Netdev List
 help / color / mirror / Atom feed
* RE: rtl_nic: request for rtl8159-1.fw (RTL8159 PHY firmware)
       [not found] ` <7b6ef8fb2c14485380c90f1f7521ea9b@realtek.com>
@ 2026-11-16 10:37   ` oraveeps
  2026-11-16 12:07     ` Birger Koblitz
  2026-11-18  6:58     ` Hayes Wang
  0 siblings, 2 replies; 4+ messages in thread
From: oraveeps @ 2026-11-16 10:37 UTC (permalink / raw)
  To: Hayes Wang
  Cc: javen_xu@realsil.com.cn, mail@birger-koblitz.de,
	linux-firmware@kernel.org, netdev@vger.kernel.org

[Resent as plain text, the lists rejected the first copy as HTML.Sorry for the duplicate.]

On Mon, 16 Nov 2026 at 09:30, Hayes Wang <hayeswang@realtek.com> wrote:
> Hi,
>
> Could you please try out the attached files?
>
> Best Regards,
> Hayes
>
> From: oraveeps@pm.me
> Sent: Sunday, November 8, 2026 5:42 PM
> Subject: rtl_nic: request for rtl8159-1.fw (RTL8159 PHY firmware)
>
> Hi Hayes,
>
> Since commit a51e171e176c ("r8152: Add firmware upload capability for
> RTL8157/RTL8159"), mainline r8152 requests rtl_nic/rtl8159-1.fw for the
> RTL8159 PHY. Without it the link stays at 5 Gbit/s, and linux-firmware
> has no rtl8159-1.fw yet (nor rtl8157-1.fw).
>
> I run three RTL8159 adapters (USB ID 0bda:815a) on kernel 7.2.4. They
> reach 10 Gbit/s only with Realtek's out-of-tree r8152 v2.22.1, which
> carries the PHY patch inside the driver. With rtl8159-1.fw in
> linux-firmware, the in-tree driver could do the same and users could
> drop the out-of-tree module.
>
> Could Realtek submit rtl8159-1.fw, and rtl8157-1.fw if there is one, to
> linux-firmware? I am happy to test a candidate file on these adapters.
>
> Thanks,
> Jonas Autzen

Hi Hayes,

Thanks for the files (rtl8157-1 and rtl8159-1, v1 11/13/26). Mainline
r8152 (v7.2.4 here, same code in v7.4-rc1) rejects them at the first
block: rtl8152_is_fw_mac_ok() has no RTL_VER_16/RTL_VER_17 case for
RTL_FW_PLA or RTL_FW_USB, so rtl8152_check_firmware() fails with
"check USB firmware failed".

An rtl8159-1.fw with only your PHY_SPEED_UP block (checksum
recomputed) loads with the in-tree driver on 7.2.4 and links my RTL8159
adapter (0bda:815a) at 10G.

The series "r8152: split RTL_VER_17 into QFN68 and QFN100 package
variants" (v8) does not change rtl8152_is_fw_mac_ok() either, so the
full files are rejected with it applied too. Is a driver change for the
PLA/USB blocks planned? If not, could rtl8159-1.fw carry only the PHY
block?

Thanks,
Jonas Autzen


On Monday, 16 November 2026 at 09:30, Hayes Wang <hayeswang@realtek.com> wrote:

> Hi,
> 
> Could you please try out the attached files?
> 
> Best Regards,
> 
> Hayes
> 
> > From: oraveeps@pm.me <oraveeps@pm.me>
> > Sent: Sunday, November 8, 2026 5:42 PM
> > To: Hayes Wang <hayeswang@realtek.com>
> > Cc: javen_xu@realsil.com.cn; mail@birger-koblitz.de; linux-firmware@kernel.org; netdev@vger.kernel.org
> > Subject: rtl_nic: request for rtl8159-1.fw (RTL8159 PHY firmware)
> > 
> > External mail : This email originated from outside the organization. Do not reply, click links, or open attachments unless you recognize the sender and know the content is safe.
> > 
> > Hi Hayes,
> > 
> > Since commit a51e171e176c ("r8152: Add firmware upload capability for
> > 
> > RTL8157/RTL8159"), mainline r8152 requests rtl_nic/rtl8159-1.fw for the
> > 
> > RTL8159 PHY. Without it the link stays at 5 Gbit/s, and linux-firmware
> > 
> > has no rtl8159-1.fw yet (nor rtl8157-1.fw).
> > 
> > I run three RTL8159 adapters (USB ID 0bda:815a) on kernel 7.2.4. They
> > 
> > reach 10 Gbit/s only with Realtek's out-of-tree r8152 v2.22.1, which
> > 
> > carries the PHY patch inside the driver. With rtl8159-1.fw in
> > 
> > linux-firmware, the in-tree driver could do the same and users could
> > 
> > drop the out-of-tree module.
> > 
> > Could Realtek submit rtl8159-1.fw, and rtl8157-1.fw if there is one, to
> > 
> > linux-firmware? I am happy to test a candidate file on these adapters.
> > 
> > Thanks,
> > 
> > Jonas Autzen

