2012-01-13 12:00:17

by Stanislaw Gruszka

Subject: [PATCH] rt2800pci: fix spurious interrupts generation

Same devices can generate interrupt without properly setting bit in
INT_SOURCE_CSR register (spurious interrupt), what will cause IRQ line
will be disabled by interrupts controller driver.

We discovered that clearing INT_MASK_CSR stops such behaviour. We
previously first read that register, and then clear all know interrupt
sources bits and do not touch reserved bits. After this patch, we write
to all register content (I believe writing to reserved bits on that
register will not cause any problems, I tested that on my rt2800pci
device).

This fix very bad performance problem, practically making device
unusable (since worked without interrupts), reported in:
https://bugzilla.redhat.com/show_bug.cgi?id=658451

We previously tried to workaround that issue in commit
4ba7d9997869d25bd223dea7536fc1ce9fab3b3b "rt2800pci: handle spurious
interrupts", but it was reverted in commit
82e5fc2a34fa9ffea38f00c4066b7e600a0ca5e6
as thing, that will prevent to detect real spurious interrupts.

Reported-and-tested-by: Amir Hedayaty <[email protected]>
Cc: [email protected]
Signed-off-by: Stanislaw Gruszka <[email protected]>
---
drivers/net/wireless/rt2x00/rt2800pci.c | 28 ++++++++--------------------
1 files changed, 8 insertions(+), 20 deletions(-)

diff --git a/drivers/net/wireless/rt2x00/rt2800pci.c b/drivers/net/wireless/rt2x00/rt2800pci.c
index 4941a1a..dc88bae 100644
--- a/drivers/net/wireless/rt2x00/rt2800pci.c
+++ b/drivers/net/wireless/rt2x00/rt2800pci.c
@@ -422,7 +422,6 @@ static int rt2800pci_init_queues(struct rt2x00_dev *rt2x00dev)
static void rt2800pci_toggle_irq(struct rt2x00_dev *rt2x00dev,
enum dev_state state)
{
- int mask = (state == STATE_RADIO_IRQ_ON);
u32 reg;
unsigned long flags;

@@ -436,25 +435,14 @@ static void rt2800pci_toggle_irq(struct rt2x00_dev *rt2x00dev,
}

spin_lock_irqsave(&rt2x00dev->irqmask_lock, flags);
- rt2x00pci_register_read(rt2x00dev, INT_MASK_CSR, &reg);
- rt2x00_set_field32(&reg, INT_MASK_CSR_RXDELAYINT, 0);
- rt2x00_set_field32(&reg, INT_MASK_CSR_TXDELAYINT, 0);
- rt2x00_set_field32(&reg, INT_MASK_CSR_RX_DONE, mask);
- rt2x00_set_field32(&reg, INT_MASK_CSR_AC0_DMA_DONE, 0);
- rt2x00_set_field32(&reg, INT_MASK_CSR_AC1_DMA_DONE, 0);
- rt2x00_set_field32(&reg, INT_MASK_CSR_AC2_DMA_DONE, 0);
- rt2x00_set_field32(&reg, INT_MASK_CSR_AC3_DMA_DONE, 0);
- rt2x00_set_field32(&reg, INT_MASK_CSR_HCCA_DMA_DONE, 0);
- rt2x00_set_field32(&reg, INT_MASK_CSR_MGMT_DMA_DONE, 0);
- rt2x00_set_field32(&reg, INT_MASK_CSR_MCU_COMMAND, 0);
- rt2x00_set_field32(&reg, INT_MASK_CSR_RXTX_COHERENT, 0);
- rt2x00_set_field32(&reg, INT_MASK_CSR_TBTT, mask);
- rt2x00_set_field32(&reg, INT_MASK_CSR_PRE_TBTT, mask);
- rt2x00_set_field32(&reg, INT_MASK_CSR_TX_FIFO_STATUS, mask);
- rt2x00_set_field32(&reg, INT_MASK_CSR_AUTO_WAKEUP, mask);
- rt2x00_set_field32(&reg, INT_MASK_CSR_GPTIMER, 0);
- rt2x00_set_field32(&reg, INT_MASK_CSR_RX_COHERENT, 0);
- rt2x00_set_field32(&reg, INT_MASK_CSR_TX_COHERENT, 0);
+ reg = 0;
+ if (state == STATE_RADIO_IRQ_ON) {
+ rt2x00_set_field32(&reg, INT_MASK_CSR_RX_DONE, 1);
+ rt2x00_set_field32(&reg, INT_MASK_CSR_TBTT, 1);
+ rt2x00_set_field32(&reg, INT_MASK_CSR_PRE_TBTT, 1);
+ rt2x00_set_field32(&reg, INT_MASK_CSR_TX_FIFO_STATUS, 1);
+ rt2x00_set_field32(&reg, INT_MASK_CSR_AUTO_WAKEUP, 1);
+ }
rt2x00pci_register_write(rt2x00dev, INT_MASK_CSR, reg);
spin_unlock_irqrestore(&rt2x00dev->irqmask_lock, flags);

