r8152: add vendor/device ID pair for Microsoft Devkit
Subsystems (4)
PATCH 1/1 r8152: add vendor/device ID pair for Microsoft Devkit
The Microsoft Devkit 2023 is a an ARM64 based machine featuring a Realtek 8153 USB3.0-to-GBit Ethernet adapter. As in their other machines, Microsoft uses a custom USB device ID.
Add the respective ID values to the driver. This makes Ethernet work on the MS Devkit device. The chip has been visually confirmed to be a RTL8153.
--- drivers/net/usb/r8152.c | 1 + 1 file changed, 1 insertion(+)
drivers/net/usb/r8152.c +1 -0
@@ -9836,6 +9836,7 @@ static const struct usb_device_id rtl8152_table[] = {REALTEK_USB_DEVICE (VENDOR_ID_MICROSOFT ,0x07ab ),REALTEK_USB_DEVICE (VENDOR_ID_MICROSOFT ,0x07c6 ),REALTEK_USB_DEVICE (VENDOR_ID_MICROSOFT ,0x0927 ), +REALTEK_USB_DEVICE (VENDOR_ID_MICROSOFT ,0x0c5e ),REALTEK_USB_DEVICE (VENDOR_ID_SAMSUNG ,0xa101 ),REALTEK_USB_DEVICE (VENDOR_ID_LENOVO ,0x304f ), REALTEK_USB_DEVICE(VENDOR_ID_LENOVO, 0x3054),
Signature · 1 line
2.25.1
Jakub Kicinski
maintainerQuoted 7 lines · On Wed, 11 Jan 2023 13:32:28 +0000 Andre Przywara wrote:
The Microsoft Devkit 2023 is a an ARM64 based machine featuring a Realtek 8153 USB3.0-to-GBit Ethernet adapter. As in their other machines, Microsoft uses a custom USB device ID.
Add the respective ID values to the driver. This makes Ethernet work on the MS Devkit device. The chip has been visually confirmed to be a RTL8153.
Hm, we have a patch in net-next which reformats the entries:ec51fbd1b8a2bca2948dede99c14ec63dc57ff6b
Would you like this ID to be also added in stable? We could just apply it to net, and deal with the conflict locally. But if you don't care about older kernels then better if you rebase.
Quoted 9 lines · Jakub Kicinski
Quoted 7 lines · On Wed, 11 Jan 2023 13:32:28 +0000 Andre Przywara wrote:
The Microsoft Devkit 2023 is a an ARM64 based machine featuring a Realtek 8153 USB3.0-to-GBit Ethernet adapter. As in their other machines, Microsoft uses a custom USB device ID.
Add the respective ID values to the driver. This makes Ethernet work on the MS Devkit device. The chip has been visually confirmed to be a RTL8153.
Quoted 6 lines
Hm, we have a patch in net-next which reformats the entries:
ec51fbd1b8a2bca2948dede99c14ec63dc57ff6bWould you like this ID to be also added in stable? We could just apply it to net, and deal with the conflict locally. But if you don't care about older kernels then better if you rebase.
And now I started worrying about consequences of that reformatting... Maybe I didn't give this enough thought?
Please let me know if you prefer to have the old macro name back. We can avoid reformatting the list.
Bjørn
Greg KH
maintainerQuoted 13 lines · Bjørn Mork
Quoted 12 lines · Jakub Kicinski
Quoted 7 lines · On Wed, 11 Jan 2023 13:32:28 +0000 Andre Przywara wrote:
The Microsoft Devkit 2023 is a an ARM64 based machine featuring a Realtek 8153 USB3.0-to-GBit Ethernet adapter. As in their other machines, Microsoft uses a custom USB device ID.
Add the respective ID values to the driver. This makes Ethernet work on the MS Devkit device. The chip has been visually confirmed to be a RTL8153.
Hm, we have a patch in net-next which reformats the entries:
ec51fbd1b8a2bca2948dede99c14ec63dc57ff6b
Quoted 6 lines
Quoted 3 lines
Would you like this ID to be also added in stable? We could just apply it to net, and deal with the conflict locally. But if you don't care about older kernels then better if you rebase.
And now I started worrying about consequences of that reformatting... Maybe I didn't give this enough thought?
