| Message ID | 20191106231650.1580-1-jeffrey.l.hugo@gmail.com (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
| Series | ath10k: Fix qmi init error handling | expand |
On Wed, Nov 06, 2019 at 03:16:50PM -0800, Jeffrey Hugo wrote: > When ath10k_qmi_init() fails, the error handling does not free the irq > resources, which causes an issue if we EPROBE_DEFER as we'll attempt to > (re-)register irqs which are already registered. > > Fixes: ba94c753ccb4 ("ath10k: add QMI message handshake for wcn3990 client") > Signed-off-by: Jeffrey Hugo <jeffrey.l.hugo@gmail.com> > --- > drivers/net/wireless/ath/ath10k/snoc.c | 2 +- > 1 file changed, 1 insertion(+), 1 deletion(-) > > diff --git a/drivers/net/wireless/ath/ath10k/snoc.c b/drivers/net/wireless/ath/ath10k/snoc.c > index fc15a0037f0e..f2a0b7aaad3b 100644 > --- a/drivers/net/wireless/ath/ath10k/snoc.c > +++ b/drivers/net/wireless/ath/ath10k/snoc.c > @@ -1729,7 +1729,7 @@ static int ath10k_snoc_probe(struct platform_device *pdev) > ret = ath10k_qmi_init(ar, msa_size); > if (ret) { > ath10k_warn(ar, "failed to register wlfw qmi client: %d\n", ret); > - goto err_core_destroy; > + goto err_free_irq; > } From a casual examination of the code this seems like a step in the right direction. But does this error path also need to call ath10k_hw_power_off() ? > > ath10k_dbg(ar, ATH10K_DBG_SNOC, "snoc probe\n"); > -- > 2.17.1 >
On Tue, Nov 12, 2019 at 1:42 AM Simon Horman <simon.horman@netronome.com> wrote: > > On Wed, Nov 06, 2019 at 03:16:50PM -0800, Jeffrey Hugo wrote: > > When ath10k_qmi_init() fails, the error handling does not free the irq > > resources, which causes an issue if we EPROBE_DEFER as we'll attempt to > > (re-)register irqs which are already registered. > > > > Fixes: ba94c753ccb4 ("ath10k: add QMI message handshake for wcn3990 client") > > Signed-off-by: Jeffrey Hugo <jeffrey.l.hugo@gmail.com> > > --- > > drivers/net/wireless/ath/ath10k/snoc.c | 2 +- > > 1 file changed, 1 insertion(+), 1 deletion(-) > > > > diff --git a/drivers/net/wireless/ath/ath10k/snoc.c b/drivers/net/wireless/ath/ath10k/snoc.c > > index fc15a0037f0e..f2a0b7aaad3b 100644 > > --- a/drivers/net/wireless/ath/ath10k/snoc.c > > +++ b/drivers/net/wireless/ath/ath10k/snoc.c > > @@ -1729,7 +1729,7 @@ static int ath10k_snoc_probe(struct platform_device *pdev) > > ret = ath10k_qmi_init(ar, msa_size); > > if (ret) { > > ath10k_warn(ar, "failed to register wlfw qmi client: %d\n", ret); > > - goto err_core_destroy; > > + goto err_free_irq; > > } > > From a casual examination of the code this seems like a step in the right > direction. But does this error path also need to call ath10k_hw_power_off() ? It probably should. I don't see any fatal errors from the step being skipped, although it might silence some regulator warnings about being left on. Unlikely to be observed by most folks as I was initing the driver pretty early to debug some things. Looks like Kalle already picked up this patch though, so I guess your suggestion would need to be a follow up.
