| lists.openwall.net | lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC | |
|
Open Source and information security mailing list archives
| ||
|
Message-ID: <5f2db9d90907212025k31c3a941v5fd1be9ed171e4c1@mail.gmail.com>
Date: Tue, 21 Jul 2009 20:25:31 -0700
From: Alexander Duyck <alexander.duyck@...il.com>
To: Andy Gospodarek <andy@...yhouse.net>
Cc: "Ronciak, John" <john.ronciak@...el.com>,
"netdev@...r.kernel.org" <netdev@...r.kernel.org>,
"Kirsher, Jeffrey T" <jeffrey.t.kirsher@...el.com>,
"Brandeburg, Jesse" <jesse.brandeburg@...el.com>,
"Allan, Bruce W" <bruce.w.allan@...el.com>,
"Waskiewicz Jr, Peter P" <peter.p.waskiewicz.jr@...el.com>,
"stable@...nel.org" <stable@...nel.org>
Subject: Re: [PATCH] igb/e1000e/e1000/e100: make wol usable
My understanding was that the can_wakeup was supposed to be
initialized by pci_pm_init or platform_pci_wakeup_init based on the
pci-e capabilities. Is this not the case? It seems like this is
where you should be looking to determine why the the can_wakeup isn't
being initialized correctly.
The patch below won't solve the problem either. The problem is that
the can_wakeup value is being disabled when the EEPROM doesn't support
WOL.
If you have to do this in the drivers, then my suggestion is to take a
look at how ixgbe is doing it. You essentially need to initialize
can_wakeup to true, and then set the should_wakeup attribute based on
the EEPROM setting or via ethtool. This way you can still toggle the
should_wakeup option without being blocked by the EEPROM disabling it.
Thanks,
Alex
On Tue, Jul 21, 2009 at 7:27 PM, Andy Gospodarek<andy@...yhouse.net> wrote:
> On Tue, Jul 21, 2009 at 02:59:31PM -0700, Ronciak, John wrote:
>> >-----Original Message-----
>> >From: Andy Gospodarek [mailto:andy@...yhouse.net]
>> >Sent: Tuesday, July 21, 2009 2:01 PM
>> >To: netdev@...r.kernel.org
>> >Cc: Kirsher, Jeffrey T; Brandeburg, Jesse; Allan, Bruce W;
>> >Waskiewicz Jr, Peter P; Ronciak, John; stable@...nel.org
>> >Subject: [PATCH] igb/e1000e/e1000/e100: make wol usable
>> >
>> >
>> >It seems that the patches that intended to use the new calls to check
>> >status and capabilities for WOL and fix it up:
>> >
>> > commit bc79fc8409b3dccbde072e8113cc1fb3fd876fc5
>> > Author: Rafael J. Wysocki <rjw@...k.pl>
>> > Date: Wed Oct 29 14:22:18 2008 -0700
>> >
>> > e100: adapt to the reworked PCI PM
>> >
>> > commit e1b86d8479f90aadee57a3d07d8e61c815c202d9
>> > Author: Rafael J. Wysocki <rjw@...k.pl>
>> > Date: Fri Nov 7 20:30:37 2008 +0000
>> >
>> > igb: Use device_set_wakeup_enable
>> >
>> > commit de1264896c8012a261c1cba17e6a61199c276ad3
>> > Author: Rafael J. Wysocki <rjw@...k.pl>
>> > Date: Fri Nov 7 20:30:19 2008 +0000
>> >
>> > e1000: Use device_set_wakeup_enable
>> >
>> > commit 6ff68026f4757d68461b7fbeca5c944e1f5f8b44
>> > Author: Rafael J. Wysocki <rjw@...k.pl>
>> > Date: Wed Nov 12 09:52:32 2008 +0000
>> >
>> > e1000e: Use device_set_wakeup_enable
>> >
>> >Unfortunately they left out some important bits. This should make sure
>> >the e100, igb, e1000e, and e1000-based devices can actually enable WOL
>> >rather than their current state which is reported as not supported.
