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
 
Hash Suite for Android: free password hash cracker in your pocket
[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Date:   Thu, 18 Jun 2020 13:18:59 +0200
From:   Lorenzo Bianconi <lorenzo@...nel.org>
To:     Oleksandr Natalenko <oleksandr@...hat.com>
Cc:     Lorenzo Bianconi <lorenzo.bianconi83@...il.com>,
        Felix Fietkau <nbd@....name>,
        Ryder Lee <ryder.lee@...iatek.com>,
        Kalle Valo <kvalo@...eaurora.org>,
        "David S. Miller" <davem@...emloft.net>,
        Jakub Kicinski <kuba@...nel.org>,
        Matthias Brugger <matthias.bgg@...il.com>,
        linux-wireless@...r.kernel.org, netdev@...r.kernel.org,
        linux-mediatek@...ts.infradead.org, linux-kernel@...r.kernel.org
Subject: Re: mt7612 suspend/resume issue

> Hello, Lorenzo et al.

Hi Oleksandr,

> 
> I'm using MT7612 mini-PCIE cards as both AP in a home server and as a client in
> a laptop. The AP works perfectly (after some fixing from your side; thanks for
> that!), and so does the client modulo it has issues during system resume.
> 

[...]

> 
> The Wi-Fi becomes unusable from this point. If I `modprobe -r` the "mt76x2e" module
> after this splat, the system hangs completely.
> 
> If the system resumes fine, the resume itself takes quite some time (more than
> 10 seconds).
> 
> I've found a workaround for this, though. It seems the system behaves fine if I
> do `modprobe -r mt76x2e` before it goes to sleep, and then `modprobe mt76x2e`
> after resume. Also, the resume time improves greatly.
> 
> I cannot say if it is some regression or not. I've installed the card
> just recently, and used it with v5.7 kernel series only.
> 
> Do you have any idea what could go wrong and how to approach the issue?

I have not reproduced the issue myself yet, but I guess we can try:
1- update to latest Felix's tree [1]
2- could you please try to apply the patch below? (compile-test only)

Regards,
Lorenzo

[1] https://github.com/nbd168/wireless

From 400268a0ee5843cf736308504dfbd2f20a291eaf Mon Sep 17 00:00:00 2001
Message-Id: <400268a0ee5843cf736308504dfbd2f20a291eaf.1592478809.git.lorenzo@...nel.org>
From: Lorenzo Bianconi <lorenzo@...nel.org>
Date: Thu, 18 Jun 2020 13:10:11 +0200
Subject: [PATCH] mt76: mt76x2: fix pci suspend

Signed-off-by: Lorenzo Bianconi <lorenzo@...nel.org>
---
 .../net/wireless/mediatek/mt76/mt76x02_dma.h  |  1 +
 .../net/wireless/mediatek/mt76/mt76x02_mmio.c | 15 +++++
 .../net/wireless/mediatek/mt76/mt76x2/pci.c   | 58 +++++++++++++++++++
 3 files changed, 74 insertions(+)

diff --git a/drivers/net/wireless/mediatek/mt76/mt76x02_dma.h b/drivers/net/wireless/mediatek/mt76/mt76x02_dma.h
index 4aff4f8e87b6..6262f2ecded5 100644
--- a/drivers/net/wireless/mediatek/mt76/mt76x02_dma.h
+++ b/drivers/net/wireless/mediatek/mt76/mt76x02_dma.h
@@ -62,5 +62,6 @@ mt76x02_wait_for_wpdma(struct mt76_dev *dev, int timeout)
 int mt76x02_dma_init(struct mt76x02_dev *dev);
 void mt76x02_dma_disable(struct mt76x02_dev *dev);
 void mt76x02_dma_cleanup(struct mt76x02_dev *dev);
+void mt76x02_dma_reset(struct mt76x02_dev *dev);
 
 #endif /* __MT76x02_DMA_H */
diff --git a/drivers/net/wireless/mediatek/mt76/mt76x02_mmio.c b/drivers/net/wireless/mediatek/mt76/mt76x02_mmio.c
index cbbe986655fe..e2631c964331 100644
--- a/drivers/net/wireless/mediatek/mt76/mt76x02_mmio.c
+++ b/drivers/net/wireless/mediatek/mt76/mt76x02_mmio.c
@@ -348,6 +348,21 @@ void mt76x02_dma_disable(struct mt76x02_dev *dev)
 }
 EXPORT_SYMBOL_GPL(mt76x02_dma_disable);
 
