Stable Archive on lore.kernel.org
 help / color / Atom feed
* [PATCH for-4.4-stable] ath10k: restore QCA9880-AR1A (v1) detection
@ 2019-11-29 20:53 Christian Lamparter
  2019-12-02 18:42 ` Greg KH
  0 siblings, 1 reply; 4+ messages in thread
From: Christian Lamparter @ 2019-11-29 20:53 UTC (permalink / raw)
  To: stable

commit f8914a14623a79b73f72b2b1ee4cd9b2cb91b735 upstream
---
From f8914a14623a79b73f72b2b1ee4cd9b2cb91b735 Mon Sep 17 00:00:00 2001
From: Christian Lamparter <chunkeey@gmail.com>
Date: Mon, 25 Mar 2019 13:50:19 +0100
Subject: [PATCH 4.4] ath10k: restore QCA9880-AR1A (v1) detection
To: linux-wireless@vger.kernel.org,
    ath10k@lists.infradead.org
Cc: Kalle Valo <kvalo@codeaurora.org>

This patch restores the old behavior that read
the chip_id on the QCA988x before resetting the
chip. This needs to be done in this order since
the unsupported QCA988x AR1A chips fall off the
bus when resetted. Otherwise the next MMIO Op
after the reset causes a BUS ERROR and panic.

Cc: stable@vger.kernel.org # 4.4
Fixes: 1a7fecb766c8 ("ath10k: reset chip before reading chip_id in probe")
Signed-off-by: Christian Lamparter <chunkeey@gmail.com>
---
 drivers/net/wireless/ath/ath10k/pci.c | 36 +++++++++++++++++++--------
 1 file changed, 25 insertions(+), 11 deletions(-)

diff --git a/drivers/net/wireless/ath/ath10k/pci.c b/drivers/net/wireless/ath/ath10k/pci.c
--- a/drivers/net/wireless/ath/ath10k/pci.c	2019-09-08 00:07:21.374565470 +0200
+++ b/drivers/net/wireless/ath/ath10k/pci.c	2019-09-08 00:17:15.365912133 +0200
@@ -2988,12 +2988,13 @@ static int ath10k_pci_probe(struct pci_d
 	struct ath10k_pci *ar_pci;
 	enum ath10k_hw_rev hw_rev;
 	u32 chip_id;
-	bool pci_ps;
+	bool pci_ps, is_qca988x = false;
 
 	switch (pci_dev->device) {
 	case QCA988X_2_0_DEVICE_ID:
 		hw_rev = ATH10K_HW_QCA988X;
 		pci_ps = false;
+		is_qca988x = true;
 		break;
 	case QCA6164_2_1_DEVICE_ID:
 	case QCA6174_2_1_DEVICE_ID:
@@ -3087,6 +3088,19 @@ static int ath10k_pci_probe(struct pci_d
 		goto err_deinit_irq;
 	}
 
+	/* Read CHIP_ID before reset to catch QCA9880-AR1A v1 devices that
+	 * fall off the bus during chip_reset. These chips have the same pci
+	 * device id as the QCA9880 BR4A or 2R4E. So that's why the check.
+	 */
+	if (is_qca988x) {
+		chip_id = ath10k_pci_soc_read32(ar, SOC_CHIP_ID_ADDRESS);
+		if (chip_id != 0xffffffff) {
+			if (!ath10k_pci_chip_is_supported(pdev->device,
+							  chip_id))
+				goto err_unsupported;
+		}
+	}
+
 	ret = ath10k_pci_chip_reset(ar);
 	if (ret) {
 		ath10k_err(ar, "failed to reset chip: %d\n", ret);
@@ -3099,11 +3113,8 @@ static int ath10k_pci_probe(struct pci_d
 		goto err_free_irq;
 	}
 
-	if (!ath10k_pci_chip_is_supported(pdev->device, chip_id)) {
-		ath10k_err(ar, "device %04x with chip_id %08x isn't supported\n",
-			   pdev->device, chip_id);
-		goto err_free_irq;
-	}
+	if (!ath10k_pci_chip_is_supported(pdev->device, chip_id))
+		goto err_unsupported;
 
 	ret = ath10k_core_register(ar, chip_id);
 	if (ret) {
@@ -3113,6 +3124,10 @@ static int ath10k_pci_probe(struct pci_d
 
 	return 0;
 
