Re: [PATCH] Revert "usb: dwc3: gadget: Prevent EP queuing while stopping transfers"
Am Donnerstag, dem 01.04.2021 um 11:09 -0700 schrieb Wesley Cheng:
>
>
> On 4/1/2021 4:55 AM, Martin Kepplinger wrote:
> > This reverts commit 9de47c3737e0c0207beb03615b320cabe495.
> >
> > Signed-off-by: Martin Kepplinger
> > ---
> >
> > I more or less blindly report:
> > commit 9de47c ("usb: dwc3: gadget: Prevent EP queuing while
> > stopping
> > transfers") results in the below error every time I connect the
> > type-c
> > connector to the dwc3, configured with serial and ethernet gadgets.
> >
> > fyi, I apply the following to dwc3 on this port:
> > dr_mode =
> > "otg";
> > snps,dis_u3_susphy_quirk;
> >
> > hnp-
> > disable;
> > srp-
> > disable;
> > adp-
> > disable;
> > usb-role-switch;
> >
> > v5.12-rc5 does not have this error so I'm not sure whether it's
> > more appropriate to add something to dwc3 than reverting. I hope
> > usb
> > people to know better and maybe even see the problem.
> >
> > thanks,
> > martin
> >
> Hi Martin,
>
> This has been fixed with the below:
> https://git.kernel.org/pub/scm/linux/kernel/git/gregkh/usb.git/commit/?h=usb-linus=5aef629704ad4d983ecf5c8a25840f16e45b6d59
>
> Can you pull that in and give it a try?
>
> Thanks
> Wesley Cheng
yes, that's a fix for my problem (and what I secretly had hoped to get
:). Thank you very much. In case it helps:
Tested-by: Martin Kepplinger
Re: [PATCH] Revert "usb: dwc3: gadget: Prevent EP queuing while stopping transfers"
On 4/1/2021 4:55 AM, Martin Kepplinger wrote:
> This reverts commit 9de47c3737e0c0207beb03615b320cabe495.
>
> Signed-off-by: Martin Kepplinger
> ---
>
> I more or less blindly report:
> commit 9de47c ("usb: dwc3: gadget: Prevent EP queuing while stopping
> transfers") results in the below error every time I connect the type-c
> connector to the dwc3, configured with serial and ethernet gadgets.
>
> fyi, I apply the following to dwc3 on this port:
> dr_mode = "otg";
> snps,dis_u3_susphy_quirk;
> hnp-disable;
> srp-disable;
> adp-disable;
> usb-role-switch;
>
> v5.12-rc5 does not have this error so I'm not sure whether it's
> more appropriate to add something to dwc3 than reverting. I hope usb
> people to know better and maybe even see the problem.
>
> thanks,
>martin
>
Hi Martin,
This has been fixed with the below:
https://git.kernel.org/pub/scm/linux/kernel/git/gregkh/usb.git/commit/?h=usb-linus=5aef629704ad4d983ecf5c8a25840f16e45b6d59
Can you pull that in and give it a try?
Thanks
Wesley Cheng
>
>
>
> [ 51.148220] [ cut here ]
> [ 51.148241] dwc3 3810.usb: No resource for ep2in
> [ 51.148376] WARNING: CPU: 0 PID: 299 at drivers/usb/dwc3/gadget.c:360
> dwc3_send_gadget_ep_cmd+0x570/0x740 [dwc3]
> [ 51.148431] Modules linked in: aes_ce_ccm rfcomm algif_hash algif_skcipher
> af_alg bnep qmi_wwan cdc_wdm usbnet mii option usb_wwan usbserial caam_jr
> caamhash_desc caamalg_desc crypto_engine mousedev redpine_sdio redpine_91x
> bluetooth uas usb_storage mac80211 st_lsm6dsx_spi aes_ce_blk crypto_simd
> crct10dif_ce cfg80211 ghash_ce sha2_ce sha1_ce bq25890_charger s5k3l6xx
> snd_soc_wm8962 hi846 vcnl4000 leds_lm3560 edt_ft5x06 ofpart
