Skip to contentSkip to conversation

musb patches for v5.8-rc1

PATCH In review Bin Liu 4 participants 11 messages May 25, 2020 – Jun 19, 2020

Subsystems (4)
lore
Share & tools
lore

Bin Liu

author maintainer
#0 lore raw

cover musb patches for v5.8-rc1

Hi Greg,

Here are musb patches for v5.8-rc1, including some fixes which were late to get in v5.7. I cc'd stable@ in some of them but wanted the rest fixes to sit in v5.8 for a while before backport to stable.

Please let me know if any change is needed.

Thanks,
-Bin.
---

Bin Liu (2):
usb: musb: return -ESHUTDOWN in urb when three-strikes error happened
usb: musb: start session in resume for host port

Dinghao Liu (1):
usb: musb: Fix runtime PM imbalance on error

Jason Yan (1):
usb: musb: use true for 'use_dma'

Macpaul Lin (1):
usb: musb: mediatek: add reset FADDR to zero in reset interrupt handle

Paul Cercueil (1):
usb: musb: jz4740: Prevent lockup when CONFIG_SMP is set

 drivers/usb/musb/jz4740.c       |  4 ++--
 drivers/usb/musb/mediatek.c     |  6 ++++++
 drivers/usb/musb/musb_core.c    |  9 ++++++++-
 drivers/usb/musb/musb_debugfs.c | 10 +++++-----
 drivers/usb/musb/musb_host.c    | 10 ++++++++--
 5 files changed, 29 insertions(+), 10 deletions(-)
Signature · 1 line
2.17.1
author maintainer
#1 lore raw

PATCH 1/6 usb: musb: return -ESHUTDOWN in urb when three-strikes error happened

When a USB device attached to a hub got disconnected, MUSB controller generates RXCSR_RX_ERROR interrupt for the 3-strikes-out error.

Currently the MUSB host driver returns -EPROTO in current URB, then the USB device driver could immediately resubmit the URB which causes MUSB generate RXCSR_RX_ERROR interrupt again. This circle causes interrupt storm then the hub never got a chance to report the USB device detach.

To fix the interrupt storm, change the URB return code to -ESHUTDOWN for MUSB_RXCSR_H_ERROR interrupt, so that the USB device driver will not immediately resubmit the URB.

---
 drivers/usb/musb/musb_host.c | 10 ++++++++--
 1 file changed, 8 insertions(+), 2 deletions(-)
drivers/usb/musb/musb_host.c +8 -2
index 8736f4251a22..8b7d22a0c0fb 100644
@@ -1774,9 +1774,15 @@ void musb_host_rx(struct musb *musb, u8 epnum)
 		status = -EPIPE;
 
 	} else if (rx_csr & MUSB_RXCSR_H_ERROR) {
-		musb_dbg(musb, "end %d RX proto error", epnum);
+		dev_err(musb->controller, "ep%d RX three-strikes error", epnum);
 
-		status = -EPROTO;
+		/*
+		 * The three-strikes error could only happen when the USB
+		 * device is not accessible, for example detached or powered
+		 * off. So return the fatal error -ESHUTDOWN so hopefully the
+		 * USB device drivers won't immediately resubmit the same URB.
+		 */
+		status = -ESHUTDOWN;
 		musb_writeb(epio, MUSB_RXINTERVAL, 0);
 
 		rx_csr &= ~MUSB_RXCSR_H_ERROR;
Signature · 1 line
2.17.1
author maintainer
#2 lore raw

PATCH 2/6 usb: musb: start session in resume for host port

In reply to Bin Liu

Commit 17539f2f4f0b ("usb: musb: fix enumeration after resume") replaced musb_start() in musb_resume() to not override softconnect bit, but it doesn't restart the session for host port which was done in musb_start(). The session could be disabled in musb_suspend(), which leads the host port doesn't stay in host mode.

So let's start the session specifically for host port in musb_resume().