+err_unsupported:
+	ath10k_err(ar, "device %04x with chip_id %08x isn't supported\n",
+		   pdev->device, bus_params.chip_id);
+
 err_free_irq:
 	ath10k_pci_free_irq(ar);
 	ath10k_pci_kill_tasklet(ar);




^ permalink raw reply	[flat|nested] 4+ messages in thread

* Re: [PATCH for-4.4-stable] ath10k: restore QCA9880-AR1A (v1) detection
  2019-11-29 20:53 [PATCH for-4.4-stable] ath10k: restore QCA9880-AR1A (v1) detection Christian Lamparter
@ 2019-12-02 18:42 ` Greg KH
  2019-12-03 20:47   ` Christian Lamparter
  0 siblings, 1 reply; 4+ messages in thread
From: Greg KH @ 2019-12-02 18:42 UTC (permalink / raw)
  To: Christian Lamparter; +Cc: stable

On Fri, Nov 29, 2019 at 09:53:50PM +0100, Christian Lamparter wrote:
> commit f8914a14623a79b73f72b2b1ee4cd9b2cb91b735 upstream
> ---
> >From f8914a14623a79b73f72b2b1ee4cd9b2cb91b735 Mon Sep 17 00:00:00 2001
> From: Christian Lamparter <chunkeey@gmail.com>
> Date: Mon, 25 Mar 2019 13:50:19 +0100
> Subject: [PATCH 4.4] ath10k: restore QCA9880-AR1A (v1) detection
> To: linux-wireless@vger.kernel.org,
>     ath10k@lists.infradead.org
> Cc: Kalle Valo <kvalo@codeaurora.org>
> 
> This patch restores the old behavior that read
> the chip_id on the QCA988x before resetting the
> chip. This needs to be done in this order since
> the unsupported QCA988x AR1A chips fall off the
> bus when resetted. Otherwise the next MMIO Op
> after the reset causes a BUS ERROR and panic.
> 
> Cc: stable@vger.kernel.org # 4.4
> Fixes: 1a7fecb766c8 ("ath10k: reset chip before reading chip_id in probe")
> Signed-off-by: Christian Lamparter <chunkeey@gmail.com>
> ---
>  drivers/net/wireless/ath/ath10k/pci.c | 36 +++++++++++++++++++--------
>  1 file changed, 25 insertions(+), 11 deletions(-)
> 
> diff --git a/drivers/net/wireless/ath/ath10k/pci.c b/drivers/net/wireless/ath/ath10k/pci.c
> --- a/drivers/net/wireless/ath/ath10k/pci.c	2019-09-08 00:07:21.374565470 +0200
> +++ b/drivers/net/wireless/ath/ath10k/pci.c	2019-09-08 00:17:15.365912133 +0200
> @@ -2988,12 +2988,13 @@ static int ath10k_pci_probe(struct pci_d
>  	struct ath10k_pci *ar_pci;
>  	enum ath10k_hw_rev hw_rev;
>  	u32 chip_id;
> -	bool pci_ps;
> +	bool pci_ps, is_qca988x = false;
>  
>  	switch (pci_dev->device) {
>  	case QCA988X_2_0_DEVICE_ID:
>  		hw_rev = ATH10K_HW_QCA988X;
>  		pci_ps = false;
> +		is_qca988x = true;
>  		break;
>  	case QCA6164_2_1_DEVICE_ID:
>  	case QCA6174_2_1_DEVICE_ID:
> @@ -3087,6 +3088,19 @@ static int ath10k_pci_probe(struct pci_d
>  		goto err_deinit_irq;
>  	}
>  
> +	/* Read CHIP_ID before reset to catch QCA9880-AR1A v1 devices that
> +	 * fall off the bus during chip_reset. These chips have the same pci
> +	 * device id as the QCA9880 BR4A or 2R4E. So that's why the check.
> +	 */
> +	if (is_qca988x) {
> +		chip_id = ath10k_pci_soc_read32(ar, SOC_CHIP_ID_ADDRESS);
> +		if (chip_id != 0xffffffff) {
> +			if (!ath10k_pci_chip_is_supported(pdev->device,
> +							  chip_id))
> +				goto err_unsupported;
> +		}
> +	}
> +
>  	ret = ath10k_pci_chip_reset(ar);
>  	if (ret) {
>  		ath10k_err(ar, "failed to reset chip: %d\n", ret);
> @@ -3099,11 +3113,8 @@ static int ath10k_pci_probe(struct pci_d
>  		goto err_free_irq;
>  	}
>  
> -	if (!ath10k_pci_chip_is_supported(pdev->device, chip_id)) {
> -		ath10k_err(ar, "device %04x with chip_id %08x isn't supported\n",
> -			   pdev->device, chip_id);
> -		goto err_free_irq;
> -	}
> +	if (!ath10k_pci_chip_is_supported(pdev->device, chip_id))
> +		goto err_unsupported;
>  
>  	ret = ath10k_core_register(ar, chip_id);
>  	if (ret) {
> @@ -3113,6 +3124,10 @@ static int ath10k_pci_probe(struct pci_d
>  
>  	return 0;
>  
> +err_unsupported:
> +	ath10k_err(ar, "device %04x with chip_id %08x isn't supported\n",
> +		   pdev->device, bus_params.chip_id);

