mbox series

[0/2] ath25: add 5.4 kernel support (not for merge)

Message ID 20200904002051.3930-1-ryazanov.s.a@gmail.com
Headers show
Series ath25: add 5.4 kernel support (not for merge) | expand

Message

Sergey Ryazanov Sept. 4, 2020, 12:20 a.m. UTC
This series introduces support for 5.4 kernel. I know that this target
was dropped earlier this week. And I think that this was a good
decision. So, this series is not for merging, but for reference if
someone still uses one of these boards and want to create a custom build
with a modern kernel version.

Sergey Ryazanov (2):
  ath25: add kernel 5.4 support
  ath25: eth: fix crash on skb DMA (un-)map

 target/linux/ath25/Makefile                   |  1 +
 .../linux/ath25/{config-4.14 => config-5.4}   | 73 +++++++++++--------
 .../107-ar5312_gpio.patch                     | 20 ++---
 .../108-ar2315_gpio.patch                     | 18 ++---
 .../110-ar2313_ethernet.patch                 | 40 +++++-----
 .../120-spiflash.patch                        |  9 +--
 .../130-watchdog.patch                        |  4 +-
 .../140-redboot_boardconfig.patch             | 14 ++--
 .../141-redboot_partition_scan.patch}         |  8 +-
 .../142-redboot_various_erase_size_fix.patch  | 16 ++--
 .../210-reset_button.patch                    |  0
 .../220-enet_micrel_workaround.patch          | 10 +--
 .../330-board_leds.patch                      |  0
 13 files changed, 112 insertions(+), 101 deletions(-)
 copy target/linux/ath25/{config-4.14 => config-5.4} (76%)
 copy target/linux/ath25/{patches-4.14 => patches-5.4}/107-ar5312_gpio.patch (92%)
 copy target/linux/ath25/{patches-4.14 => patches-5.4}/108-ar2315_gpio.patch (95%)
 copy target/linux/ath25/{patches-4.14 => patches-5.4}/110-ar2313_ethernet.patch (98%)
 copy target/linux/ath25/{patches-4.14 => patches-5.4}/120-spiflash.patch (98%)
 copy target/linux/ath25/{patches-4.14 => patches-5.4}/130-watchdog.patch (98%)
 copy target/linux/ath25/{patches-4.14 => patches-5.4}/140-redboot_boardconfig.patch (81%)
 copy target/linux/{ath79/patches-5.4/408-mtd-redboot_partition_scan.patch => ath25/patches-5.4/141-redboot_partition_scan.patch} (79%)
 copy target/linux/ath25/{patches-4.14 => patches-5.4}/142-redboot_various_erase_size_fix.patch (81%)
 copy target/linux/ath25/{patches-4.14 => patches-5.4}/210-reset_button.patch (100%)
 copy target/linux/ath25/{patches-4.14 => patches-5.4}/220-enet_micrel_workaround.patch (88%)
 copy target/linux/ath25/{patches-4.14 => patches-5.4}/330-board_leds.patch (100%)

Comments

Daniel Golle Sept. 4, 2020, 12:50 a.m. UTC | #1
On Fri, Sep 04, 2020 at 03:20:49AM +0300, Sergey Ryazanov wrote:
> This series introduces support for 5.4 kernel. I know that this target
> was dropped earlier this week. And I think that this was a good
> decision. So, this series is not for merging, but for reference if
> someone still uses one of these boards and want to create a custom build
> with a modern kernel version.

I'd say as you have apprently really managed to get it booting it is
worth reverting the removal and including it.
Sure, a bit late, but there is hardware out there which would benefit
from recent OpenWrt (old ubnt gear, fonera, ... some of them 8/32).
But in terms of resources, ath25 is more advanced than good-old
bcm47xx and also played an important role in OpenWrt history.

Maybe I'm just nostalgic...