Fixes 17539f2f4f0b ("usb: musb: fix enumeration after resume")
---
 drivers/usb/musb/musb_core.c | 7 +++++++
 1 file changed, 7 insertions(+)
drivers/usb/musb/musb_core.c +7 -0
index d590110539ab..48178aeccf5b 100644
@@ -2877,6 +2877,13 @@ static int musb_resume(struct device *dev)
 	musb_enable_interrupts(musb);
 	musb_platform_enable(musb);
 
+	/* session might be disabled in suspend */
+	if (musb->port_mode == MUSB_HOST &&
+	    !(musb->ops->quirks & MUSB_PRESERVE_SESSION)) {
+		devctl |= MUSB_DEVCTL_SESSION;
+		musb_writeb(musb->mregs, MUSB_DEVCTL, devctl);
+	}
+
 	spin_lock_irqsave(&musb->lock, flags);
 	error = musb_run_resume_work(musb);
 	if (error)
Signature · 1 line
2.17.1
author maintainer
#3 lore raw

PATCH 3/6 usb: musb: use true for 'use_dma'

In reply to Bin Liu

From: Jason Yan <[email protected]>

Fix the following coccicheck warning:

drivers/usb/musb/musb_core.c:1798:12-19: WARNING: Assignment of 0/1 to bool variable

---
 drivers/usb/musb/musb_core.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/usb/musb/musb_core.c +1 -1
index 48178aeccf5b..384a8039a7fd 100644
@@ -1795,7 +1795,7 @@ irqreturn_t musb_interrupt(struct musb *musb)
 EXPORT_SYMBOL_GPL(musb_interrupt);
 
 #ifndef CONFIG_MUSB_PIO_ONLY
-static bool use_dma = 1;
+static bool use_dma = true;
 
 /* "modprobe ... use_dma=0" etc */
 module_param(use_dma, bool, 0644);
Signature · 1 line
2.17.1
author maintainer
#4 lore raw

PATCH 4/6 usb: musb: mediatek: add reset FADDR to zero in reset interrupt handle

In reply to Bin Liu

From: Macpaul Lin <[email protected]>

When receiving reset interrupt, FADDR need to be reset to zero in peripheral mode. Otherwise ep0 cannot do enumeration when re-plugging USB cable.

---
 drivers/usb/musb/mediatek.c | 6 ++++++
 1 file changed, 6 insertions(+)
drivers/usb/musb/mediatek.c +6 -0
index 6196b0e8d77d..eebeadd26946 100644
@@ -208,6 +208,12 @@ static irqreturn_t generic_interrupt(int irq, void *__hci)
 	musb->int_rx = musb_clearw(musb->mregs, MUSB_INTRRX);
 	musb->int_tx = musb_clearw(musb->mregs, MUSB_INTRTX);
 
+	if ((musb->int_usb & MUSB_INTR_RESET) && !is_host_active(musb)) {
+		/* ep0 FADDR must be 0 when (re)entering peripheral mode */
+		musb_ep_select(musb->mregs, 0);
+		musb_writeb(musb->mregs, MUSB_FADDR, 0);
+	}
+
 	if (musb->int_usb || musb->int_tx || musb->int_rx)
 		retval = musb_interrupt(musb);
 
Signature · 1 line
2.17.1
#5 lore raw

Re: [PATCH 4/6] usb: musb: mediatek: add reset FADDR to zero in reset interrupt handle

Bin Liu <[email protected]> 於 2020年5月25日 週一 上午10:53寫道:

Quoted 32 lines

From: Macpaul Lin <[email protected]>

When receiving reset interrupt, FADDR need to be reset to zero in peripheral mode. Otherwise ep0 cannot do enumeration when re-plugging USB cable.

---
 drivers/usb/musb/mediatek.c | 6 ++++++
 1 file changed, 6 insertions(+)
drivers/usb/musb/mediatek.c +6 -1
index 6196b0e8d77d..eebeadd26946 100644
@@ -208,6 +208,12 @@ static irqreturn_t generic_interrupt(int irq, void *__hci)
        musb->int_rx = musb_clearw(musb->mregs, MUSB_INTRRX);
        musb->int_tx = musb_clearw(musb->mregs, MUSB_INTRTX);
 
+       if ((musb->int_usb & MUSB_INTR_RESET) && !is_host_active(musb)) {
+               /* ep0 FADDR must be 0 when (re)entering peripheral mode */
+               musb_ep_select(musb->mregs, 0);
+               musb_writeb(musb->mregs, MUSB_FADDR, 0);
+       }
+
        if (musb->int_usb || musb->int_tx || musb->int_rx)
                retval = musb_interrupt(musb);
 
--
Quoted 2 lines

2.17.1

Could this bug fix also been applied to stable kernel? Thanks!