Backports are great, but as I mentioned before, this breaks the build,
so we can't take it:

drivers/net/wireless/ath/ath10k/pci.c: In function ‘ath10k_pci_probe’:
drivers/net/wireless/ath/ath10k/pci.c:3129:20: error: ‘bus_params’ undeclared (first use in this function); did you mean ‘key_params’?
 3129 |      pdev->device, bus_params.chip_id);
      |                    ^~~~~~~~~~
      |                    key_params


Please fix this up and resend backports again.

thanks,

greg k-h

^ permalink raw reply	[flat|nested] 4+ messages in thread

* Re: [PATCH for-4.4-stable] ath10k: restore QCA9880-AR1A (v1) detection
  2019-12-02 18:42 ` Greg KH
@ 2019-12-03 20:47   ` Christian Lamparter
  2019-12-03 21:12     ` Greg KH
  0 siblings, 1 reply; 4+ messages in thread
From: Christian Lamparter @ 2019-12-03 20:47 UTC (permalink / raw)
  To: Greg KH; +Cc: Stable

On Mon, Dec 2, 2019 at 7:42 PM Greg KH <gregkh@linuxfoundation.org> wrote:
>
> On Fri, Nov 29, 2019 at 09:53:50PM +0100, Christian Lamparter wrote:
> > commit f8914a14623a79b73f72b2b1ee4cd9b2cb91b735 upstream
> > ---
> > >From f8914a14623a79b73f72b2b1ee4cd9b2cb91b735 Mon Sep 17 00:00:00 2001
> > From: Christian Lamparter <chunkeey@gmail.com>
> > Date: Mon, 25 Mar 2019 13:50:19 +0100
> > Subject: [PATCH 4.4] ath10k: restore QCA9880-AR1A (v1) detection
> > To: linux-wireless@vger.kernel.org,
> >     ath10k@lists.infradead.org
> > Cc: Kalle Valo <kvalo@codeaurora.org>
> >
> > This patch restores the old behavior that read
> > the chip_id on the QCA988x before resetting the
> > chip. This needs to be done in this order since
> > the unsupported QCA988x AR1A chips fall off the
> > bus when resetted. Otherwise the next MMIO Op
> > after the reset causes a BUS ERROR and panic.
> >
> > Cc: stable@vger.kernel.org # 4.4
> > Fixes: 1a7fecb766c8 ("ath10k: reset chip before reading chip_id in probe")
> > Signed-off-by: Christian Lamparter <chunkeey@gmail.com>
> > ---
> >  drivers/net/wireless/ath/ath10k/pci.c | 36 +++++++++++++++++++--------
> >  1 file changed, 25 insertions(+), 11 deletions(-)
> >
> > diff --git a/drivers/net/wireless/ath/ath10k/pci.c b/drivers/net/wireless/ath/ath10k/pci.c
> > --- a/drivers/net/wireless/ath/ath10k/pci.c   2019-09-08 00:07:21.374565470 +0200
> > +++ b/drivers/net/wireless/ath/ath10k/pci.c   2019-09-08 00:17:15.365912133 +0200
> > @@ -2988,12 +2988,13 @@ static int ath10k_pci_probe(struct pci_d
> >       struct ath10k_pci *ar_pci;
> >       enum ath10k_hw_rev hw_rev;
> >       u32 chip_id;
> > -     bool pci_ps;
> > +     bool pci_ps, is_qca988x = false;
> >
> >       switch (pci_dev->device) {
> >       case QCA988X_2_0_DEVICE_ID:
> >               hw_rev = ATH10K_HW_QCA988X;
> >               pci_ps = false;