--
1.7.1

2012-01-13 17:47:55

by Gertjan van Wingerde

Subject: Re: [PATCH] rt2800pci: fix spurious interrupts generation

On 13 jan. 2012, at 12:59, Stanislaw Gruszka <[email protected]> wrote:

> Same devices can generate interrupt without properly setting bit in
> INT_SOURCE_CSR register (spurious interrupt), what will cause IRQ line
> will be disabled by interrupts controller driver.
>
> We discovered that clearing INT_MASK_CSR stops such behaviour. We
> previously first read that register, and then clear all know interrupt
> sources bits and do not touch reserved bits. After this patch, we write
> to all register content (I believe writing to reserved bits on that
> register will not cause any problems, I tested that on my rt2800pci
> device).
>
> This fix very bad performance problem, practically making device
> unusable (since worked without interrupts), reported in:
> https://bugzilla.redhat.com/show_bug.cgi?id=658451
>
> We previously tried to workaround that issue in commit
> 4ba7d9997869d25bd223dea7536fc1ce9fab3b3b "rt2800pci: handle spurious
> interrupts", but it was reverted in commit
> 82e5fc2a34fa9ffea38f00c4066b7e600a0ca5e6
> as thing, that will prevent to detect real spurious interrupts.
>
> Reported-and-tested-by: Amir Hedayaty <[email protected]>
> Cc: [email protected]
> Signed-off-by: Stanislaw Gruszka <[email protected]>

Acked-by: Gertjan van Wingerde <[email protected]>

> ---
> drivers/net/wireless/rt2x00/rt2800pci.c | 28 ++++++++--------------------
> 1 files changed, 8 insertions(+), 20 deletions(-)
>
> diff --git a/drivers/net/wireless/rt2x00/rt2800pci.c b/drivers/net/wireless/rt2x00/rt2800pci.c
> index 4941a1a..dc88bae 100644
> --- a/drivers/net/wireless/rt2x00/rt2800pci.c
> +++ b/drivers/net/wireless/rt2x00/rt2800pci.c
> @@ -422,7 +422,6 @@ static int rt2800pci_init_queues(struct rt2x00_dev *rt2x00dev)
> static void rt2800pci_toggle_irq(struct rt2x00_dev *rt2x00dev,
> enum dev_state state)
> {
> - int mask = (state == STATE_RADIO_IRQ_ON);
> u32 reg;
> unsigned long flags;
>
> @@ -436,25 +435,14 @@ static void rt2800pci_toggle_irq(struct rt2x00_dev *rt2x00dev,
> }
>
> spin_lock_irqsave(&rt2x00dev->irqmask_lock, flags);
> - rt2x00pci_register_read(rt2x00dev, INT_MASK_CSR, &reg);
> - rt2x00_set_field32(&reg, INT_MASK_CSR_RXDELAYINT, 0);
> - rt2x00_set_field32(&reg, INT_MASK_CSR_TXDELAYINT, 0);
> - rt2x00_set_field32(&reg, INT_MASK_CSR_RX_DONE, mask);
> - rt2x00_set_field32(&reg, INT_MASK_CSR_AC0_DMA_DONE, 0);
> - rt2x00_set_field32(&reg, INT_MASK_CSR_AC1_DMA_DONE, 0);
> - rt2x00_set_field32(&reg, INT_MASK_CSR_AC2_DMA_DONE, 0);
> - rt2x00_set_field32(&reg, INT_MASK_CSR_AC3_DMA_DONE, 0);
> - rt2x00_set_field32(&reg, INT_MASK_CSR_HCCA_DMA_DONE, 0);
> - rt2x00_set_field32(&reg, INT_MASK_CSR_MGMT_DMA_DONE, 0);
> - rt2x00_set_field32(&reg, INT_MASK_CSR_MCU_COMMAND, 0);
> - rt2x00_set_field32(&reg, INT_MASK_CSR_RXTX_COHERENT, 0);
> - rt2x00_set_field32(&reg, INT_MASK_CSR_TBTT, mask);
> - rt2x00_set_field32(&reg, INT_MASK_CSR_PRE_TBTT, mask);
> - rt2x00_set_field32(&reg, INT_MASK_CSR_TX_FIFO_STATUS, mask);
> - rt2x00_set_field32(&reg, INT_MASK_CSR_AUTO_WAKEUP, mask);
> - rt2x00_set_field32(&reg, INT_MASK_CSR_GPTIMER, 0);
> - rt2x00_set_field32(&reg, INT_MASK_CSR_RX_COHERENT, 0);
> - rt2x00_set_field32(&reg, INT_MASK_CSR_TX_COHERENT, 0);
> + reg = 0;
> + if (state == STATE_RADIO_IRQ_ON) {
> + rt2x00_set_field32(&reg, INT_MASK_CSR_RX_DONE, 1);
> + rt2x00_set_field32(&reg, INT_MASK_CSR_TBTT, 1);
> + rt2x00_set_field32(&reg, INT_MASK_CSR_PRE_TBTT, 1);
> + rt2x00_set_field32(&reg, INT_MASK_CSR_TX_FIFO_STATUS, 1);
> + rt2x00_set_field32(&reg, INT_MASK_CSR_AUTO_WAKEUP, 1);
> + }
> rt2x00pci_register_write(rt2x00dev, INT_MASK_CSR, reg);
> spin_unlock_irqrestore(&rt2x00dev->irqmask_lock, flags);
>
> --
> 1.7.1
>