> 
> Sergey Ryazanov (2):
>   ath25: add kernel 5.4 support
>   ath25: eth: fix crash on skb DMA (un-)map
> 
>  target/linux/ath25/Makefile                   |  1 +
>  .../linux/ath25/{config-4.14 => config-5.4}   | 73 +++++++++++--------
>  .../107-ar5312_gpio.patch                     | 20 ++---
>  .../108-ar2315_gpio.patch                     | 18 ++---
>  .../110-ar2313_ethernet.patch                 | 40 +++++-----
>  .../120-spiflash.patch                        |  9 +--
>  .../130-watchdog.patch                        |  4 +-
>  .../140-redboot_boardconfig.patch             | 14 ++--
>  .../141-redboot_partition_scan.patch}         |  8 +-
>  .../142-redboot_various_erase_size_fix.patch  | 16 ++--
>  .../210-reset_button.patch                    |  0
>  .../220-enet_micrel_workaround.patch          | 10 +--
>  .../330-board_leds.patch                      |  0
>  13 files changed, 112 insertions(+), 101 deletions(-)
>  copy target/linux/ath25/{config-4.14 => config-5.4} (76%)
>  copy target/linux/ath25/{patches-4.14 => patches-5.4}/107-ar5312_gpio.patch (92%)
>  copy target/linux/ath25/{patches-4.14 => patches-5.4}/108-ar2315_gpio.patch (95%)
>  copy target/linux/ath25/{patches-4.14 => patches-5.4}/110-ar2313_ethernet.patch (98%)
>  copy target/linux/ath25/{patches-4.14 => patches-5.4}/120-spiflash.patch (98%)
>  copy target/linux/ath25/{patches-4.14 => patches-5.4}/130-watchdog.patch (98%)
>  copy target/linux/ath25/{patches-4.14 => patches-5.4}/140-redboot_boardconfig.patch (81%)
>  copy target/linux/{ath79/patches-5.4/408-mtd-redboot_partition_scan.patch => ath25/patches-5.4/141-redboot_partition_scan.patch} (79%)
>  copy target/linux/ath25/{patches-4.14 => patches-5.4}/142-redboot_various_erase_size_fix.patch (81%)
>  copy target/linux/ath25/{patches-4.14 => patches-5.4}/210-reset_button.patch (100%)
>  copy target/linux/ath25/{patches-4.14 => patches-5.4}/220-enet_micrel_workaround.patch (88%)
>  copy target/linux/ath25/{patches-4.14 => patches-5.4}/330-board_leds.patch (100%)
> 
> -- 
> 2.26.2
> 
> 
> _______________________________________________
> openwrt-devel mailing list
> openwrt-devel@lists.openwrt.org
> https://lists.openwrt.org/mailman/listinfo/openwrt-devel
Russell Senior Sept. 4, 2020, 4:14 a.m. UTC | #2
>>>>> "Daniel" == Daniel Golle <daniel@makrotopia.org> writes:

Daniel> On Fri, Sep 04, 2020 at 03:20:49AM +0300, Sergey Ryazanov wrote:
>> This series introduces support for 5.4 kernel. I know that this
>> target was dropped earlier this week. And I think that this was a
>> good decision. So, this series is not for merging, but for reference
>> if someone still uses one of these boards and want to create a custom
>> build with a modern kernel version.

Daniel> I'd say as you have apprently really managed to get it booting
Daniel> it is worth reverting the removal and including it.  Sure, a bit
Daniel> late, but there is hardware out there which would benefit from
Daniel> recent OpenWrt (old ubnt gear, fonera, ... some of them 8/32).
Daniel> But in terms of resources, ath25 is more advanced than good-old
Daniel> bcm47xx and also played an important role in OpenWrt history.

Daniel> Maybe I'm just nostalgic...

+1
Adrian Schmutzler Sept. 4, 2020, 9:25 a.m. UTC | #3
> -----Original Message-----
> From: Daniel Golle [mailto:daniel@makrotopia.org]
> Sent: Freitag, 4. September 2020 02:50
> To: Sergey Ryazanov <ryazanov.s.a@gmail.com>
> Cc: OpenWrt Development List <openwrt-devel@lists.openwrt.org>; Adrian
> Schmutzler <mail@adrianschmutzler.de>
> Subject: Re: [PATCH 0/2] ath25: add 5.4 kernel support (not for merge)
> 
> On Fri, Sep 04, 2020 at 03:20:49AM +0300, Sergey Ryazanov wrote:
> > This series introduces support for 5.4 kernel. I know that this target
> > was dropped earlier this week. And I think that this was a good
> > decision. So, this series is not for merging, but for reference if
> > someone still uses one of these boards and want to create a custom
> > build with a modern kernel version.
> 
> I'd say as you have apprently really managed to get it booting it is worth
> reverting the removal and including it.
> Sure, a bit late, but there is hardware out there which would benefit from
> recent OpenWrt (old ubnt gear, fonera, ... some of them 8/32).
> But in terms of resources, ath25 is more advanced than good-old bcm47xx
> and also played an important role in OpenWrt history.

I'm not opposed to adding it back in general, but from a quick look it appears to me that devices are either at least 4/32 or broken?

So, effectively we would add it just for custom builds and packages.