> > +             is_qca988x = true;
> >               break;
> >       case QCA6164_2_1_DEVICE_ID:
> >       case QCA6174_2_1_DEVICE_ID:
> > @@ -3087,6 +3088,19 @@ static int ath10k_pci_probe(struct pci_d
> >               goto err_deinit_irq;
> >       }
> >
> > +     /* Read CHIP_ID before reset to catch QCA9880-AR1A v1 devices that
> > +      * fall off the bus during chip_reset. These chips have the same pci
> > +      * device id as the QCA9880 BR4A or 2R4E. So that's why the check.
> > +      */
> > +     if (is_qca988x) {
> > +             chip_id = ath10k_pci_soc_read32(ar, SOC_CHIP_ID_ADDRESS);
> > +             if (chip_id != 0xffffffff) {
> > +                     if (!ath10k_pci_chip_is_supported(pdev->device,
> > +                                                       chip_id))
> > +                             goto err_unsupported;
> > +             }
> > +     }
> > +
> >       ret = ath10k_pci_chip_reset(ar);
> >       if (ret) {
> >               ath10k_err(ar, "failed to reset chip: %d\n", ret);
> > @@ -3099,11 +3113,8 @@ static int ath10k_pci_probe(struct pci_d
> >               goto err_free_irq;
> >       }
> >
> > -     if (!ath10k_pci_chip_is_supported(pdev->device, chip_id)) {
> > -             ath10k_err(ar, "device %04x with chip_id %08x isn't supported\n",
> > -                        pdev->device, chip_id);
> > -             goto err_free_irq;
> > -     }
> > +     if (!ath10k_pci_chip_is_supported(pdev->device, chip_id))
> > +             goto err_unsupported;
> >
> >       ret = ath10k_core_register(ar, chip_id);
> >       if (ret) {
> > @@ -3113,6 +3124,10 @@ static int ath10k_pci_probe(struct pci_d
> >
> >       return 0;
> >
> > +err_unsupported:
> > +     ath10k_err(ar, "device %04x with chip_id %08x isn't supported\n",
> > +                pdev->device, bus_params.chip_id);
>
> Backports are great, but as I mentioned before, this breaks the build,
> so we can't take it:
>
> drivers/net/wireless/ath/ath10k/pci.c: In function ‘ath10k_pci_probe’:
> drivers/net/wireless/ath/ath10k/pci.c:3129:20: error: ‘bus_params’ undeclared (first use in this function); did you mean ‘key_params’?
>  3129 |      pdev->device, bus_params.chip_id);
>       |                    ^~~~~~~~~~
>       |                    key_params
>
Ok, I see. That "bus_params." there has to go so just "chip_id" remains.

But, I'm sorry that I caused this havoc.
For a little bit of background: Getting the QCA9880-AR1A and the board
it works back, will
take until next year. Since this card either doesn't fit (it's
oversized mini-pcie) or flat-out
doesn't work (it crashes the vast majority of x86/AMD64 system during post).
So doing tests with the hardware involves going through custom OpenWrt-builds.

