When only PHY1 is used (for example on Odroid-HC4), the regmap init code
uses the usb2 ports when doesn't initialize the PHY1 regmap entry.
This fixes:
Unable to handle kernel NULL pointer dereference at virtual address 0000000000000020
...
pc : regmap_update_bits_base+0x40/0xa0
lr : dwc3_meson_g12a_usb2_init_phy+0x4c/0xf8
...
Call trace:
regmap_update_bits_base+0x40/0xa0
dwc3_meson_g12a_usb2_init_phy+0x4c/0xf8
dwc3_meson_g12a_usb2_init+0x7c/0xc8
dwc3_meson_g12a_usb_init+0x28/0x48
dwc3_meson_g12a_probe+0x298/0x540
platform_probe+0x70/0xe0
really_probe+0xf0/0x4d8
driver_probe_device+0xfc/0x168
...
Fixes: 013af227f58a97 ("usb: dwc3: meson-g12a: handle the phy and glue registers separately")
Signed-off-by: Neil Armstrong <[email protected]>
---
drivers/usb/dwc3/dwc3-meson-g12a.c | 5 ++++-
1 file changed, 4 insertions(+), 1 deletion(-)
diff --git a/drivers/usb/dwc3/dwc3-meson-g12a.c b/drivers/usb/dwc3/dwc3-meson-g12a.c
index bdf1f98dfad8..9c944bc6d555 100644
--- a/drivers/usb/dwc3/dwc3-meson-g12a.c
+++ b/drivers/usb/dwc3/dwc3-meson-g12a.c
@@ -651,7 +651,7 @@ static int dwc3_meson_g12a_setup_regmaps(struct dwc3_meson_g12a *priv,
return PTR_ERR(priv->usb_glue_regmap);
/* Create a regmap for each USB2 PHY control register set */
- for (i = 0; i < priv->usb2_ports; i++) {
+ for (i = 0; i < priv->drvdata->num_phys; i++) {
struct regmap_config u2p_regmap_config = {
.reg_bits = 8,
.val_bits = 32,
@@ -659,6 +659,9 @@ static int dwc3_meson_g12a_setup_regmaps(struct dwc3_meson_g12a *priv,
.max_register = U2P_R1,
};
+ if (!strstr(priv->drvdata->phy_names[i], "usb2"))
+ continue;
+
u2p_regmap_config.name = devm_kasprintf(priv->dev, GFP_KERNEL,
"u2p-%d", i);
if (!u2p_regmap_config.name)
--
2.25.1
On Tue, Jun 1, 2021 at 10:49 AM Neil Armstrong <[email protected]> wrote:
>
> When only PHY1 is used (for example on Odroid-HC4), the regmap init code
> uses the usb2 ports when doesn't initialize the PHY1 regmap entry.
>
> This fixes:
> Unable to handle kernel NULL pointer dereference at virtual address 0000000000000020
> ...
> pc : regmap_update_bits_base+0x40/0xa0
> lr : dwc3_meson_g12a_usb2_init_phy+0x4c/0xf8
> ...
> Call trace:
> regmap_update_bits_base+0x40/0xa0
> dwc3_meson_g12a_usb2_init_phy+0x4c/0xf8
> dwc3_meson_g12a_usb2_init+0x7c/0xc8
> dwc3_meson_g12a_usb_init+0x28/0x48
> dwc3_meson_g12a_probe+0x298/0x540
> platform_probe+0x70/0xe0
> really_probe+0xf0/0x4d8
> driver_probe_device+0xfc/0x168
> ...
>
> Fixes: 013af227f58a97 ("usb: dwc3: meson-g12a: handle the phy and glue registers separately")
> Signed-off-by: Neil Armstrong <[email protected]>
Reviewed-by: Martin Blumenstingl <[email protected]>
On Tue, Jun 01, 2021 at 10:48:30AM +0200, Neil Armstrong wrote:
> When only PHY1 is used (for example on Odroid-HC4), the regmap init code
> uses the usb2 ports when doesn't initialize the PHY1 regmap entry.
<snip>
Meta comment, you signed this, and it worked!