> industrialio_triggered_buffer tps6598x st_lsm6dsx_i2c st_lsm6dsx kfifo_buf
> gnss_mtk typec mxc_mipi_csi2_yav gnss_serial mx6s_capture gnss
> videobuf2_dma_contig videobuf2_memops spi_nor v4l2_fwnode videobuf2_v4l2
> videobuf2_common snd_soc_gtm601 snd_soc_simple_card videodev mtd
> snd_soc_simple_card_utils mc snd_soc_fsl_sai imx_pcm_dma caam error
> snvs_pwrkey imx_sdma virt_dma snd_soc_core imx2_wdt snd_pcm_dmaengine snd_pcm
> watchdog snd_timer snd soundcore pwm_vibra rfkill_hks rfkill ledtrig_timer
> usb_f_acm u_serial usb_f_rndis g_multi usb_f_mass_storage u_ether
> [ 51.148759] libcomposite ledtrig_pattern fuse ip_tables x_tables ipv6
> xhci_plat_hcd xhci_hcd usbcore dwc3 cdns_mhdp_imx ulpi cdns_mhdp_drmcore
> udc_core imx_dcss roles usb_common phy_fsl_imx8mq_usb clk_bd718x7
> [ 51.148837] CPU: 0 PID: 299 Comm: irq/64-dwc3 Not tainted
> 5.11.11-librem5-00334-ge4c4ff3624e9 #218
> [ 51.148848] Hardware name: Purism Librem 5r4 (DT)
> [ 51.148854] pstate: 6085 (nZCv daIf -PAN -UAO -TCO BTYPE=--)
> [ 51.148863] pc : dwc3_send_gadget_ep_cmd+0x570/0x740 [dwc3]
> [ 51.148894] lr : dwc3_send_gadget_ep_cmd+0x570/0x740 [dwc3]
> [ 51.148924] sp : 800011cb3ac0
> [ 51.148929] x29: 800011cb3ac0 x28: 032a7b00
> [ 51.148942] x27: 0327da00 x26:
> [ 51.148954] x25: ffea x24: 0006
> [ 51.148967] x23: bee1c080 x22: 800011cb3b7c
> [ 51.148979] x21: 0406 x20: bf17
> [ 51.148992] x19: 0001 x18:
> [ 51.149004] x17: x16:
> [ 51.149016] x15: x14: 8000114512c0
> [ 51.149028] x13: 1698 x12: 0040
> [ 51.149040] x11: 80001151a6f8 x10: e000
> [ 51.149052] x9 : 8000100b2b7c x8 : 80001146a6f8
> [ 51.149065] x7 : 80001151a6f8 x6 :
> [ 51.149077] x5 : bf939948 x4 :
> [ 51.149089] x3 : 0027 x2 :
> [ 51.149101] x1 : x0 : 049ae040
> [ 51.149114] Call trace:
> [ 51.149120] dwc3_send_gadget_ep_cmd+0x570/0x740 [dwc3]
> [ 51.149150] __dwc3_gadget_ep_enable+0x288/0x4fc [dwc3]
> [ 51.149179] dwc3_gadget_ep_enable+0x6c/0x15c [dwc3]
> [ 51.149209] usb_ep_enable+0x48/0x110 [udc_core]
> [ 51.149251] rndis_set_alt+0x138/0x1c0 [usb_f_rndis]
> [ 51.149276] composite_setup+0x674/0x194c [libcomposite]
> [ 51.149314] dwc3_ep0_interrupt+0x9c4/0xb9c [dwc3]
> [ 51.149344] dwc3_thread_interrupt+0x8bc/0xe74 [dwc3]
> [ 51.149374] irq_thread_fn+0x38/0xb0
> [ 51.149388] irq_thread+0x170/0x294
> [ 51.149397] kthread+0x164/0x16c
> [ 51.149407] ret_from_fork+0x10/0x34
> [ 51.149419] ---[ end trace 62c6cc2ebfb18047 ]---
> [ 51.149436] [ cut here ]
> [ 51.149440] dwc3 3810.usb: No resource for
[PATCH] Revert "usb: dwc3: gadget: Prevent EP queuing while stopping transfers"
This reverts commit 9de47c3737e0c0207beb03615b320cabe495.
Signed-off-by: Martin Kepplinger
---
I more or less blindly report:
commit 9de47c ("usb: dwc3: gadget: Prevent EP queuing while stopping
transfers") results in the below error every time I connect the type-c
connector to the dwc3, configured with serial and ethernet gadgets.
fyi, I apply the following to dwc3 on this port:
dr_mode = "otg";
snps,dis_u3_susphy_quirk;
hnp-disable;
srp-disable;
adp-disable;
usb-role-switch;
v5.12-rc5 does not have this error so I'm not sure whether it's
more appropriate to add something to dwc3 than reverting. I hope usb
people to know better and maybe even see the problem.