Just send a reformatting patch for stable as well. I've taken patches like that many times for other drivers/subsystems to make backports trivial.
thanks,
greg k-h
PATCH 1/1 r8152; preserve device list format
In reply to Bjørn MorkThis is a partial revert of commit ec51fbd1b8a2 ("r8152:
add USB device driver for config selection")
Keep a simplified version of the REALTEK_USB_DEVICE macro to avoid unnecessary reformatting of the device list. This makes new device ID additions apply cleanly across driver versions.
---
The patch in
https://lore.kernel.org/lkml/[email protected]/
will apply cleanly on top of this.
This fix will also prevent a lot of stable backporting hassle.
Sorry for not thinking the change completely through before submitting. I should never have touched the rtl8152_table.
Bjørn
drivers/net/usb/r8152.c | 48 +++++++++++++++++++++-------------------- 1 file changed, 25 insertions(+), 23 deletions(-)
drivers/net/usb/r8152.c +25 -23
@@ -9817,34 +9817,36 @@ static void rtl8152_disconnect(struct usb_interface *intf) } } +#define REALTEK_USB_DEVICE (vend, prod) { USB_DEVICE(vend, prod) } +/* table of devices that work with this driver */ static const struct usb_device_id rtl8152_table[] = {/* Realtek */ - {USB_DEVICE (VENDOR_ID_REALTEK ,0x8050 ) }, - {USB_DEVICE (VENDOR_ID_REALTEK ,0x8053 ) }, - {USB_DEVICE (VENDOR_ID_REALTEK ,0x8152 ) }, - {USB_DEVICE (VENDOR_ID_REALTEK ,0x8153 ) }, - {USB_DEVICE (VENDOR_ID_REALTEK ,0x8155 ) }, - {USB_DEVICE (VENDOR_ID_REALTEK ,0x8156 ) }, +REALTEK_USB_DEVICE (VENDOR_ID_REALTEK ,0x8050 ), +REALTEK_USB_DEVICE (VENDOR_ID_REALTEK ,0x8053 ), +REALTEK_USB_DEVICE (VENDOR_ID_REALTEK ,0x8152 ), +REALTEK_USB_DEVICE (VENDOR_ID_REALTEK ,0x8153 ), +REALTEK_USB_DEVICE (VENDOR_ID_REALTEK ,0x8155 ), +REALTEK_USB_DEVICE (VENDOR_ID_REALTEK ,0x8156 ),/* Microsoft */ - {USB_DEVICE (VENDOR_ID_MICROSOFT ,0x07ab ) }, - {USB_DEVICE (VENDOR_ID_MICROSOFT ,0x07c6 ) }, - {USB_DEVICE (VENDOR_ID_MICROSOFT ,0x0927 ) }, - {USB_DEVICE (VENDOR_ID_SAMSUNG ,0xa101 ) }, - {USB_DEVICE (VENDOR_ID_LENOVO ,0x304f ) }, - {USB_DEVICE (VENDOR_ID_LENOVO ,0x3054 ) }, - {USB_DEVICE (VENDOR_ID_LENOVO ,0x3062 ) }, - {USB_DEVICE (VENDOR_ID_LENOVO ,0x3069 ) }, - {USB_DEVICE (VENDOR_ID_LENOVO ,0x3082 ) }, - {USB_DEVICE (VENDOR_ID_LENOVO ,0x7205 ) }, - {USB_DEVICE (VENDOR_ID_LENOVO ,0x720c ) }, - {USB_DEVICE (VENDOR_ID_LENOVO ,0x7214 ) }, - {USB_DEVICE (VENDOR_ID_LENOVO ,0x721e ) }, - {USB_DEVICE (VENDOR_ID_LENOVO ,0xa387 ) }, - {USB_DEVICE (VENDOR_ID_LINKSYS ,0x0041 ) }, - {USB_DEVICE (VENDOR_ID_NVIDIA ,0x09ff ) }, - {USB_DEVICE (VENDOR_ID_TPLINK ,0x0601 ) }, +REALTEK_USB_DEVICE (VENDOR_ID_MICROSOFT ,0x07ab ), +REALTEK_USB_DEVICE (VENDOR_ID_MICROSOFT ,0x07c6 ), +REALTEK_USB_DEVICE (VENDOR_ID_MICROSOFT ,0x0927 ), +REALTEK_USB_DEVICE (VENDOR_ID_SAMSUNG ,0xa101 ), +REALTEK_USB_DEVICE (VENDOR_ID_LENOVO ,0x304f ), +REALTEK_USB_DEVICE (VENDOR_ID_LENOVO ,0x3054 ), +REALTEK_USB_DEVICE (VENDOR_ID_LENOVO ,0x3062 ), +REALTEK_USB_DEVICE (VENDOR_ID_LENOVO ,0x3069 ), +REALTEK_USB_DEVICE (VENDOR_ID_LENOVO ,0x3082 ), +REALTEK_USB_DEVICE (VENDOR_ID_LENOVO ,0x7205 ), +REALTEK_USB_DEVICE (VENDOR_ID_LENOVO ,0x720c ), +REALTEK_USB_DEVICE (VENDOR_ID_LENOVO ,0x7214 ), +REALTEK_USB_DEVICE (VENDOR_ID_LENOVO ,0x721e ), +REALTEK_USB_DEVICE (VENDOR_ID_LENOVO ,0xa387 ), +REALTEK_USB_DEVICE (VENDOR_ID_LINKSYS ,0x0041 ), +REALTEK_USB_DEVICE (VENDOR_ID_NVIDIA ,0x09ff ), +REALTEK_USB_DEVICE (VENDOR_ID_TPLINK ,0x0601 ), {} };