> Please fix this up and resend backports again.
I guess, I can do that. Question is, given the bug has been around
since 2015 and the
response for this patch has been negative (even from those who have
crashes from it).
I'm thinking that I'll rather drop this and take the time off by doing
something else for
a while.

Yours,
Christian

^ permalink raw reply	[flat|nested] 4+ messages in thread

* Re: [PATCH for-4.4-stable] ath10k: restore QCA9880-AR1A (v1) detection
  2019-12-03 20:47   ` Christian Lamparter
@ 2019-12-03 21:12     ` Greg KH
  0 siblings, 0 replies; 4+ messages in thread
From: Greg KH @ 2019-12-03 21:12 UTC (permalink / raw)
  To: Christian Lamparter; +Cc: Stable

On Tue, Dec 03, 2019 at 09:47:38PM +0100, Christian Lamparter wrote:
> On Mon, Dec 2, 2019 at 7:42 PM Greg KH <gregkh@linuxfoundation.org> wrote:
> >
> > On Fri, Nov 29, 2019 at 09:53:50PM +0100, Christian Lamparter wrote:
> > > commit f8914a14623a79b73f72b2b1ee4cd9b2cb91b735 upstream
> > > ---
> > > >From f8914a14623a79b73f72b2b1ee4cd9b2cb91b735 Mon Sep 17 00:00:00 2001
> > > From: Christian Lamparter <chunkeey@gmail.com>
> > > Date: Mon, 25 Mar 2019 13:50:19 +0100
> > > Subject: [PATCH 4.4] ath10k: restore QCA9880-AR1A (v1) detection
> > > To: linux-wireless@vger.kernel.org,
> > >     ath10k@lists.infradead.org
> > > Cc: Kalle Valo <kvalo@codeaurora.org>
> > >
> > > This patch restores the old behavior that read
> > > the chip_id on the QCA988x before resetting the
> > > chip. This needs to be done in this order since
> > > the unsupported QCA988x AR1A chips fall off the
> > > bus when resetted. Otherwise the next MMIO Op
> > > after the reset causes a BUS ERROR and panic.
> > >
> > > Cc: stable@vger.kernel.org # 4.4
> > > Fixes: 1a7fecb766c8 ("ath10k: reset chip before reading chip_id in probe")
> > > Signed-off-by: Christian Lamparter <chunkeey@gmail.com>
> > > ---
> > >  drivers/net/wireless/ath/ath10k/pci.c | 36 +++++++++++++++++++--------
> > >  1 file changed, 25 insertions(+), 11 deletions(-)
> > >
> > > diff --git a/drivers/net/wireless/ath/ath10k/pci.c b/drivers/net/wireless/ath/ath10k/pci.c
> > > --- a/drivers/net/wireless/ath/ath10k/pci.c   2019-09-08 00:07:21.374565470 +0200
> > > +++ b/drivers/net/wireless/ath/ath10k/pci.c   2019-09-08 00:17:15.365912133 +0200
> > > @@ -2988,12 +2988,13 @@ static int ath10k_pci_probe(struct pci_d
> > >       struct ath10k_pci *ar_pci;
> > >       enum ath10k_hw_rev hw_rev;
> > >       u32 chip_id;
> > > -     bool pci_ps;
> > > +     bool pci_ps, is_qca988x = false;
> > >
> > >       switch (pci_dev->device) {
> > >       case QCA988X_2_0_DEVICE_ID:
> > >               hw_rev = ATH10K_HW_QCA988X;
> > >               pci_ps = false;
> > > +             is_qca988x = true;
> > >               break;
> > >       case QCA6164_2_1_DEVICE_ID:
> > >       case QCA6174_2_1_DEVICE_ID:
> > > @@ -3087,6 +3088,19 @@ static int ath10k_pci_probe(struct pci_d
> > >               goto err_deinit_irq;
> > >       }
> > >
> > > +     /* Read CHIP_ID before reset to catch QCA9880-AR1A v1 devices that
> > > +      * fall off the bus during chip_reset. These chips have the same pci
> > > +      * device id as the QCA9880 BR4A or 2R4E. So that's why the check.
> > > +      */
> > > +     if (is_qca988x) {
> > > +             chip_id = ath10k_pci_soc_read32(ar, SOC_CHIP_ID_ADDRESS);
> > > +             if (chip_id != 0xffffffff) {
> > > +                     if (!ath10k_pci_chip_is_supported(pdev->device,