^ permalink raw reply	[flat|nested] 4+ messages in thread

* Re: rtl_nic: request for rtl8159-1.fw (RTL8159 PHY firmware)
  2026-11-16 10:37   ` rtl_nic: request for rtl8159-1.fw (RTL8159 PHY firmware) oraveeps
@ 2026-11-16 12:07     ` Birger Koblitz
  2026-11-16 15:33       ` oraveeps
  2026-11-18  6:58     ` Hayes Wang
  1 sibling, 1 reply; 4+ messages in thread
From: Birger Koblitz @ 2026-11-16 12:07 UTC (permalink / raw)
  To: oraveeps, Hayes Wang
  Cc: javen_xu@realsil.com.cn, linux-firmware@kernel.org,
	netdev@vger.kernel.org

Could you try:
https://gitlab.com/koblitz-rtlnic/rtlnic_fw
I used that testing the drivers. It is based on the source of the 
out-of-tree driver and creates firmware files compatible to the Linux 
driver. The firmware needs to be generated from the source.

Birger

On 16/11/2026 11:37 am, oraveeps@pm.me wrote:
> [Resent as plain text, the lists rejected the first copy as HTML.Sorry for the duplicate.]
> 
> On Mon, 16 Nov 2026 at 09:30, Hayes Wang <hayeswang@realtek.com> wrote:
>> Hi,
>>
>> Could you please try out the attached files?
>>
>> Best Regards,
>> Hayes
>>
>> From: oraveeps@pm.me
>> Sent: Sunday, November 8, 2026 5:42 PM
>> Subject: rtl_nic: request for rtl8159-1.fw (RTL8159 PHY firmware)
>>
>> Hi Hayes,
>>
>> Since commit a51e171e176c ("r8152: Add firmware upload capability for
>> RTL8157/RTL8159"), mainline r8152 requests rtl_nic/rtl8159-1.fw for the
>> RTL8159 PHY. Without it the link stays at 5 Gbit/s, and linux-firmware
>> has no rtl8159-1.fw yet (nor rtl8157-1.fw).
>>
>> I run three RTL8159 adapters (USB ID 0bda:815a) on kernel 7.2.4. They
>> reach 10 Gbit/s only with Realtek's out-of-tree r8152 v2.22.1, which
>> carries the PHY patch inside the driver. With rtl8159-1.fw in
>> linux-firmware, the in-tree driver could do the same and users could
>> drop the out-of-tree module.
>>
>> Could Realtek submit rtl8159-1.fw, and rtl8157-1.fw if there is one, to
>> linux-firmware? I am happy to test a candidate file on these adapters.
>>
>> Thanks,
>> Jonas Autzen
> 
> Hi Hayes,
> 
> Thanks for the files (rtl8157-1 and rtl8159-1, v1 11/13/26). Mainline
> r8152 (v7.2.4 here, same code in v7.4-rc1) rejects them at the first
> block: rtl8152_is_fw_mac_ok() has no RTL_VER_16/RTL_VER_17 case for
> RTL_FW_PLA or RTL_FW_USB, so rtl8152_check_firmware() fails with
> "check USB firmware failed".
> 
> An rtl8159-1.fw with only your PHY_SPEED_UP block (checksum
> recomputed) loads with the in-tree driver on 7.2.4 and links my RTL8159
> adapter (0bda:815a) at 10G.
> 
> The series "r8152: split RTL_VER_17 into QFN68 and QFN100 package
> variants" (v8) does not change rtl8152_is_fw_mac_ok() either, so the
> full files are rejected with it applied too. Is a driver change for the
> PLA/USB blocks planned? If not, could rtl8159-1.fw carry only the PHY
> block?
> 
> Thanks,
> Jonas Autzen
> 
> 
> On Monday, 16 November 2026 at 09:30, Hayes Wang <hayeswang@realtek.com> wrote:
> 
>> Hi,
>>
>> Could you please try out the attached files?
>>
>> Best Regards,
>>
>> Hayes
>>
>>> From: oraveeps@pm.me <oraveeps@pm.me>
>>> Sent: Sunday, November 8, 2026 5:42 PM
>>> To: Hayes Wang <hayeswang@realtek.com>
>>> Cc: javen_xu@realsil.com.cn; mail@birger-koblitz.de; linux-firmware@kernel.org; netdev@vger.kernel.org
>>> Subject: rtl_nic: request for rtl8159-1.fw (RTL8159 PHY firmware)
>>>
>>> External mail : This email originated from outside the organization. Do not reply, click links, or open attachments unless you recognize the sender and know the content is safe.
>>>
>>> Hi Hayes,
>>>
>>> Since commit a51e171e176c ("r8152: Add firmware upload capability for
>>>
>>> RTL8157/RTL8159"), mainline r8152 requests rtl_nic/rtl8159-1.fw for the
>>>
>>> RTL8159 PHY. Without it the link stays at 5 Gbit/s, and linux-firmware
>>>
>>> has no rtl8159-1.fw yet (nor rtl8157-1.fw).
>>>
>>> I run three RTL8159 adapters (USB ID 0bda:815a) on kernel 7.2.4. They
>>>
>>> reach 10 Gbit/s only with Realtek's out-of-tree r8152 v2.22.1, which
>>>
>>> carries the PHY patch inside the driver. With rtl8159-1.fw in
>>>
>>> linux-firmware, the in-tree driver could do the same and users could
>>>
>>> drop the out-of-tree module.
>>>
>>> Could Realtek submit rtl8159-1.fw, and rtl8157-1.fw if there is one, to
>>>
>>> linux-firmware? I am happy to test a candidate file on these adapters.
>>>
>>> Thanks,
>>>
>>> Jonas Autzen