Jeffrey Hugo <jeffrey.l.hugo@gmail.com> writes: > On Tue, Nov 12, 2019 at 1:42 AM Simon Horman <simon.horman@netronome.com> wrote: >> >> On Wed, Nov 06, 2019 at 03:16:50PM -0800, Jeffrey Hugo wrote: >> > When ath10k_qmi_init() fails, the error handling does not free the irq >> > resources, which causes an issue if we EPROBE_DEFER as we'll attempt to >> > (re-)register irqs which are already registered. >> > >> > Fixes: ba94c753ccb4 ("ath10k: add QMI message handshake for wcn3990 client") >> > Signed-off-by: Jeffrey Hugo <jeffrey.l.hugo@gmail.com> >> > --- >> > drivers/net/wireless/ath/ath10k/snoc.c | 2 +- >> > 1 file changed, 1 insertion(+), 1 deletion(-) >> > >> > diff --git a/drivers/net/wireless/ath/ath10k/snoc.c b/drivers/net/wireless/ath/ath10k/snoc.c >> > index fc15a0037f0e..f2a0b7aaad3b 100644 >> > --- a/drivers/net/wireless/ath/ath10k/snoc.c >> > +++ b/drivers/net/wireless/ath/ath10k/snoc.c >> > @@ -1729,7 +1729,7 @@ static int ath10k_snoc_probe(struct platform_device *pdev) >> > ret = ath10k_qmi_init(ar, msa_size); >> > if (ret) { >> > ath10k_warn(ar, "failed to register wlfw qmi client: %d\n", ret); >> > - goto err_core_destroy; >> > + goto err_free_irq; >> > } >> >> From a casual examination of the code this seems like a step in the right >> direction. But does this error path also need to call ath10k_hw_power_off() ? > > It probably should. I don't see any fatal errors from the step being > skipped, although it might silence some regulator warnings about being > left on. Unlikely to be observed by most folks as I was initing the > driver pretty early to debug some things. Looks like Kalle already > picked up this patch though, so I guess your suggestion would need to > be a follow up. Actually it's only in the pending branch, which means that the patch can be changed or a new version can be submitted: https://wireless.wiki.kernel.org/en/users/drivers/ath10k/submittingpatches#patch_flow The easiest way to check the state of a wireless patch is from patchwork: https://patchwork.kernel.org/patch/11231325/ https://wireless.wiki.kernel.org/en/developers/documentation/submittingpatches#checking_state_of_patches_from_patchwork
On Tue, Nov 12, 2019 at 9:57 PM Kalle Valo <kvalo@codeaurora.org> wrote: > > Jeffrey Hugo <jeffrey.l.hugo@gmail.com> writes: > > > On Tue, Nov 12, 2019 at 1:42 AM Simon Horman <simon.horman@netronome.com> wrote: > >> > >> On Wed, Nov 06, 2019 at 03:16:50PM -0800, Jeffrey Hugo wrote: > >> > When ath10k_qmi_init() fails, the error handling does not free the irq > >> > resources, which causes an issue if we EPROBE_DEFER as we'll attempt to > >> > (re-)register irqs which are already registered. > >> > > >> > Fixes: ba94c753ccb4 ("ath10k: add QMI message handshake for wcn3990 client") > >> > Signed-off-by: Jeffrey Hugo <jeffrey.l.hugo@gmail.com> > >> > --- > >> > drivers/net/wireless/ath/ath10k/snoc.c | 2 +- > >> > 1 file changed, 1 insertion(+), 1 deletion(-) > >> > > >> > diff --git a/drivers/net/wireless/ath/ath10k/snoc.c b/drivers/net/wireless/ath/ath10k/snoc.c > >> > index fc15a0037f0e..f2a0b7aaad3b 100644 > >> > --- a/drivers/net/wireless/ath/ath10k/snoc.c > >> > +++ b/drivers/net/wireless/ath/ath10k/snoc.c > >> > @@ -1729,7 +1729,7 @@ static int ath10k_snoc_probe(struct platform_device *pdev) > >> > ret = ath10k_qmi_init(ar, msa_size); > >> > if (ret) { > >> > ath10k_warn(ar, "failed to register wlfw qmi client: %d\n", ret); > >> > - goto err_core_destroy; > >> > + goto err_free_irq; > >> > } > >> > >> From a casual examination of the code this seems like a step in the right > >> direction. But does this error path also need to call ath10k_hw_power_off() ? > > > > It probably should. I don't see any fatal errors from the step being > > skipped, although it might silence some regulator warnings about being > > left on. Unlikely to be observed by most folks as I was initing the > > driver pretty early to debug some things. Looks like Kalle already > > picked up this patch though, so I guess your suggestion would need to > > be a follow up. > > Actually it's only in the pending branch, which means that the patch can > be changed or a new version can be submitted: Thats an interesting flow. Ok. > > https://wireless.wiki.kernel.org/en/users/drivers/ath10k/submittingpatches#patch_flow > > The easiest way to check the state of a wireless patch is from > patchwork: > > https://patchwork.kernel.org/patch/11231325/ > > https://wireless.wiki.kernel.org/en/developers/documentation/submittingpatches#checking_state_of_patches_from_patchwork > > -- > https://wireless.wiki.kernel.org/en/developers/documentation/submittingpatches
diff --git a/drivers/net/wireless/ath/ath10k/snoc.c b/drivers/net/wireless/ath/ath10k/snoc.c index fc15a0037f0e..f2a0b7aaad3b 100644 --- a/drivers/net/wireless/ath/ath10k/snoc.c +++ b/drivers/net/wireless/ath/ath10k/snoc.c @@ -1729,7 +1729,7 @@ static int ath10k_snoc_probe(struct platform_device *pdev) ret = ath10k_qmi_init(ar, msa_size); if (ret) { ath10k_warn(ar, "failed to register wlfw qmi client: %d\n", ret); - goto err_core_destroy; + goto err_free_irq; } ath10k_dbg(ar, ATH10K_DBG_SNOC, "snoc probe\n");
When ath10k_qmi_init() fails, the error handling does not free the irq resources, which causes an issue if we EPROBE_DEFER as we'll attempt to (re-)register irqs which are already registered. Fixes: ba94c753ccb4 ("ath10k: add QMI message handshake for wcn3990 client") Signed-off-by: Jeffrey Hugo <jeffrey.l.hugo@gmail.com> --- drivers/net/wireless/ath/ath10k/snoc.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)