Best regards,
Macpaul Lin

#6 lore raw

Re: [PATCH 4/6] usb: musb: mediatek: add reset FADDR to zero in reset interrupt handle

Quoted 32 lines · Macpaul Lin

Bin Liu <[email protected]> 於 2020年5月25日 週一 上午10:53寫道:

Quoted 31 lines

From: Macpaul Lin <[email protected]>

When receiving reset interrupt, FADDR need to be reset to zero in peripheral mode. Otherwise ep0 cannot do enumeration when re-plugging USB cable.

---
 drivers/usb/musb/mediatek.c | 6 ++++++
 1 file changed, 6 insertions(+)
drivers/usb/musb/mediatek.c +6 -0
index 6196b0e8d77d..eebeadd26946 100644
@@ -208,6 +208,12 @@ static irqreturn_t generic_interrupt(int irq, void *__hci)
        musb->int_rx = musb_clearw(musb->mregs, MUSB_INTRRX);
        musb->int_tx = musb_clearw(musb->mregs, MUSB_INTRTX);
 
+       if ((musb->int_usb & MUSB_INTR_RESET) && !is_host_active(musb)) {
+               /* ep0 FADDR must be 0 when (re)entering peripheral mode */
+               musb_ep_select(musb->mregs, 0);
+               musb_writeb(musb->mregs, MUSB_FADDR, 0);
+       }
+
        if (musb->int_usb || musb->int_tx || musb->int_rx)
                retval = musb_interrupt(musb);
 
Quoted 4 lines
Quoted 3 lines

--
2.17.1

Could this bug fix also been applied to stable kernel?

Sure, what is the git commit of it in Linus's tree?

thanks,

greg k-h

#7 lore raw

Re: [PATCH 4/6] usb: musb: mediatek: add reset FADDR to zero in reset interrupt handle

Quoted 38 lines · Greg Kroah-Hartman
Quoted 32 lines · Macpaul Lin

Bin Liu <[email protected]> 於 2020年5月25日 週一 上午10:53寫道:

Quoted 31 lines

From: Macpaul Lin <[email protected]>

When receiving reset interrupt, FADDR need to be reset to zero in peripheral mode. Otherwise ep0 cannot do enumeration when re-plugging USB cable.

---
 drivers/usb/musb/mediatek.c | 6 ++++++
 1 file changed, 6 insertions(+)
drivers/usb/musb/mediatek.c +6 -0
index 6196b0e8d77d..eebeadd26946 100644
@@ -208,6 +208,12 @@ static irqreturn_t generic_interrupt(int irq, void *__hci)
        musb->int_rx = musb_clearw(musb->mregs, MUSB_INTRRX);
        musb->int_tx = musb_clearw(musb->mregs, MUSB_INTRTX);
 
+       if ((musb->int_usb & MUSB_INTR_RESET) && !is_host_active(musb)) {
+               /* ep0 FADDR must be 0 when (re)entering peripheral mode */
+               musb_ep_select(musb->mregs, 0);
+               musb_writeb(musb->mregs, MUSB_FADDR, 0);
+       }
+
        if (musb->int_usb || musb->int_tx || musb->int_rx)
                retval = musb_interrupt(musb);
 
Quoted 4 lines
Quoted 3 lines

--
2.17.1

Could this bug fix also been applied to stable kernel?

Quoted 5 lines

Sure, what is the git commit of it in Linus's tree?

thanks,

greg k-h

The commit id of this patch should be
402bcac4b25b520c89ba60db85eb6316f36e797f

Thank you very much.

BR
Macpaul Lin

#8 lore raw

Re: [PATCH 4/6] usb: musb: mediatek: add reset FADDR to zero in reset interrupt handle

Quoted 39 lines · Macpaul Lin
Quoted 38 lines · Greg Kroah-Hartman
Quoted 36 lines · Macpaul Lin

Bin Liu <[email protected]> 於 2020年5月25日 週一 上午10:53寫道:

Quoted 32 lines

From: Macpaul Lin <[email protected]>

When receiving reset interrupt, FADDR need to be reset to zero in peripheral mode. Otherwise ep0 cannot do enumeration when re-plugging USB cable.