thanks,
martin
[ 51.148220] [ cut here ]
[ 51.148241] dwc3 3810.usb: No resource for ep2in
[ 51.148376] WARNING: CPU: 0 PID: 299 at drivers/usb/dwc3/gadget.c:360
dwc3_send_gadget_ep_cmd+0x570/0x740 [dwc3]
[ 51.148431] Modules linked in: aes_ce_ccm rfcomm algif_hash algif_skcipher
af_alg bnep qmi_wwan cdc_wdm usbnet mii option usb_wwan usbserial caam_jr
caamhash_desc caamalg_desc crypto_engine mousedev redpine_sdio redpine_91x
bluetooth uas usb_storage mac80211 st_lsm6dsx_spi aes_ce_blk crypto_simd
crct10dif_ce cfg80211 ghash_ce sha2_ce sha1_ce bq25890_charger s5k3l6xx
snd_soc_wm8962 hi846 vcnl4000 leds_lm3560 edt_ft5x06 ofpart
industrialio_triggered_buffer tps6598x st_lsm6dsx_i2c st_lsm6dsx kfifo_buf
gnss_mtk typec mxc_mipi_csi2_yav gnss_serial mx6s_capture gnss
videobuf2_dma_contig videobuf2_memops spi_nor v4l2_fwnode videobuf2_v4l2
videobuf2_common snd_soc_gtm601 snd_soc_simple_card videodev mtd
snd_soc_simple_card_utils mc snd_soc_fsl_sai imx_pcm_dma caam error snvs_pwrkey
imx_sdma virt_dma snd_soc_core imx2_wdt snd_pcm_dmaengine snd_pcm watchdog
snd_timer snd soundcore pwm_vibra rfkill_hks rfkill ledtrig_timer usb_f_acm
u_serial usb_f_rndis g_multi usb_f_mass_storage u_ether
[ 51.148759] libcomposite ledtrig_pattern fuse ip_tables x_tables ipv6
xhci_plat_hcd xhci_hcd usbcore dwc3 cdns_mhdp_imx ulpi cdns_mhdp_drmcore
udc_core imx_dcss roles usb_common phy_fsl_imx8mq_usb clk_bd718x7
[ 51.148837] CPU: 0 PID: 299 Comm: irq/64-dwc3 Not tainted
5.11.11-librem5-00334-ge4c4ff3624e9 #218
[ 51.148848] Hardware name: Purism Librem 5r4 (DT)
[ 51.148854] pstate: 6085 (nZCv daIf -PAN -UAO -TCO BTYPE=--)
[ 51.148863] pc : dwc3_send_gadget_ep_cmd+0x570/0x740 [dwc3]
[ 51.148894] lr : dwc3_send_gadget_ep_cmd+0x570/0x740 [dwc3]
[ 51.148924] sp : 800011cb3ac0
[ 51.148929] x29: 800011cb3ac0 x28: 032a7b00
[ 51.148942] x27: 0327da00 x26:
[ 51.148954] x25: ffea x24: 0006
[ 51.148967] x23: bee1c080 x22: 800011cb3b7c
[ 51.148979] x21: 0406 x20: bf17
[ 51.148992] x19: 0001 x18:
[ 51.149004] x17: x16:
[ 51.149016] x15: x14: 8000114512c0
[ 51.149028] x13: 1698 x12: 0040
[ 51.149040] x11: 80001151a6f8 x10: e000
[ 51.149052] x9 : 8000100b2b7c x8 : 80001146a6f8
[ 51.149065] x7 : 80001151a6f8 x6 :
[ 51.149077] x5 : bf939948 x4 :
[ 51.149089] x3 : 0027 x2 :
[ 51.149101] x1 : x0 : 049ae040
[ 51.149114] Call trace:
[ 51.149120] dwc3_send_gadget_ep_cmd+0x570/0x740 [dwc3]
[ 51.149150] __dwc3_gadget_ep_enable+0x288/0x4fc [dwc3]
[ 51.149179] dwc3_gadget_ep_enable+0x6c/0x15c [dwc3]
[ 51.149209] usb_ep_enable+0x48/0x110 [udc_core]
[ 51.149251] rndis_set_alt+0x138/0x1c0 [usb_f_rndis]
[ 51.149276] composite_setup+0x674/0x194c [libcomposite]
[ 51.149314] dwc3_ep0_interrupt+0x9c4/0xb9c [dwc3]
[ 51.149344] dwc3_thread_interrupt+0x8bc/0xe74 [dwc3]
[ 51.149374] irq_thread_fn+0x38/0xb0
[ 51.149388] irq_thread+0x170/0x294
[ 51.149397] kthread+0x164/0x16c
[ 51.149407] ret_from_fork+0x10/0x34
[ 51.149419] ---[ end trace 62c6cc2ebfb18047 ]---
[ 51.149436] [ cut here ]
[ 51.149440] dwc3 3810.usb: No resource for ep1in
[ 51.149507] WARNING: CPU: 0 PID: 299 at drivers/usb/dwc3/gadget.c:360
dwc3_send_gadget_ep_cmd+0x570/0x740 [dwc3]
[ 51.149543] Modules linked in: aes_ce_ccm rfcomm algif_hash algif_skcipher
af_alg bnep qmi_wwan cdc_wdm usbnet mii option usb_wwan usbserial caam_jr
caamhash_desc caamalg_desc crypto_engine mousedev redpine_sdio redpine_91x
bluetooth uas usb_storage mac80211 st_lsm6dsx_spi aes_ce_blk crypto_simd
crct10dif_ce cfg80211 ghash_ce
Re: [PATCH] Revert "usb: dwc3: gadget: Prevent EP queuing while stopping transfers"
On Thu, Apr 01, 2021 at 01:55:58PM +0200, Martin Kepplinger wrote: > This reverts commit 9de47c3737e0c0207beb03615b320cabe495. > You do not put a reason here, so I can not take this :( Please fix up and resend. thanks, greg k-h
