[OpenWrt-Devel] [LEDE-DEV] [PATCH 0/4] Gemini forward-port to kernel v4.14
roman at advem.lv
Wed May 2 08:02:22 EDT 2018
On 2018-05-02 11:01, Linus Walleij wrote:
> On Wed, May 2, 2018 at 12:18 AM, Linus Walleij
> <linus.walleij at linaro.org> wrote:
>> On Sun, Apr 29, 2018 at 8:32 PM, Roman Yeryomin <roman at advem.lv>
>>> Linus, could you test that branch on your device and see if network
>>> working by default?
>> I've pulled in the branch and building it for D-Link DNS-313 right
>> Will report back.
> I found some initial snags and sent you a patch for it, mostly
> some chicken-and-egg problem for harddisk boot and clumsy
> attempt to patch the command line from my side that I replaced
> with bootargs in the device tree.
> It boots, mounts root and gets to prompt now.
> I will submit the bootargs patches upstream.
> There is some moaning in dmesg:
> [ 10.071561] Btrfs loaded, crc32c=crc32c-generic
> [ 10.105705] mbcache: exports duplicate symbol __mb_cache_entry_free
> (owned by kernel)
> [ 10.161586] jbd2: exports duplicate symbol jbd2__journal_restart
> (owned by kernel)
Oh.. wait... I've missed that DNS-313 doesn't have any onboard flash and
boots from hdd!
Seems we need CONFIG_JBD2=y and CONFIG_FS_MBCACHE=y (along with ext4 and
other fs) back.
> I guess you see this too? Related to the USB mess I guess, we can live
> with it.
No, I don't have the same, Raidsonic (and other boards, afaik) have
onboard flash which can be used for system.
So hard drives can be used for pure storage, which is more convenient,
> Network does not come up either, I suspect it is because of missing the
> Realtek PHY driver, so will investigate this further.
Could be... what does dmesg say about that? does it read any phy_id ?
Raidsonic has Marvell 88e1116 phy but works as generic phy just fine:
[ 5.731760] Generic PHY gpio-0:01: attached PHY driver [Generic PHY]
[ 5.731781] phy_id=0x01410e11, phy_mode=rgmii
openwrt-devel mailing list
openwrt-devel at lists.openwrt.org
More information about the openwrt-devel