---
 drivers/usb/musb/mediatek.c | 6 ++++++
 1 file changed, 6 insertions(+)
drivers/usb/musb/mediatek.c +6 -1
index 6196b0e8d77d..eebeadd26946 100644
@@ -208,6 +208,12 @@ static irqreturn_t generic_interrupt(int irq, void *__hci)
        musb->int_rx = musb_clearw(musb->mregs, MUSB_INTRRX);
        musb->int_tx = musb_clearw(musb->mregs, MUSB_INTRTX);
 
+       if ((musb->int_usb & MUSB_INTR_RESET) && !is_host_active(musb)) {
+               /* ep0 FADDR must be 0 when (re)entering peripheral mode */
+               musb_ep_select(musb->mregs, 0);
+               musb_writeb(musb->mregs, MUSB_FADDR, 0);
+       }
+
        if (musb->int_usb || musb->int_tx || musb->int_rx)
                retval = musb_interrupt(musb);
 
--
Quoted 2 lines

2.17.1

Could this bug fix also been applied to stable kernel?

Quoted 8 lines
Quoted 5 lines

Sure, what is the git commit of it in Linus's tree?

thanks,

greg k-h

The commit id of this patch should be
402bcac4b25b520c89ba60db85eb6316f36e797f

Now queued up, thanks.

greg k-h

author maintainer
#9 lore raw

PATCH 5/6 usb: musb: jz4740: Prevent lockup when CONFIG_SMP is set

In reply to Bin Liu

From: Paul Cercueil <[email protected]>

The function dma_controller_irq() locks up the exact same spinlock we locked before calling it, which obviously resulted in a deadlock when CONFIG_SMP was enabled. This flew under the radar as none of the boards supported by this driver needs SMP.

Fixes 57aadb46bd63 ("usb: musb: jz4740: Add support for DMA")
---
 drivers/usb/musb/jz4740.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)
drivers/usb/musb/jz4740.c +2 -2
index e64dd30e80e7..c4fe1f4cd17a 100644
@@ -30,11 +30,11 @@ static irqreturn_t jz4740_musb_interrupt(int irq, void *__hci)
 	irqreturn_t	retval = IRQ_NONE, retval_dma = IRQ_NONE;
 	struct musb	*musb = __hci;
 
-	spin_lock_irqsave(&musb->lock, flags);
-
 	if (IS_ENABLED(CONFIG_USB_INVENTRA_DMA) && musb->dma_controller)
 		retval_dma = dma_controller_irq(irq, musb->dma_controller);
 
+	spin_lock_irqsave(&musb->lock, flags);
+
 	musb->int_usb = musb_readb(musb->mregs, MUSB_INTRUSB);
 	musb->int_tx = musb_readw(musb->mregs, MUSB_INTRTX);
 	musb->int_rx = musb_readw(musb->mregs, MUSB_INTRRX);
Signature · 1 line
2.17.1
author maintainer
#10 lore raw

PATCH 6/6 usb: musb: Fix runtime PM imbalance on error

In reply to Bin Liu

From: Dinghao Liu <[email protected]>

When copy_from_user() returns an error code, there is a runtime PM usage counter imbalance.

Fix this by moving copy_from_user() to the beginning of this function.

Fixes 7b6c1b4c0e1e ("usb: musb: fix runtime PM in debugfs")
---
 drivers/usb/musb/musb_debugfs.c | 10 +++++-----
 1 file changed, 5 insertions(+), 5 deletions(-)
drivers/usb/musb/musb_debugfs.c +5 -5
index 7b6281ab62ed..30a89aa8a3e7 100644
@@ -168,6 +168,11 @@ static ssize_t musb_test_mode_write(struct file *file,
 	u8			test;
 	char			buf[24];
 
+	memset(buf, 0x00, sizeof(buf));
+
+	if (copy_from_user(buf, ubuf, min_t(size_t, sizeof(buf) - 1, count)))
+		return -EFAULT;
+
 	pm_runtime_get_sync(musb->controller);
 	test = musb_readb(musb->mregs, MUSB_TESTMODE);
 	if (test) {
@@ -176,11 +181,6 @@ static ssize_t musb_test_mode_write(struct file *file,
 		goto ret;
 	}
 
-	memset(buf, 0x00, sizeof(buf));
-
-	if (copy_from_user(buf, ubuf, min_t(size_t, sizeof(buf) - 1, count)))
-		return -EFAULT;
-
 	if (strstarts(buf, "force host full-speed"))
 		test = MUSB_TEST_FORCE_HOST | MUSB_TEST_FORCE_FS;
 
Signature · 1 line
2.17.1