^ permalink raw reply	[flat|nested] 4+ messages in thread

* Re: rtl_nic: request for rtl8159-1.fw (RTL8159 PHY firmware)
  2026-11-16 12:07     ` Birger Koblitz
@ 2026-11-16 15:33       ` oraveeps
  0 siblings, 0 replies; 4+ messages in thread
From: oraveeps @ 2026-11-16 15:33 UTC (permalink / raw)
  To: Birger Koblitz
  Cc: Hayes Wang, javen_xu@realsil.com.cn, linux-firmware@kernel.org,
	netdev@vger.kernel.org

Hi Birger,

thanks, that works. I built rtl8159-1.fw the way rtl_fw.c does, but from the ram17 array of the out-of-tree driver 2.22.1 (code version 0x0022) instead of the 2.21.4 one in your repo. With it the in-tree r8152 on 7.2.4 loads the PHY code after a full power cycle without errors, and the RTL8159 links at 10 Gbit/s.

The PHY_SPEED_UP block in Realtek's candidate file
(pwr_on_patch_v22_pr_test_ramcode_usb_20260428) matches the 2.22.1 ram17 for its first 16536 bytes, then ends. It lacks the write of the PHY firmware version (SRAM 0x801e), which rtl_ram_code_speed_up() checks after loading, and the parameter table writes that follow it.

Best regards
Jonas

On Monday, 16 November 2026 at 13:07, Birger Koblitz <mail@birger-koblitz.de> wrote:

