[PATCH 0/5][SRU][F] Add r8117 support and fix firmware base issue

Aaron Ma aaron.ma at canonical.com
Mon Jun 29 10:23:13 UTC 2020



On 6/29/20 5:29 PM, Stefan Bader wrote:
> On 28.06.20 09:58, Aaron Ma wrote:
>> BugLink: https://bugs.launchpad.net/bugs/1885072
>>
>> [Impact]
>> Realtek 8117 is failed to be initialized on 5.4- kernel,
>> even failed to work after loading firmware on 5.5+ kernel.
>>
>> [Fix]
>> Add rtl8117 support for 5.4- kernel.
>> On 5.5+ kernel, firmware should reset ocp_base, but not on r8117.
>> Correct the ocp_base in the driver.
>> Then it works well.
>>
>> [Test]
>> Verified on hardware, link status is OK, iperf test result is good.
>>
>> [Regression Potential]
>> Low.
>> Fixed issue, cherry-picked from linux-next tree.
>> SRU for OEM-5.6/U was already done. this is for focal because of more
>> commits.
>> Bionic kernel set rtl8117 as default r8169. It just works. no need these
>> changes.
>>
>> Backport 0002 patch because of another upstream patch is already merged.
>>
>> Heiner Kallweit (5):
>>   r8169: add helper r8168g_phy_param
>>   r8169: add support for RTL8117
>>   r8169: load firmware for RTL8168fp/RTL8117
>>   r8169: fix OCP access on RTL8117
>>   r8169: fix firmware not resetting tp->ocp_base
>>
>>  drivers/net/ethernet/realtek/r8169_main.c | 445 ++++++++++++----------
>>  1 file changed, 241 insertions(+), 204 deletions(-)
>>
> 
> For the last 3 patches, I cannot without doubt say that this would not also have
> some effect on some of the other chips which the r8169 driver supports. Was
> there any regression testing on any hw beside the actual target platform?
> 

Yes, tested on two other hw with RTL8168ep/8111ep.

Aaron

> -Stefan
> 



More information about the kernel-team mailing list