Signature · 1 line
2.30.2
Re: [PATCH net-next] r8152; preserve device list format
Quoted 8 lines · Bjørn Mork
This is a partial revert of commit
ec51fbd1b8a2("r8152:
add USB device driver for config selection")Keep a simplified version of the
REALTEK_USB_DEVICEmacro to avoid unnecessary reformatting of the device list. This makes new device ID additions apply cleanly across driver versions.
Quoted 8 lines
Fixes ec51fbd1b8a2 ("r8152: add USB device driver for config selection")Signed-off-by Bjørn Mork <[email protected]>---
The patch in
https://lore.kernel.org/lkml/[email protected]/
will apply cleanly on top of this.This fix will also prevent a lot of stable backporting hassle.
No need for this, just backport the original change to older kernels and all will be fine.
Don't live with stuff you don't want to because of stable kernels, that's not how this whole process works at all :)
thanks,
greg k-h
Re: [PATCH net-next] r8152; preserve device list format
Quoted 16 lines · Greg KH
Quoted 15 lines · Bjørn Mork
This is a partial revert of commit
ec51fbd1b8a2("r8152:
add USB device driver for config selection")Keep a simplified version of the
REALTEK_USB_DEVICEmacro to avoid unnecessary reformatting of the device list. This makes new device ID additions apply cleanly across driver versions.Fixes ec51fbd1b8a2 ("r8152: add USB device driver for config selection")Signed-off-by Bjørn Mork <[email protected]>---
The patch in
https://lore.kernel.org/lkml/[email protected]/
will apply cleanly on top of this.
Quoted 7 lines
Quoted 1 line
This fix will also prevent a lot of stable backporting hassle.
No need for this, just backport the original change to older kernels and all will be fine.
Don't live with stuff you don't want to because of stable kernels, that's not how this whole process works at all :)
OK, thanks. Will prepare a patch for stable instead then.
But I guess the original patch is unacceptable for stable as-is? It changes how Linux react to these devces, and includes a completely new USB device driver (i.e not interface driver).
Bjørn
Re: [PATCH net-next] r8152; preserve device list format
Quoted 22 lines · Bjørn Mork
Quoted 21 lines · Greg KH
Quoted 8 lines · Bjørn Mork
This is a partial revert of commit
ec51fbd1b8a2("r8152:
add USB device driver for config selection")Keep a simplified version of the
REALTEK_USB_DEVICEmacro to avoid unnecessary reformatting of the device list. This makes new device ID additions apply cleanly across driver versions.Quoted 8 lines
Fixes ec51fbd1b8a2 ("r8152: add USB device driver for config selection")Signed-off-by Bjørn Mork <[email protected]>---
The patch in
https://lore.kernel.org/lkml/[email protected]/
will apply cleanly on top of this.This fix will also prevent a lot of stable backporting hassle.
No need for this, just backport the original change to older kernels and all will be fine.
Quoted 8 lines
Quoted 2 lines
Don't live with stuff you don't want to because of stable kernels, that's not how this whole process works at all :)
OK, thanks. Will prepare a patch for stable instead then.