2012-01-13 12:19:11

by Helmut Schaa

Subject: Re: [PATCH] rt2800pci: fix spurious interrupts generation

On Fri, Jan 13, 2012 at 12:59 PM, Stanislaw Gruszka <[email protected]> wrote:
> Same devices can generate interrupt without properly setting bit in
> INT_SOURCE_CSR register (spurious interrupt), what will cause IRQ line
> will be disabled by interrupts controller driver.
>
> We discovered that clearing INT_MASK_CSR stops such behaviour. We
> previously first read that register, and then clear all know interrupt
> sources bits and do not touch reserved bits. After this patch, we write
> to all register content (I believe writing to reserved bits on that
> register will not cause any problems, I tested that on my rt2800pci
> device).
>
> This fix very bad performance problem, practically making device
> unusable (since worked without interrupts), reported in:
> https://bugzilla.redhat.com/show_bug.cgi?id=658451
>
> We previously tried to workaround that issue in commit
> 4ba7d9997869d25bd223dea7536fc1ce9fab3b3b "rt2800pci: handle spurious
> interrupts", but it was reverted in commit
> 82e5fc2a34fa9ffea38f00c4066b7e600a0ca5e6
> as thing, that will prevent to detect real spurious interrupts.
>
> Reported-and-tested-by: Amir Hedayaty <[email protected]>
> Cc: [email protected]
> Signed-off-by: Stanislaw Gruszka <[email protected]>
> ---
> ?drivers/net/wireless/rt2x00/rt2800pci.c | ? 28 ++++++++--------------------
> ?1 files changed, 8 insertions(+), 20 deletions(-)
>
> diff --git a/drivers/net/wireless/rt2x00/rt2800pci.c b/drivers/net/wireless/rt2x00/rt2800pci.c
> index 4941a1a..dc88bae 100644
> --- a/drivers/net/wireless/rt2x00/rt2800pci.c
> +++ b/drivers/net/wireless/rt2x00/rt2800pci.c
> @@ -422,7 +422,6 @@ static int rt2800pci_init_queues(struct rt2x00_dev *rt2x00dev)
> ?static void rt2800pci_toggle_irq(struct rt2x00_dev *rt2x00dev,
> ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? enum dev_state state)
> ?{
> - ? ? ? int mask = (state == STATE_RADIO_IRQ_ON);
> ? ? ? ?u32 reg;
> ? ? ? ?unsigned long flags;
>
> @@ -436,25 +435,14 @@ static void rt2800pci_toggle_irq(struct rt2x00_dev *rt2x00dev,
> ? ? ? ?}
>
> ? ? ? ?spin_lock_irqsave(&rt2x00dev->irqmask_lock, flags);
> - ? ? ? rt2x00pci_register_read(rt2x00dev, INT_MASK_CSR, &reg);
> - ? ? ? rt2x00_set_field32(&reg, INT_MASK_CSR_RXDELAYINT, 0);
> - ? ? ? rt2x00_set_field32(&reg, INT_MASK_CSR_TXDELAYINT, 0);
> - ? ? ? rt2x00_set_field32(&reg, INT_MASK_CSR_RX_DONE, mask);
> - ? ? ? rt2x00_set_field32(&reg, INT_MASK_CSR_AC0_DMA_DONE, 0);
> - ? ? ? rt2x00_set_field32(&reg, INT_MASK_CSR_AC1_DMA_DONE, 0);
> - ? ? ? rt2x00_set_field32(&reg, INT_MASK_CSR_AC2_DMA_DONE, 0);
> - ? ? ? rt2x00_set_field32(&reg, INT_MASK_CSR_AC3_DMA_DONE, 0);
> - ? ? ? rt2x00_set_field32(&reg, INT_MASK_CSR_HCCA_DMA_DONE, 0);
> - ? ? ? rt2x00_set_field32(&reg, INT_MASK_CSR_MGMT_DMA_DONE, 0);
> - ? ? ? rt2x00_set_field32(&reg, INT_MASK_CSR_MCU_COMMAND, 0);
> - ? ? ? rt2x00_set_field32(&reg, INT_MASK_CSR_RXTX_COHERENT, 0);
> - ? ? ? rt2x00_set_field32(&reg, INT_MASK_CSR_TBTT, mask);
> - ? ? ? rt2x00_set_field32(&reg, INT_MASK_CSR_PRE_TBTT, mask);
> - ? ? ? rt2x00_set_field32(&reg, INT_MASK_CSR_TX_FIFO_STATUS, mask);
> - ? ? ? rt2x00_set_field32(&reg, INT_MASK_CSR_AUTO_WAKEUP, mask);
> - ? ? ? rt2x00_set_field32(&reg, INT_MASK_CSR_GPTIMER, 0);
> - ? ? ? rt2x00_set_field32(&reg, INT_MASK_CSR_RX_COHERENT, 0);
> - ? ? ? rt2x00_set_field32(&reg, INT_MASK_CSR_TX_COHERENT, 0);
> + ? ? ? reg = 0;
> + ? ? ? if (state == STATE_RADIO_IRQ_ON) {
> + ? ? ? ? ? ? ? rt2x00_set_field32(&reg, INT_MASK_CSR_RX_DONE, 1);
> + ? ? ? ? ? ? ? rt2x00_set_field32(&reg, INT_MASK_CSR_TBTT, 1);
> + ? ? ? ? ? ? ? rt2x00_set_field32(&reg, INT_MASK_CSR_PRE_TBTT, 1);
> + ? ? ? ? ? ? ? rt2x00_set_field32(&reg, INT_MASK_CSR_TX_FIFO_STATUS, 1);
> + ? ? ? ? ? ? ? rt2x00_set_field32(&reg, INT_MASK_CSR_AUTO_WAKEUP, 1);
> + ? ? ? }
> ? ? ? ?rt2x00pci_register_write(rt2x00dev, INT_MASK_CSR, reg);
> ? ? ? ?spin_unlock_irqrestore(&rt2x00dev->irqmask_lock, flags);