Looking up https://lore.kernel.org/r/20210601084830.260196-1-narmstrong%40baylibre.com
Grabbing thread from lore.kernel.org/linux-amlogic/20210601084830.260196-1-narmstrong%40baylibre.com/t.mbox.gz
Analyzing 2 messages in the thread
Checking attestation on all messages, may take a moment...
---
✓ [PATCH] usb: dwc3-meson-g12a: fix usb2 PHY glue init when phy0 is disabled
+ Reviewed-by: Martin Blumenstingl <[email protected]> (✓ DKIM/lists.infradead.org)
---
✓ Signed: openpgp/[email protected]
✗ BADSIG: DKIM/baylibre-com.20150623.gappssmtp.com
✓ Signed: DKIM/lists.infradead.org (From: [email protected])
---
Total patches: 1
---
Link: https://lore.kernel.org/r/[email protected]
Base: not found
Applying: usb: dwc3-meson-g12a: fix usb2 PHY glue init when phy0 is disabled
Odd that DKIM didn't work for baylibre-com, but hey, I'll take a real
signature over DKIM anyday!
thanks for doing this.
greg k-h
On Fri, Jun 04, 2021 at 12:58:16PM +0200, Greg KH wrote:
> Meta comment, you signed this, and it worked!
NOICE. \o/
> Looking up https://lore.kernel.org/r/20210601084830.260196-1-narmstrong%40baylibre.com
> Grabbing thread from lore.kernel.org/linux-amlogic/20210601084830.260196-1-narmstrong%40baylibre.com/t.mbox.gz
> Analyzing 2 messages in the thread
> Checking attestation on all messages, may take a moment...
> ---
> ✓ [PATCH] usb: dwc3-meson-g12a: fix usb2 PHY glue init when phy0 is disabled
> + Reviewed-by: Martin Blumenstingl <[email protected]> (✓ DKIM/lists.infradead.org)
> ---
> ✓ Signed: openpgp/[email protected]
> ✗ BADSIG: DKIM/baylibre-com.20150623.gappssmtp.com
> ✓ Signed: DKIM/lists.infradead.org (From: [email protected])
> ---
> Total patches: 1
> ---
> Link: https://lore.kernel.org/r/[email protected]
> Base: not found
> Applying: usb: dwc3-meson-g12a: fix usb2 PHY glue init when phy0 is disabled
>
>
> Odd that DKIM didn't work for baylibre-com, but hey, I'll take a real
> signature over DKIM anyday!
That lookup happened to grab the thread from linux-amlogic, which is
mailman2-managed and is known to break DKIM. I'll try to fix our configuration
so that known-DKIM-friendly sources are given priority. This way, when a thread
exists on multiple lists, you'll get the one more likely to pass DKIM checks.
> thanks for doing this.
Yes, thanks! Hopefully, more people will start doing this.
-K
On Fri, Jun 04, 2021 at 11:07:42AM -0400, Konstantin Ryabitsev wrote:
> > Odd that DKIM didn't work for baylibre-com, but hey, I'll take a real
> > signature over DKIM anyday!
>
> That lookup happened to grab the thread from linux-amlogic, which is
> mailman2-managed and is known to break DKIM. I'll try to fix our configuration
> so that known-DKIM-friendly sources are given priority. This way, when a thread
> exists on multiple lists, you'll get the one more likely to pass DKIM checks.
This is now in place -- lore will now prefer results from DKIM-friendly
sources. E.g. grabbing the same message-id will now return the thread from
linux-usb (via vger) instead of linux-amlogic (via infradead).
$ b4 am -o/tmp [email protected]
Looking up https://lore.kernel.org/r/20210601084830.260196-1-narmstrong%40baylibre.com
Grabbing thread from lore.kernel.org/linux-usb/20210601084830.260196-1-narmstrong%40baylibre.com/t.mbox.gz
Analyzing 4 messages in the thread
Checking attestation on all messages, may take a moment...