>> >
>> >This looks like it has been broken since 2.6.28.
>> >
>> >Signed-off-by: Andy Gospodarek <andy@...yhouse.net>
>> >---
>> >
>> > e100.c | 1 +
>> > e1000/e1000_main.c | 1 +
>> > e1000e/netdev.c | 1 +
>> > igb/igb_main.c | 1 +
>> > 4 files changed, 4 insertions(+)
>> >
>> >diff --git a/drivers/net/e100.c b/drivers/net/e100.c
>> >index 569df19..8824952 100644
>> >--- a/drivers/net/e100.c
>> >+++ b/drivers/net/e100.c
>> >@@ -2816,6 +2816,7 @@ static int __devinit e100_probe(struct
>> >pci_dev *pdev,
>> > if ((nic->mac >= mac_82558_D101_A4) &&
>> > (nic->eeprom[eeprom_id] & eeprom_id_wol)) {
>> > nic->flags |= wol_magic;
>> >+ device_init_wakeup(&pdev->dev, true);
>> > device_set_wakeup_enable(&pdev->dev, true);
>> > }
>> >
>> >diff --git a/drivers/net/e1000/e1000_main.c
>> >b/drivers/net/e1000/e1000_main.c
>> >index d7df00c..229d874 100644
>> >--- a/drivers/net/e1000/e1000_main.c
>> >+++ b/drivers/net/e1000/e1000_main.c
>> >@@ -1208,6 +1208,7 @@ static int __devinit e1000_probe(struct
>> >pci_dev *pdev,
>> >
>> > /* initialize the wol settings based on the eeprom settings */
>> > adapter->wol = adapter->eeprom_wol;
>> >+ device_init_wakeup(&adapter->pdev->dev, adapter->wol);
>> > device_set_wakeup_enable(&adapter->pdev->dev, adapter->wol);
>> >
>> > /* print bus type/speed/width info */
>> >diff --git a/drivers/net/e1000e/netdev.c b/drivers/net/e1000e/netdev.c
>> >index 63415bb..5c2edaa 100644
>> >--- a/drivers/net/e1000e/netdev.c
>> >+++ b/drivers/net/e1000e/netdev.c
>> >@@ -5196,6 +5196,7 @@ static int __devinit e1000_probe(struct
>> >pci_dev *pdev,
>> >
>> > /* initialize the wol settings based on the eeprom settings */
>> > adapter->wol = adapter->eeprom_wol;
>> >+ device_init_wakeup(&adapter->pdev->dev, adapter->wol);
>> > device_set_wakeup_enable(&adapter->pdev->dev, adapter->wol);
>> >
>> > /* save off EEPROM version number */
>> >diff --git a/drivers/net/igb/igb_main.c b/drivers/net/igb/igb_main.c
>> >index adb09d3..8512a40 100644
>> >--- a/drivers/net/igb/igb_main.c
>> >+++ b/drivers/net/igb/igb_main.c
>> >@@ -1475,6 +1475,7 @@ static int __devinit igb_probe(struct
>> >pci_dev *pdev,
>> >
>> > /* initialize the wol settings based on the eeprom settings */
>> > adapter->wol = adapter->eeprom_wol;
>> >+ device_init_wakeup(&adapter->pdev->dev, adapter->wol);
>> > device_set_wakeup_enable(&adapter->pdev->dev, adapter->wol);
>> >
>> > /* reset the hardware with the new settings */
>> Andy,
>>
>> WoL not supported is reported back when WoL is disabled in the EEPROM. So if this is happening to you, that system (or at least that interface) has WoL disabled. How are you testing this?
>>
>> Cheers,
>> John
>
> John,
>
> I'm not doing anything differently than what the eeprom is reporting as
> support with these patches. In every one of the drivers above, the
> enabling of ethtool first calls device_can_wakeup. That inline function
> is pretty simple and all it does is return the value of
> dev->power.can_wakeup. If never set, dev->power.can_wakeup is false and
> wol status can never be set.
>
> The proper way to set it would be to call dev_init_wakeup first (or
> simply call it). Take a look at the code and you will see what I'm
> talking about.