Looks good to me. Thanks for finding the root cause for this issue!

Helmut

2012-01-15 09:20:44

by Ivo Van Doorn

Subject: Re: [PATCH] rt2800pci: fix spurious interrupts generation

On Fri, Jan 13, 2012 at 12:59 PM, Stanislaw Gruszka <[email protected]> wrote:
> Same devices can generate interrupt without properly setting bit in
> INT_SOURCE_CSR register (spurious interrupt), what will cause IRQ line
> will be disabled by interrupts controller driver.
>
> We discovered that clearing INT_MASK_CSR stops such behaviour. We
> previously first read that register, and then clear all know interrupt
> sources bits and do not touch reserved bits. After this patch, we write
> to all register content (I believe writing to reserved bits on that
> register will not cause any problems, I tested that on my rt2800pci
> device).
>
> This fix very bad performance problem, practically making device
> unusable (since worked without interrupts), reported in:
> https://bugzilla.redhat.com/show_bug.cgi?id=658451
>
> We previously tried to workaround that issue in commit
> 4ba7d9997869d25bd223dea7536fc1ce9fab3b3b "rt2800pci: handle spurious
> interrupts", but it was reverted in commit
> 82e5fc2a34fa9ffea38f00c4066b7e600a0ca5e6
> as thing, that will prevent to detect real spurious interrupts.
>
> Reported-and-tested-by: Amir Hedayaty <[email protected]>
> Cc: [email protected]
> Signed-off-by: Stanislaw Gruszka <[email protected]>