But I guess the original patch is unacceptable for stable as-is? It changes how Linux react to these devces, and includes a completely new USB device driver (i.e not interface driver).
That's up to you all. We don't add new support for new hardware to older kernels _UNLESS_ it's just a new device id. Otherwise it's just better to tell people to upgrade to the newer kernel.
If you split things up and added a whole new driver, then just leave it alone, no need to backport anything, sorry, I didn't realize that.
greg k-h
Re: [PATCH net-next] r8152; preserve device list format
In reply to Bjørn MorkQuoted 5 lines · Bjørn Mork
Quoted 3 lines · Greg KH
No need for this, just backport the original change to older kernels and all will be fine.
Quoted 8 lines
Quoted 2 lines
Don't live with stuff you don't want to because of stable kernels, that's not how this whole process works at all :)
OK, thanks. Will prepare a patch for stable instead then.
But I guess the original patch is unacceptable for stable as-is? It changes how Linux react to these devces, and includes a completely new USB device driver (i.e not interface driver).
Doh! I gotta start thinking before I send email. Will start right after sending this one ;-)
We cannot backport the device-id table change to stable without taking the rest of the patch. The strategy used by the old driver needs two entries per device ID, which is why the macro was there.
So the question is: Can commit ec51fbd1b8a2 ("r8152: add USB device driver for config selection") be accepted in stable?
( Direct link for convenience since it's not yet in mainline:
https://git.kernel.org/pub/scm/linux/kernel/git/netdev/net-next.git/commit/drivers/net/usb/r8152.c?id=ec51fbd1b8a2bca2948dede99c14ec63dc57ff6b
)
This is not within the rules as I read them, but it's your call...
Bjørn
Re: [PATCH net-next] r8152; preserve device list format
Quoted 22 lines · Bjørn Mork
Quoted 5 lines · Bjørn Mork
Quoted 3 lines · Greg KH
No need for this, just backport the original change to older kernels and all will be fine.
Quoted 8 lines
Quoted 2 lines
Don't live with stuff you don't want to because of stable kernels, that's not how this whole process works at all :)
OK, thanks. Will prepare a patch for stable instead then.
But I guess the original patch is unacceptable for stable as-is? It changes how Linux react to these devces, and includes a completely new USB device driver (i.e not interface driver).
Doh! I gotta start thinking before I send email. Will start right after sending this one ;-)
We cannot backport the device-id table change to stable without taking the rest of the patch. The strategy used by the old driver needs two entries per device ID, which is why the macro was there.
Quoted 8 lines
So the question is: Can commit
ec51fbd1b8a2("r8152: add USB device driver for config selection") be accepted in stable?( Direct link for convenience since it's not yet in mainline:
https://git.kernel.org/pub/scm/linux/kernel/git/netdev/net-next.git/commit/drivers/net/usb/r8152.c?id=ec51fbd1b8a2bca2948dede99c14ec63dc57ff6b
)This is not within the rules as I read them, but it's your call...
Ah, yeah, that's simple enough, I'd take it if you send it to me :)
greg k-h
Re: [PATCH net-next] r8152; preserve device list format
Quoted 26 lines · Greg KH
Quoted 25 lines · Bjørn Mork
Quoted 5 lines · Bjørn Mork
Quoted 3 lines · Greg KH
No need for this, just backport the original change to older kernels and all will be fine.
Quoted 8 lines
Quoted 2 lines
Don't live with stuff you don't want to because of stable kernels, that's not how this whole process works at all :)
OK, thanks. Will prepare a patch for stable instead then.
But I guess the original patch is unacceptable for stable as-is? It changes how Linux react to these devces, and includes a completely new USB device driver (i.e not interface driver).
Doh! I gotta start thinking before I send email. Will start right after sending this one ;-)
We cannot backport the device-id table change to stable without taking the rest of the patch. The strategy used by the old driver needs two entries per device ID, which is why the macro was there.
So the question is: Can commit
ec51fbd1b8a2("r8152: add USB device driver for config selection") be accepted in stable?
Quoted 7 lines
Quoted 5 lines
( Direct link for convenience since it's not yet in mainline:
https://git.kernel.org/pub/scm/linux/kernel/git/netdev/net-next.git/commit/drivers/net/usb/r8152.c?id=ec51fbd1b8a2bca2948dede99c14ec63dc57ff6b
)This is not within the rules as I read them, but it's your call...
