/X3aF4e5TU3micPaYjVCq3n8=; h=Date:From:To:CC:Subject:References:In-Reply-To; b=i4w+/JiM2lJDJ7C0NwithyB9DyEeEuEbXaMFVQTeb5Oz545z24IprF5xk89hrEwvC 3tVVvzK3jMBimOhII92lhWzWYRtj7PhYYVdyHnhSfUv6WrVtsTn5qyL3F2D9kAKUjb /s01N2f4jovjP5/rC1zaM1cmmTho6jen6Lh1v3Is= Received: from DLEE70.ent.ti.com (dlee70.ent.ti.com [157.170.170.113]) by dlelxv90.itg.ti.com (8.14.3/8.13.8) with ESMTP id v2B2rO30022758; Fri, 10 Mar 2017 20:53:24 -0600 Received: from dlep32.itg.ti.com (157.170.170.100) by DLEE70.ent.ti.com (157.170.170.113) with Microsoft SMTP Server id 14.3.294.0; Fri, 10 Mar 2017 20:53:24 -0600 Received: from localhost (ileax41-snat.itg.ti.com [10.172.224.153]) by dlep32.itg.ti.com (8.14.3/8.13.8) with ESMTP id v2B2rOvg003851; Fri, 10 Mar 2017 20:53:24 -0600 Date: Fri, 10 Mar 2017 21:53:23 -0500 From: Denys Dmytriyenko To: Khem Raj Message-ID: <20170311025323.GB578@edge> References: <20170308170328.GA2586@haswell> <20170311014705.GY578@edge> <8777103b-9a6f-c1fb-f5ad-f04dc7f44401@gmail.com> <20170311022715.GA578@edge> <9aaf9e56-7ad8-ede6-600d-eac5879a2c7f@gmail.com> MIME-Version: 1.0 In-Reply-To: <9aaf9e56-7ad8-ede6-600d-eac5879a2c7f@gmail.com> User-Agent: Mutt/1.5.20 (2009-06-14) Cc: meta-ti@yoctoproject.org, Gary Thomas Subject: Re: Getting TI tools into [host] SDK X-BeenThere: meta-ti@yoctoproject.org X-Mailman-Version: 2.1.13 Precedence: list List-Id: Usage and development list for the meta-ti layer List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Sat, 11 Mar 2017 02:54:14 -0000 Content-Type: text/plain; charset="us-ascii" Content-Disposition: inline On Fri, Mar 10, 2017 at 06:45:00PM -0800, Khem Raj wrote: > > > On 3/10/17 6:27 PM, Denys Dmytriyenko wrote: > > On Fri, Mar 10, 2017 at 06:00:20PM -0800, Khem Raj wrote: > >> > >> > >> On 3/10/17 5:47 PM, Denys Dmytriyenko wrote: > >>> On Wed, Mar 08, 2017 at 09:03:28AM -0800, Khem Raj wrote: > >>>> On 17-03-08 09:10:18, Gary Thomas wrote: > >>>>> I'm trying to build a [host] SDK for my target which uses some > >>>>> TI tools, via -c populate_sdk_ext > >>> > >>> Gary, > >>> > >>> We don't use -c populate_sdk to produce SDKs. Instead, we still rely on the > >>> old meta-toolchain mechanism. First, because it can be incorporated into other > >>> recipes - i.e. we have a recipe that combines rootfs and SDK into one tarball. > >>> Second, it provides more control to what get's packaged into SDK. Third, it > >>> doesn't have this undefined magic to packaging stuff - i.e. this post is an > >>> example of it failing. And only one slight drawback - requires keeping rootfs > >>> image recipe in sync with meta-toolchain SDK recipe... > >>> > >>> > >>>>> In particular, I'm using recipes-ti/devtools/ti-cgt-pru_2.1.4.bb > >>>>> I include ti-cgt-pru-native to use the tools in my bitbake recipes - that works fine. > >>>>> I include ti-cgt-pru in my target image to get the tools on my board - that also works. > >>>>> > >>>>> The problem is that in my SDK, none of the binaries are included, including clpru: > >>>>> > >>>>> $ ls -l /home/gthomas/my_ti_sdk/tmp/sysroots/rainier-p8701/usr/share/ti/cgt-pru > >>>>> total 16 > >>>>> drwxrwxr-x 2 gthomas gthomas 12288 Mar 8 08:22 include > >>>>> drwxrwxr-x 3 gthomas gthomas 4096 Mar 8 08:22 lib > >>>>> > >>>>> I added this to try and include the files: > >>>>> TOOLCHAIN_HOST_TASK += "'nativesdk-ti-cgt-pru" > >>>>> and I can see that there was a nativesdk-ti-cgt-pru built which does have all the > >>>>> executables included, they just aren't being packaged. > >>>>> > >>>>> $ ls tmp/work/x86_64-nativesdk-mysdk-linux/nativesdk-ti-cgt-pru/2.1.4-r0/image/opt/mydistro/2.2+snapshot/sysroots/x86_64-amltdsdk-linux/usr/share/ti/cgt-pru/bin > >>>> > >>>> perhaps you need to correct the install location for nativesdk may be > >>>> via overriding do_install > >>> > >>> Why? > >> > >> because /usr/share doesnt look normal place for putting executables. > > > > True in general, but that shouldn't matter. FILES_${PN} lists those and that > > should be enough. It appears you are just guessing... > > > > Sidenote: > > This PRU CodeGen toolchain is kind of weird - it targets PRUs (Programmable > > Real-time Units http://elinux.org/Ti_AM33XX_PRUSSv2) and provides binaries in > > bin/, headers in include/ and libraries in lib/ directories. Some of the names > > collide with system (e.g. libc.a) and hence cannot be installed in standard > > locations, thus /usr/share. In fact, many of non-Linux specific components > > that deal with specialized cores (i.e. non-general-purpose ARM, such as DSP, > > PRU, IPU, numerous Cortex-M accelerators, etc.) are in fact as weird and are > > also getting installed in /usr/share to not confuse people and not conflict > > with the rest of the system... > > it would be better to have multiple sysroots for all these variants > would be way simpler and understandable, but why do I care. Unfortunately, OE doesn't easily support separate sysroots for "secondary" processors, besides host and target, like DSP. We had looked into that long time ago w/o much success. If you have any suggestions or recommendations, I'm all ears! -- Denys