+void mt76x02_dma_reset(struct mt76x02_dev *dev)
+{
+	int i;
+
+	mt76x02_dma_disable(dev);
+	usleep_range(1000, 2000);
+
+	for (i = 0; i < __MT_TXQ_MAX; i++)
+		mt76_queue_tx_cleanup(dev, i, true);
+	mt76_for_each_q_rx(&dev->mt76, i)
+		mt76_queue_rx_reset(dev, i);
+
+	mt76x02_dma_enable(dev);
+}
+
 void mt76x02_mac_start(struct mt76x02_dev *dev)
 {
 	mt76x02_mac_reset_counters(dev);
diff --git a/drivers/net/wireless/mediatek/mt76/mt76x2/pci.c b/drivers/net/wireless/mediatek/mt76/mt76x2/pci.c
index 53ca0cedf026..5543e242fb9b 100644
--- a/drivers/net/wireless/mediatek/mt76/mt76x2/pci.c
+++ b/drivers/net/wireless/mediatek/mt76/mt76x2/pci.c
@@ -103,6 +103,60 @@ mt76pci_remove(struct pci_dev *pdev)
 	mt76_free_device(mdev);
 }
 
+static int __maybe_unused
+mt76x2e_suspend(struct pci_dev *pdev, pm_message_t state)
+{
+	struct mt76_dev *mdev = pci_get_drvdata(pdev);
+	struct mt76x02_dev *dev = container_of(mdev, struct mt76x02_dev, mt76);
+	int i, err;
+
+	napi_disable(&mdev->tx_napi);
+	tasklet_kill(&mdev->pre_tbtt_tasklet);
+	tasklet_kill(&mdev->tx_tasklet);
+
+	mt76_for_each_q_rx(mdev, i)
+		napi_disable(&mdev->napi[i]);
+
+	mt76x02_dma_reset(dev);
+
+	pci_enable_wake(pdev, pci_choose_state(pdev, state), true);
+	pci_save_state(pdev);
+	err = pci_set_power_state(pdev, pci_choose_state(pdev, state));
+	if (err)
+		goto restore;
+
+	return 0;
+
+restore:
+	mt76_for_each_q_rx(mdev, i)
+		napi_enable(&mdev->napi[i]);
+	napi_enable(&mdev->tx_napi);
+
+	return err;
+}
+
+static int __maybe_unused
+mt76x2e_resume(struct pci_dev *pdev)
+{
+	struct mt76_dev *mdev = pci_get_drvdata(pdev);
+	int i, err;
+
+	err = pci_set_power_state(pdev, PCI_D0);
+	if (err)
+		return err;
+
+	pci_restore_state(pdev);
+
+	mt76_for_each_q_rx(mdev, i) {
+		napi_enable(&mdev->napi[i]);
+		napi_schedule(&mdev->napi[i]);
+	}
+	napi_enable(&mdev->tx_napi);
+	napi_schedule(&mdev->tx_napi);
+
+	return 0;
+}
+
 MODULE_DEVICE_TABLE(pci, mt76pci_device_table);
 MODULE_FIRMWARE(MT7662_FIRMWARE);
 MODULE_FIRMWARE(MT7662_ROM_PATCH);
@@ -113,6 +167,10 @@ static struct pci_driver mt76pci_driver = {
 	.id_table	= mt76pci_device_table,
 	.probe		= mt76pci_probe,
 	.remove		= mt76pci_remove,
+#ifdef CONFIG_PM
+	.suspend	= mt76x2e_suspend,
+	.resume		= mt76x2e_resume,
+#endif /* CONFIG_PM */
 };
 
 module_pci_driver(mt76pci_driver);
-- 
2.26.2


> 
> Thanks.
> 
> -- 
>   Best regards,
>     Oleksandr Natalenko (post-factum)
>     Principal Software Maintenance Engineer
> 

Download attachment "signature.asc" of type "application/pgp-signature" (229 bytes)

Powered by blists - more mailing lists

Powered by Openwall GNU/*/Linux Powered by OpenVZ