Ah, yeah, that's simple enough, I'd take it if you send it to me :)
Great!
There is no point backporting to anything older than v5.15 since the patch depend on significant driver changes between v5.10 and v5.15. The good news is that those changes also modified the macro in question so any device ID patch for v5.10 or older will have to be fixed up in any case. So we don't lose anything by ignoring the older longterm kernels here.
IIUC the special netdev stable rules are gone. But if this is going to stable, then I believe it still has to go to "net" first.
David/Jakub - Would you please pick
ec51fbd1b8a2 ("r8152: add USB device driver for config selection") 69649ef84053 ("cdc_ether: no need to blacklist any r8152 devices")
from net-next to net? With a "CC: stable" preferably. Or do you prefer some other solution?
Bjørn
Jakub Kicinski
maintainerRe: [PATCH net-next] r8152; preserve device list format
Quoted 10 lines · On Fri, 13 Jan 2023 11:16:47 +0100 Bjørn Mork wrote:
There is no point backporting to anything older than v5.15 since the patch depend on significant driver changes between v5.10 and v5.15. The good news is that those changes also modified the macro in question so any device ID patch for v5.10 or older will have to be fixed up in any case. So we don't lose anything by ignoring the older longterm kernels here.
IIUC the special netdev stable rules are gone. But if this is going to stable, then I believe it still has to go to "net" first.
Quoted 7 lines
David/Jakub - Would you please pick
ec51fbd1b8a2("r8152: add USB device driver for config selection")69649ef84053("cdc_ether: no need to blacklist any r8152 devices")from net-next to net? With a "CC: stable" preferably. Or do you prefer some other solution?
Well.. we already shipped the patch from this thread as is to Linus. Greg will be able to take be53771c87f4 into stable directly, with no dependencies.
And now the refactoring won't cherry-pick cleanly :( Maybe let's leave it be?
I'll keep in mind that Greg is okay with taking this sort of refactoring in in the future. I made an unnecessary commotion here.
On Wed, 11 Jan 2023 21:31:43 -0800
Jakub Kicinski <[email protected]> wrote:
Hi,
Quoted 9 lines · On Wed, 11 Jan 2023 13:32:28 +0000 Andre Przywara wrote:
Quoted 7 lines
The Microsoft Devkit 2023 is a an ARM64 based machine featuring a Realtek 8153 USB3.0-to-GBit Ethernet adapter. As in their other machines, Microsoft uses a custom USB device ID.
Add the respective ID values to the driver. This makes Ethernet work on the MS Devkit device. The chip has been visually confirmed to be a RTL8153.
Quoted 6 lines
Hm, we have a patch in net-next which reformats the entries:
ec51fbd1b8a2bca2948dede99c14ec63dc57ff6bWould you like this ID to be also added in stable? We could just apply it to net, and deal with the conflict locally. But if you don't care about older kernels then better if you rebase.
Stable would be nice, but only to v6.1. I think I don't care about older kernels.
So what about if I resend this one here, based on top of the reformat patch, with a:
Cc: <[email protected]> # 6.1.x
line in there, and then reply to the email that the automatic backport failed, with a tailored patch for v6.1?
Alternatively I can send an explicit stable backport email once this one is merged.
Cheers,
Andre
Paolo Abeni
maintainerQuoted 8 lines · Andre Przywara
Quoted 6 lines · On Wed, 11 Jan 2023 21:31:43 -0800 Jakub Kicinski <[email protected]> wrote:
Hm, we have a patch in net-next which reformats the entries:
ec51fbd1b8a2bca2948dede99c14ec63dc57ff6bWould you like this ID to be also added in stable? We could just apply it to net, and deal with the conflict locally. But if you don't care about older kernels then better if you rebase.
Quoted 9 lines
Stable would be nice, but only to v6.1. I think I don't care about older kernels.
So what about if I resend this one here, based on top of the reformat patch, with a:
Cc: <[email protected]> # 6.1.x
line in there, and then reply to the email that the automatic backport failed, with a tailored patch for v6.1?
Alternatively I can send an explicit stable backport email once this one is merged.