>
> Now that I'm looking at it again, this patch would probably be better
> anyway:
>
> Replace device_set_wakeup_enable calls with device_init_wakeup, so that
> calls to device_can_wakeup in each drivers' code will not fail each
> time. The intent of the calls to device_can_wakeup were added to ensure
> the device was enabled for WOL, but without the proper call to
> device_init_wakeup, any call to device_can_wakeup will always return
> false.
>
> Signed-off-by: Andy Gospodarek <andy@...yhouse.net>
> ---
>
> e100.c | 2 +-
> e1000/e1000_main.c | 2 +-
> e1000e/netdev.c | 2 +-
> igb/igb_main.c | 2 +-
> 4 files changed, 4 insertions(+), 4 deletions(-)
>
> diff --git a/drivers/net/e100.c b/drivers/net/e100.c
> index 569df19..1d82d25 100644
> --- a/drivers/net/e100.c
> +++ b/drivers/net/e100.c
> @@ -2816,7 +2816,7 @@ static int __devinit e100_probe(struct pci_dev *pdev,
> if ((nic->mac >= mac_82558_D101_A4) &&
> (nic->eeprom[eeprom_id] & eeprom_id_wol)) {
> nic->flags |= wol_magic;
> - device_set_wakeup_enable(&pdev->dev, true);
> + device_init_wakeup(&pdev->dev, true);
> }
>
> /* ack any pending wake events, disable PME */
> diff --git a/drivers/net/e1000/e1000_main.c b/drivers/net/e1000/e1000_main.c
> index d7df00c..5f22a85 100644
> --- a/drivers/net/e1000/e1000_main.c
> +++ b/drivers/net/e1000/e1000_main.c
> @@ -1208,7 +1208,7 @@ static int __devinit e1000_probe(struct pci_dev *pdev,
>
> /* initialize the wol settings based on the eeprom settings */
> adapter->wol = adapter->eeprom_wol;
> - device_set_wakeup_enable(&adapter->pdev->dev, adapter->wol);
> + device_init_wakeup(&adapter->pdev->dev, adapter->wol);
>
> /* print bus type/speed/width info */
> DPRINTK(PROBE, INFO, "(PCI%s:%s:%s) ",
> diff --git a/drivers/net/e1000e/netdev.c b/drivers/net/e1000e/netdev.c
> index 63415bb..62de3ad 100644
> --- a/drivers/net/e1000e/netdev.c
> +++ b/drivers/net/e1000e/netdev.c
> @@ -5196,7 +5196,7 @@ static int __devinit e1000_probe(struct pci_dev *pdev,
>
> /* initialize the wol settings based on the eeprom settings */
> adapter->wol = adapter->eeprom_wol;
> - device_set_wakeup_enable(&adapter->pdev->dev, adapter->wol);
> + device_init_wakeup(&adapter->pdev->dev, adapter->wol);
>
> /* save off EEPROM version number */
> e1000_read_nvm(&adapter->hw, 5, 1, &adapter->eeprom_vers);
> diff --git a/drivers/net/igb/igb_main.c b/drivers/net/igb/igb_main.c
> index adb09d3..f761e67 100644
> --- a/drivers/net/igb/igb_main.c
> +++ b/drivers/net/igb/igb_main.c
> @@ -1475,7 +1475,7 @@ static int __devinit igb_probe(struct pci_dev *pdev,
>
> /* initialize the wol settings based on the eeprom settings */
> adapter->wol = adapter->eeprom_wol;
> - device_set_wakeup_enable(&adapter->pdev->dev, adapter->wol);
> + device_init_wakeup(&adapter->pdev->dev, adapter->wol);
>
> /* reset the hardware with the new settings */
> igb_reset(adapter);
>
>
>
>
>
>
> --
> To unsubscribe from this list: send the line "unsubscribe netdev" in
> the body of a message to majordomo@...r.kernel.org
> More majordomo info at http://vger.kernel.org/majordomo-info.html
>
--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Powered by blists - more mailing lists