Best

Adrian

> 
> Maybe I'm just nostalgic...
> 
> >
> > Sergey Ryazanov (2):
> >   ath25: add kernel 5.4 support
> >   ath25: eth: fix crash on skb DMA (un-)map
> >
> >  target/linux/ath25/Makefile                   |  1 +
> >  .../linux/ath25/{config-4.14 => config-5.4}   | 73 +++++++++++--------
> >  .../107-ar5312_gpio.patch                     | 20 ++---
> >  .../108-ar2315_gpio.patch                     | 18 ++---
> >  .../110-ar2313_ethernet.patch                 | 40 +++++-----
> >  .../120-spiflash.patch                        |  9 +--
> >  .../130-watchdog.patch                        |  4 +-
> >  .../140-redboot_boardconfig.patch             | 14 ++--
> >  .../141-redboot_partition_scan.patch}         |  8 +-
> >  .../142-redboot_various_erase_size_fix.patch  | 16 ++--
> >  .../210-reset_button.patch                    |  0
> >  .../220-enet_micrel_workaround.patch          | 10 +--
> >  .../330-board_leds.patch                      |  0
> >  13 files changed, 112 insertions(+), 101 deletions(-)  copy
> > target/linux/ath25/{config-4.14 => config-5.4} (76%)  copy
> > target/linux/ath25/{patches-4.14 => patches-5.4}/107-ar5312_gpio.patch
> > (92%)  copy target/linux/ath25/{patches-4.14 =>
> > patches-5.4}/108-ar2315_gpio.patch (95%)  copy
> > target/linux/ath25/{patches-4.14 =>
> > patches-5.4}/110-ar2313_ethernet.patch (98%)  copy
> > target/linux/ath25/{patches-4.14 => patches-5.4}/120-spiflash.patch
> > (98%)  copy target/linux/ath25/{patches-4.14 =>
> > patches-5.4}/130-watchdog.patch (98%)  copy
> > target/linux/ath25/{patches-4.14 =>
> > patches-5.4}/140-redboot_boardconfig.patch (81%)  copy
> > target/linux/{ath79/patches-5.4/408-mtd-redboot_partition_scan.patch
> > => ath25/patches-5.4/141-redboot_partition_scan.patch} (79%)  copy
> > target/linux/ath25/{patches-4.14 =>
> > patches-5.4}/142-redboot_various_erase_size_fix.patch (81%)  copy
> > target/linux/ath25/{patches-4.14 =>
> > patches-5.4}/210-reset_button.patch (100%)  copy
> > target/linux/ath25/{patches-4.14 =>
> > patches-5.4}/220-enet_micrel_workaround.patch (88%)  copy
> > target/linux/ath25/{patches-4.14 => patches-5.4}/330-board_leds.patch
> > (100%)
> >
> > --
> > 2.26.2
> >
> >
> > _______________________________________________
> > openwrt-devel mailing list
> > openwrt-devel@lists.openwrt.org
> > https://lists.openwrt.org/mailman/listinfo/openwrt-devel
Adrian Schmutzler Sept. 4, 2020, 9:56 a.m. UTC | #4
> Indeed some of the devices are 4/32 or even 4/16 in ath25, but there are a number of 8/32 devices (like Fon2202, Meraki Mini, Ubiquiti PS5, and likely other custom boards). So above the nostalgic feeling, there is some technical reasoning to keep the target around, I'm in favour of that. (Thanks Sergey for finding the cause of the eth crash.)

The non-Ubiquiti are served by the generic image then?

Best

Adrian
Zoltan HERPAI Sept. 4, 2020, 10:26 a.m. UTC | #5
On 09/04/2020 11:56, Adrian Schmutzler wrote:
>> Indeed some of the devices are 4/32 or even 4/16 in ath25, but there are a number of 8/32 devices (like Fon2202, Meraki Mini, Ubiquiti PS5, and likely other custom boards). So above the nostalgic feeling, there is some technical reasoning to keep the target around, I'm in favour of that. (Thanks Sergey for finding the cause of the eth crash.)
> The non-Ubiquiti are served by the generic image then?
>
Correct, most of the devices are supported by the generic image.

Regards,
Zoltan H
Sergey Ryazanov Sept. 4, 2020, 10:47 a.m. UTC | #6
On Fri, Sep 4, 2020 at 12:54 PM Zoltan HERPAI <wigyori@uid0.hu> wrote:
>> On 09/04/2020 11:25, Adrian Schmutzler wrote:
>> Zoltan HERPAI <wigyori@uid0.hu> wrote:
>>> Sure, a bit late, but there is hardware out there which would benefit from
>>> recent OpenWrt (old ubnt gear, fonera, ... some of them 8/32).
>>> But in terms of resources, ath25 is more advanced than good-old bcm47xx
>>> and also played an important role in OpenWrt history.
>>
>> I'm not opposed to adding it back in general, but from a quick look it appears to me that devices are either at least 4/32 or broken?
>>
>> So, effectively we would add it just for custom builds and packages.
>>
> Indeed some of the devices are 4/32 or even 4/16 in ath25, but there are a number of 8/32 devices (like Fon2202, Meraki Mini, Ubiquiti PS5, and likely other custom boards). So above the nostalgic feeling, there is some technical reasoning to keep the target around, I'm in favour of that.

Looks like I was wrong in assuming there are no active users of these boards.

I will send another series that reverts target dropping and updates
the kernel to 5.4.