> > > +                                                       chip_id))
> > > +                             goto err_unsupported;
> > > +             }
> > > +     }
> > > +
> > >       ret = ath10k_pci_chip_reset(ar);
> > >       if (ret) {
> > >               ath10k_err(ar, "failed to reset chip: %d\n", ret);
> > > @@ -3099,11 +3113,8 @@ static int ath10k_pci_probe(struct pci_d
> > >               goto err_free_irq;
> > >       }
> > >
> > > -     if (!ath10k_pci_chip_is_supported(pdev->device, chip_id)) {
> > > -             ath10k_err(ar, "device %04x with chip_id %08x isn't supported\n",
> > > -                        pdev->device, chip_id);
> > > -             goto err_free_irq;
> > > -     }
> > > +     if (!ath10k_pci_chip_is_supported(pdev->device, chip_id))
> > > +             goto err_unsupported;
> > >
> > >       ret = ath10k_core_register(ar, chip_id);
> > >       if (ret) {
> > > @@ -3113,6 +3124,10 @@ static int ath10k_pci_probe(struct pci_d
> > >
> > >       return 0;
> > >
> > > +err_unsupported:
> > > +     ath10k_err(ar, "device %04x with chip_id %08x isn't supported\n",
> > > +                pdev->device, bus_params.chip_id);
> >
> > Backports are great, but as I mentioned before, this breaks the build,
> > so we can't take it:
> >
> > drivers/net/wireless/ath/ath10k/pci.c: In function ‘ath10k_pci_probe’:
> > drivers/net/wireless/ath/ath10k/pci.c:3129:20: error: ‘bus_params’ undeclared (first use in this function); did you mean ‘key_params’?
> >  3129 |      pdev->device, bus_params.chip_id);
> >       |                    ^~~~~~~~~~
> >       |                    key_params
> >
> Ok, I see. That "bus_params." there has to go so just "chip_id" remains.
> 
> But, I'm sorry that I caused this havoc.
> For a little bit of background: Getting the QCA9880-AR1A and the board
> it works back, will
> take until next year. Since this card either doesn't fit (it's
> oversized mini-pcie) or flat-out
> doesn't work (it crashes the vast majority of x86/AMD64 system during post).
> So doing tests with the hardware involves going through custom OpenWrt-builds.
> 
> > Please fix this up and resend backports again.
> I guess, I can do that. Question is, given the bug has been around
> since 2015 and the
> response for this patch has been negative (even from those who have
> crashes from it).
> I'm thinking that I'll rather drop this and take the time off by doing
> something else for
> a while.

That's fine, if someone has noticed this, then they just can use a new
kernel version.

thanks,

greg k-h

^ permalink raw reply	[flat|nested] 4+ messages in thread

end of thread, back to index

Thread overview: 4+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2019-11-29 20:53 [PATCH for-4.4-stable] ath10k: restore QCA9880-AR1A (v1) detection Christian Lamparter
2019-12-02 18:42 ` Greg KH
2019-12-03 20:47   ` Christian Lamparter
2019-12-03 21:12     ` Greg KH

Stable Archive on lore.kernel.org

Archives are clonable:
	git clone --mirror https://lore.kernel.org/stable/0 stable/git/0.git

	# If you have public-inbox 1.1+ installed, you may
	# initialize and index your mirror using the following commands:
	public-inbox-init -V2 stable stable/ https://lore.kernel.org/stable \
		stable@vger.kernel.org
	public-inbox-index stable

Example config snippet for mirrors

Newsgroup available over NNTP:
	nntp://nntp.lore.kernel.org/org.kernel.vger.stable


AGPL code for this site: git clone https://public-inbox.org/public-inbox.git