Acked-by: Ivo van Doorn <[email protected]>

> ---
> ?drivers/net/wireless/rt2x00/rt2800pci.c | ? 28 ++++++++--------------------
> ?1 files changed, 8 insertions(+), 20 deletions(-)
>
> diff --git a/drivers/net/wireless/rt2x00/rt2800pci.c b/drivers/net/wireless/rt2x00/rt2800pci.c
> index 4941a1a..dc88bae 100644
> --- a/drivers/net/wireless/rt2x00/rt2800pci.c
> +++ b/drivers/net/wireless/rt2x00/rt2800pci.c
> @@ -422,7 +422,6 @@ static int rt2800pci_init_queues(struct rt2x00_dev *rt2x00dev)
> ?static void rt2800pci_toggle_irq(struct rt2x00_dev *rt2x00dev,
> ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? enum dev_state state)
> ?{
> - ? ? ? int mask = (state == STATE_RADIO_IRQ_ON);
> ? ? ? ?u32 reg;
> ? ? ? ?unsigned long flags;
>
> @@ -436,25 +435,14 @@ static void rt2800pci_toggle_irq(struct rt2x00_dev *rt2x00dev,
> ? ? ? ?}
>
> ? ? ? ?spin_lock_irqsave(&rt2x00dev->irqmask_lock, flags);
> - ? ? ? rt2x00pci_register_read(rt2x00dev, INT_MASK_CSR, &reg);
> - ? ? ? rt2x00_set_field32(&reg, INT_MASK_CSR_RXDELAYINT, 0);
> - ? ? ? rt2x00_set_field32(&reg, INT_MASK_CSR_TXDELAYINT, 0);
> - ? ? ? rt2x00_set_field32(&reg, INT_MASK_CSR_RX_DONE, mask);
> - ? ? ? rt2x00_set_field32(&reg, INT_MASK_CSR_AC0_DMA_DONE, 0);
> - ? ? ? rt2x00_set_field32(&reg, INT_MASK_CSR_AC1_DMA_DONE, 0);
> - ? ? ? rt2x00_set_field32(&reg, INT_MASK_CSR_AC2_DMA_DONE, 0);
> - ? ? ? rt2x00_set_field32(&reg, INT_MASK_CSR_AC3_DMA_DONE, 0);
> - ? ? ? rt2x00_set_field32(&reg, INT_MASK_CSR_HCCA_DMA_DONE, 0);
> - ? ? ? rt2x00_set_field32(&reg, INT_MASK_CSR_MGMT_DMA_DONE, 0);
> - ? ? ? rt2x00_set_field32(&reg, INT_MASK_CSR_MCU_COMMAND, 0);
> - ? ? ? rt2x00_set_field32(&reg, INT_MASK_CSR_RXTX_COHERENT, 0);
> - ? ? ? rt2x00_set_field32(&reg, INT_MASK_CSR_TBTT, mask);
> - ? ? ? rt2x00_set_field32(&reg, INT_MASK_CSR_PRE_TBTT, mask);
> - ? ? ? rt2x00_set_field32(&reg, INT_MASK_CSR_TX_FIFO_STATUS, mask);
> - ? ? ? rt2x00_set_field32(&reg, INT_MASK_CSR_AUTO_WAKEUP, mask);
> - ? ? ? rt2x00_set_field32(&reg, INT_MASK_CSR_GPTIMER, 0);
> - ? ? ? rt2x00_set_field32(&reg, INT_MASK_CSR_RX_COHERENT, 0);
> - ? ? ? rt2x00_set_field32(&reg, INT_MASK_CSR_TX_COHERENT, 0);
> + ? ? ? reg = 0;
> + ? ? ? if (state == STATE_RADIO_IRQ_ON) {
> + ? ? ? ? ? ? ? rt2x00_set_field32(&reg, INT_MASK_CSR_RX_DONE, 1);
> + ? ? ? ? ? ? ? rt2x00_set_field32(&reg, INT_MASK_CSR_TBTT, 1);
> + ? ? ? ? ? ? ? rt2x00_set_field32(&reg, INT_MASK_CSR_PRE_TBTT, 1);
> + ? ? ? ? ? ? ? rt2x00_set_field32(&reg, INT_MASK_CSR_TX_FIFO_STATUS, 1);
> + ? ? ? ? ? ? ? rt2x00_set_field32(&reg, INT_MASK_CSR_AUTO_WAKEUP, 1);
> + ? ? ? }
> ? ? ? ?rt2x00pci_register_write(rt2x00dev, INT_MASK_CSR, reg);
> ? ? ? ?spin_unlock_irqrestore(&rt2x00dev->irqmask_lock, flags);
>
> --
> 1.7.1
>