---
✓ [PATCH] usb: dwc3-meson-g12a: fix usb2 PHY glue init when phy0 is disabled
+ Reviewed-by: Martin Blumenstingl <[email protected]> (✓ DKIM/googlemail.com)
---
✓ Signed: openpgp/[email protected]
✓ Signed: DKIM/baylibre-com.20150623.gappssmtp.com (From: [email protected])
---
Total patches: 1
---
Link: https://lore.kernel.org/r/[email protected]
Base: not found
git am /tmp/20210601_narmstrong_usb_dwc3_meson_g12a_fix_usb2_phy_glue_init_when_phy0_is_disabled.mbx
-K
On Fri, Jun 04, 2021 at 12:46:01PM -0400, Konstantin Ryabitsev wrote:
> On Fri, Jun 04, 2021 at 11:07:42AM -0400, Konstantin Ryabitsev wrote:
> > > Odd that DKIM didn't work for baylibre-com, but hey, I'll take a real
> > > signature over DKIM anyday!
> >
> > That lookup happened to grab the thread from linux-amlogic, which is
> > mailman2-managed and is known to break DKIM. I'll try to fix our configuration
> > so that known-DKIM-friendly sources are given priority. This way, when a thread
> > exists on multiple lists, you'll get the one more likely to pass DKIM checks.
>
> This is now in place -- lore will now prefer results from DKIM-friendly
> sources. E.g. grabbing the same message-id will now return the thread from
> linux-usb (via vger) instead of linux-amlogic (via infradead).
>
> $ b4 am -o/tmp [email protected]
> Looking up https://lore.kernel.org/r/20210601084830.260196-1-narmstrong%40baylibre.com
> Grabbing thread from lore.kernel.org/linux-usb/20210601084830.260196-1-narmstrong%40baylibre.com/t.mbox.gz
> Analyzing 4 messages in the thread
> Checking attestation on all messages, may take a moment...
> ---
> ✓ [PATCH] usb: dwc3-meson-g12a: fix usb2 PHY glue init when phy0 is disabled
> + Reviewed-by: Martin Blumenstingl <[email protected]> (✓ DKIM/googlemail.com)
> ---
> ✓ Signed: openpgp/[email protected]
> ✓ Signed: DKIM/baylibre-com.20150623.gappssmtp.com (From: [email protected])
> ---
> Total patches: 1
> ---
> Link: https://lore.kernel.org/r/[email protected]
> Base: not found
> git am /tmp/20210601_narmstrong_usb_dwc3_meson_g12a_fix_usb2_phy_glue_init_when_phy0_is_disabled.mbx
This is great, thanks for changing this on the lore backend, and for all
of this work in the firstplace.
greg k-h
On 04/06/2021 18:46, Konstantin Ryabitsev wrote:
> On Fri, Jun 04, 2021 at 11:07:42AM -0400, Konstantin Ryabitsev wrote:
>>> Odd that DKIM didn't work for baylibre-com, but hey, I'll take a real
>>> signature over DKIM anyday!
>>
>> That lookup happened to grab the thread from linux-amlogic, which is
>> mailman2-managed and is known to break DKIM. I'll try to fix our configuration
>> so that known-DKIM-friendly sources are given priority. This way, when a thread
>> exists on multiple lists, you'll get the one more likely to pass DKIM checks.
>
> This is now in place -- lore will now prefer results from DKIM-friendly
> sources. E.g. grabbing the same message-id will now return the thread from
> linux-usb (via vger) instead of linux-amlogic (via infradead).
>
> $ b4 am -o/tmp [email protected]
> Looking up https://lore.kernel.org/r/20210601084830.260196-1-narmstrong%40baylibre.com
> Grabbing thread from lore.kernel.org/linux-usb/20210601084830.260196-1-narmstrong%40baylibre.com/t.mbox.gz
> Analyzing 4 messages in the thread
> Checking attestation on all messages, may take a moment...
> ---
> ✓ [PATCH] usb: dwc3-meson-g12a: fix usb2 PHY glue init when phy0 is disabled
> + Reviewed-by: Martin Blumenstingl <[email protected]> (✓ DKIM/googlemail.com)
> ---
> ✓ Signed: openpgp/[email protected]
> ✓ Signed: DKIM/baylibre-com.20150623.gappssmtp.com (From: [email protected])
> ---
> Total patches: 1
> ---
> Link: https://lore.kernel.org/r/[email protected]
> Base: not found
> git am /tmp/20210601_narmstrong_usb_dwc3_meson_g12a_fix_usb2_phy_glue_init_when_phy0_is_disabled.mbx
>
> -K
>
Nice, thanks for this !
Neil