Note that we can merge this kind of changes via the -net tree. No repost will be needed. We can merge it as is on -net and you can follow the option 2 from the stable kernel rules doc, with no repost nor additional mangling for stable will be needed.
If you are ok with the above let me know.
Thanks,
Paolo
On Thu, 12 Jan 2023 12:39:01 +0100
Paolo Abeni <[email protected]> wrote:
Hi,
Quoted 19 lines · Andre Przywara
Quoted 8 lines · On Wed, 11 Jan 2023 21:31:43 -0800 Jakub Kicinski <[email protected]> wrote:
Quoted 6 lines
Hm, we have a patch in net-next which reformats the entries:
ec51fbd1b8a2bca2948dede99c14ec63dc57ff6bWould you like this ID to be also added in stable? We could just apply it to net, and deal with the conflict locally. But if you don't care about older kernels then better if you rebase.
Quoted 9 lines
Stable would be nice, but only to v6.1. I think I don't care about older kernels.
So what about if I resend this one here, based on top of the reformat patch, with a:
Cc: <[email protected]> # 6.1.x
line in there, and then reply to the email that the automatic backport failed, with a tailored patch for v6.1?
Alternatively I can send an explicit stable backport email once this one is merged.
Quoted 6 lines
Note that we can merge this kind of changes via the -net tree. No repost will be needed. We can merge it as is on -net and you can follow the option 2 from the stable kernel rules doc, with no repost nor additional mangling for stable will be needed.
If you are ok with the above let me know.
That sounds good to me, but that will then trigger a merge conflict when net-next (with the reformat patch) is merged? I guess it's easy enough to solve, but that would be extra work on your side. If you are fine with that, it's OK for me.
Thanks,
Andre
Paolo Abeni
maintainerQuoted 29 lines · Andre Przywara
On Thu, 12 Jan 2023 12:39:01 +0100
Paolo Abeni <[email protected]> wrote:Hi,
Quoted 24 lines · Andre Przywara
Quoted 8 lines · On Wed, 11 Jan 2023 21:31:43 -0800 Jakub Kicinski <[email protected]> wrote:
Quoted 6 lines
Hm, we have a patch in net-next which reformats the entries:
ec51fbd1b8a2bca2948dede99c14ec63dc57ff6bWould you like this ID to be also added in stable? We could just apply it to net, and deal with the conflict locally. But if you don't care about older kernels then better if you rebase.
Quoted 9 lines
Stable would be nice, but only to v6.1. I think I don't care about older kernels.
So what about if I resend this one here, based on top of the reformat patch, with a:
Cc: <[email protected]> # 6.1.x
line in there, and then reply to the email that the automatic backport failed, with a tailored patch for v6.1?
Alternatively I can send an explicit stable backport email once this one is merged.Note that we can merge this kind of changes via the -net tree. No repost will be needed. We can merge it as is on -net and you can follow the option 2 from the stable kernel rules doc, with no repost nor additional mangling for stable will be needed.
Quoted 6 lines
Quoted 1 line
If you are ok with the above let me know.
That sounds good to me, but that will then trigger a merge conflict when net-next (with the reformat patch) is merged? I guess it's easy enough to solve, but that would be extra work on your side. If you are fine with that, it's OK for me.
Fine by us (well, probably poor Jakub will end-up handling the conflict, but AFAIK he is ok with this specific case).
I'll merge the patch on net.
Cheers,
Paolo
This patch was applied to netdev/net.git (master) by Paolo Abeni <[email protected]>:
Hello:
This patch was applied to netdev/net.git (master)
by Paolo Abeni <[email protected]>:
Quoted 9 lines · On Wed, 11 Jan 2023 13:32:28 +0000 you wrote:
The Microsoft Devkit 2023 is a an ARM64 based machine featuring a Realtek 8153 USB3.0-to-GBit Ethernet adapter. As in their other machines, Microsoft uses a custom USB device ID.
Add the respective ID values to the driver. This makes Ethernet work on the MS Devkit device. The chip has been visually confirmed to be a RTL8153.
[...]
Here is the summary with links:
- [net-next] r8152: add vendor/device ID pair for Microsoft Devkit
https://git.kernel.org/netdev/net/c/be53771c87f4
You are awesome, thank you!
Signature · 2 lines
Deet-doot-dot, I am a bot. https://korg.docs.kernel.org/patchwork/pwbot.html