> Could you try:
> https://gitlab.com/koblitz-rtlnic/rtlnic_fw
> I used that testing the drivers. It is based on the source of the
> out-of-tree driver and creates firmware files compatible to the Linux
> driver. The firmware needs to be generated from the source.
> 
> Birger
> 
> On 16/11/2026 11:37 am, oraveeps@pm.me wrote:
> > [Resent as plain text, the lists rejected the first copy as HTML.Sorry for the duplicate.]
> >
> > On Mon, 16 Nov 2026 at 09:30, Hayes Wang <hayeswang@realtek.com> wrote:
> >> Hi,
> >>
> >> Could you please try out the attached files?
> >>
> >> Best Regards,
> >> Hayes
> >>
> >> From: oraveeps@pm.me
> >> Sent: Sunday, November 8, 2026 5:42 PM
> >> Subject: rtl_nic: request for rtl8159-1.fw (RTL8159 PHY firmware)
> >>
> >> Hi Hayes,
> >>
> >> Since commit a51e171e176c ("r8152: Add firmware upload capability for
> >> RTL8157/RTL8159"), mainline r8152 requests rtl_nic/rtl8159-1.fw for the
> >> RTL8159 PHY. Without it the link stays at 5 Gbit/s, and linux-firmware
> >> has no rtl8159-1.fw yet (nor rtl8157-1.fw).
> >>
> >> I run three RTL8159 adapters (USB ID 0bda:815a) on kernel 7.2.4. They
> >> reach 10 Gbit/s only with Realtek's out-of-tree r8152 v2.22.1, which
> >> carries the PHY patch inside the driver. With rtl8159-1.fw in
> >> linux-firmware, the in-tree driver could do the same and users could
> >> drop the out-of-tree module.
> >>
> >> Could Realtek submit rtl8159-1.fw, and rtl8157-1.fw if there is one, to
> >> linux-firmware? I am happy to test a candidate file on these adapters.
> >>
> >> Thanks,
> >> Jonas Autzen
> >
> > Hi Hayes,
> >
> > Thanks for the files (rtl8157-1 and rtl8159-1, v1 11/13/26). Mainline
> > r8152 (v7.2.4 here, same code in v7.4-rc1) rejects them at the first
> > block: rtl8152_is_fw_mac_ok() has no RTL_VER_16/RTL_VER_17 case for
> > RTL_FW_PLA or RTL_FW_USB, so rtl8152_check_firmware() fails with
> > "check USB firmware failed".
> >
> > An rtl8159-1.fw with only your PHY_SPEED_UP block (checksum
> > recomputed) loads with the in-tree driver on 7.2.4 and links my RTL8159
> > adapter (0bda:815a) at 10G.
> >
> > The series "r8152: split RTL_VER_17 into QFN68 and QFN100 package
> > variants" (v8) does not change rtl8152_is_fw_mac_ok() either, so the
> > full files are rejected with it applied too. Is a driver change for the
> > PLA/USB blocks planned? If not, could rtl8159-1.fw carry only the PHY
> > block?
> >
> > Thanks,
> > Jonas Autzen
> >
> >
> > On Monday, 16 November 2026 at 09:30, Hayes Wang <hayeswang@realtek.com> wrote:
> >
> >> Hi,
> >>
> >> Could you please try out the attached files?
> >>
> >> Best Regards,
> >>
> >> Hayes
> >>
> >>> From: oraveeps@pm.me <oraveeps@pm.me>
> >>> Sent: Sunday, November 8, 2026 5:42 PM
> >>> To: Hayes Wang <hayeswang@realtek.com>
> >>> Cc: javen_xu@realsil.com.cn; mail@birger-koblitz.de; linux-firmware@kernel.org; netdev@vger.kernel.org
> >>> Subject: rtl_nic: request for rtl8159-1.fw (RTL8159 PHY firmware)
> >>>
> >>> External mail : This email originated from outside the organization. Do not reply, click links, or open attachments unless you recognize the sender and know the content is safe.
> >>>
> >>> Hi Hayes,
> >>>
> >>> Since commit a51e171e176c ("r8152: Add firmware upload capability for
> >>>
> >>> RTL8157/RTL8159"), mainline r8152 requests rtl_nic/rtl8159-1.fw for the
> >>>
> >>> RTL8159 PHY. Without it the link stays at 5 Gbit/s, and linux-firmware
> >>>
> >>> has no rtl8159-1.fw yet (nor rtl8157-1.fw).
> >>>
> >>> I run three RTL8159 adapters (USB ID 0bda:815a) on kernel 7.2.4. They
> >>>
> >>> reach 10 Gbit/s only with Realtek's out-of-tree r8152 v2.22.1, which
> >>>
> >>> carries the PHY patch inside the driver. With rtl8159-1.fw in
> >>>
> >>> linux-firmware, the in-tree driver could do the same and users could
> >>>
> >>> drop the out-of-tree module.
> >>>
> >>> Could Realtek submit rtl8159-1.fw, and rtl8157-1.fw if there is one, to
> >>>
> >>> linux-firmware? I am happy to test a candidate file on these adapters.
> >>>
> >>> Thanks,
> >>>
> >>> Jonas Autzen
> 
>

^ permalink raw reply	[flat|nested] 4+ messages in thread

* RE: rtl_nic: request for rtl8159-1.fw (RTL8159 PHY firmware)
  2026-11-16 10:37   ` rtl_nic: request for rtl8159-1.fw (RTL8159 PHY firmware) oraveeps
  2026-11-16 12:07     ` Birger Koblitz
@ 2026-11-18  6:58     ` Hayes Wang
  1 sibling, 0 replies; 4+ messages in thread
From: Hayes Wang @ 2026-11-18  6:58 UTC (permalink / raw)
  To: oraveeps@pm.me, Chih Kai Hsu
  Cc: javen_xu@realsil.com.cn, mail@birger-koblitz.de,
	linux-firmware@kernel.org, netdev@vger.kernel.org

oraveeps@pm.me <oraveeps@pm.me>
> Sent: Monday, November 16, 2026 6:38 PM
[...]
> Hi Hayes,
> 
> Thanks for the files (rtl8157-1 and rtl8159-1, v1 11/13/26). Mainline
> r8152 (v7.2.4 here, same code in v7.4-rc1) rejects them at the first
> block: rtl8152_is_fw_mac_ok() has no RTL_VER_16/RTL_VER_17 case for
> RTL_FW_PLA or RTL_FW_USB, so rtl8152_check_firmware() fails with
> "check USB firmware failed".
> 
> An rtl8159-1.fw with only your PHY_SPEED_UP block (checksum
> recomputed) loads with the in-tree driver on 7.2.4 and links my RTL8159
> adapter (0bda:815a) at 10G.
> 
> The series "r8152: split RTL_VER_17 into QFN68 and QFN100 package
> variants" (v8) does not change rtl8152_is_fw_mac_ok() either, so the
> full files are rejected with it applied too. Is a driver change for the
> PLA/USB blocks planned? If not, could rtl8159-1.fw carry only the PHY
> block?

Chih-Kai will update the driver.
However, some patches are still pending review,
so we need to wait until they are merged before proceeding.


Best Regards,
Hayes

> Thanks,
> Jonas Autzen
> 
> 
> On Monday, 16 November 2026 at 09:30, Hayes Wang
> <hayeswang@realtek.com> wrote:
> 
> > Hi,
> >
> > Could you please try out the attached files?
> >
> > Best Regards,
> >
> > Hayes
> >
> > > From: oraveeps@pm.me <oraveeps@pm.me>
> > > Sent: Sunday, November 8, 2026 5:42 PM
> > > To: Hayes Wang <hayeswang@realtek.com>
> > > Cc: javen_xu@realsil.com.cn; mail@birger-koblitz.de;
> linux-firmware@kernel.org; netdev@vger.kernel.org
> > > Subject: rtl_nic: request for rtl8159-1.fw (RTL8159 PHY firmware)
> > >
> > > External mail : This email originated from outside the organization. Do not
> reply, click links, or open attachments unless you recognize the sender and
> know the content is safe.
> > >
> > > Hi Hayes,
> > >
> > > Since commit a51e171e176c ("r8152: Add firmware upload capability for
> > >
> > > RTL8157/RTL8159"), mainline r8152 requests rtl_nic/rtl8159-1.fw for the
> > >
> > > RTL8159 PHY. Without it the link stays at 5 Gbit/s, and linux-firmware
> > >
> > > has no rtl8159-1.fw yet (nor rtl8157-1.fw).
> > >
> > > I run three RTL8159 adapters (USB ID 0bda:815a) on kernel 7.2.4. They
> > >
> > > reach 10 Gbit/s only with Realtek's out-of-tree r8152 v2.22.1, which
> > >
> > > carries the PHY patch inside the driver. With rtl8159-1.fw in
> > >
> > > linux-firmware, the in-tree driver could do the same and users could
> > >
> > > drop the out-of-tree module.
> > >
> > > Could Realtek submit rtl8159-1.fw, and rtl8157-1.fw if there is one, to
> > >
> > > linux-firmware? I am happy to test a candidate file on these adapters.
> > >
> > > Thanks,
> > >
> > > Jonas Autzen

^ permalink raw reply	[flat|nested] 4+ messages in thread

end of thread, other threads:[~2026-11-18  6:58 UTC | newest]

Thread overview: 4+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
     [not found] <PfNGg6VjMdDzTCX2LETCm65NKSR0af1EG7ZX5CI-7Fef0x8NE2MnZ8fNtgr1Hm86sNjqHXcNyyqIbbeZL5jDu7URTqMygnyj4ulG00NxR7U=@pm.me>
     [not found] ` <7b6ef8fb2c14485380c90f1f7521ea9b@realtek.com>
2026-11-16 10:37   ` rtl_nic: request for rtl8159-1.fw (RTL8159 PHY firmware) oraveeps
2026-11-16 12:07     ` Birger Koblitz
2026-11-16 15:33       ` oraveeps
2026-11-18  6:58     ` Hayes Wang

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox