| Message ID | 20210624122039.596528-1-yoshihiro.shimoda.uh@renesas.com (mailing list archive) |
|---|---|
| State | Accepted |
| Commit | 5719df243e118fb343725e8b2afb1637e1af1373 |
| Headers | show |
| Series | [PATCH/RESEND] usb: renesas_usbhs: Fix superfluous irqs happen after usb_pkt_pop() | expand |
Hi, > From: Yoshihiro Shimoda , Sent: Thursday, June 24, 2021 9:21 PM > > This driver has a potential issue which this driver is possible to > cause superfluous irqs after usb_pkt_pop() is called. So, after > the commit 3af32605289e ("usb: renesas_usbhs: fix error return > code of usbhsf_pkt_handler()") had been applied, we could observe > the following error happened when we used g_audio. > > renesas_usbhs e6590000.usb: irq_ready run_error 1 : -22 > > To fix the issue, disable the tx or rx interrupt in usb_pkt_pop(). > > Fixes: 2743e7f90dc0 ("usb: renesas_usbhs: fix the usb_pkt_pop()") > Cc: <stable@vger.kernel.org> # v4.4+ > Signed-off-by: Yoshihiro Shimoda <yoshihiro.shimoda.uh@renesas.com> > --- > I'm sorry for causing a trouble before [1], so I resend a suitable patch. > [1] > https://lore.kernel.org/linux-renesas-soc/TY2PR01MB3692555C6EAC8F02BC8B3D63D8329@TY2PR01MB3692.jpnprd01.prod.outlook.com/ I'm afraid, but would you review this patch? I confirmed this patch can be applied into v5.14-rc1. Or, should I resend this patch? Best regards, Yoshihiro Shimoda
diff --git a/drivers/usb/renesas_usbhs/fifo.c b/drivers/usb/renesas_usbhs/fifo.c index b5e7991dc7d9..a3c2b01ccf7b 100644 --- a/drivers/usb/renesas_usbhs/fifo.c +++ b/drivers/usb/renesas_usbhs/fifo.c @@ -101,6 +101,8 @@ static struct dma_chan *usbhsf_dma_chan_get(struct usbhs_fifo *fifo, #define usbhsf_dma_map(p) __usbhsf_dma_map_ctrl(p, 1) #define usbhsf_dma_unmap(p) __usbhsf_dma_map_ctrl(p, 0) static int __usbhsf_dma_map_ctrl(struct usbhs_pkt *pkt, int map); +static void usbhsf_tx_irq_ctrl(struct usbhs_pipe *pipe, int enable); +static void usbhsf_rx_irq_ctrl(struct usbhs_pipe *pipe, int enable); struct usbhs_pkt *usbhs_pkt_pop(struct usbhs_pipe *pipe, struct usbhs_pkt *pkt) { struct usbhs_priv *priv = usbhs_pipe_to_priv(pipe); @@ -123,6 +125,11 @@ struct usbhs_pkt *usbhs_pkt_pop(struct usbhs_pipe *pipe, struct usbhs_pkt *pkt) if (chan) { dmaengine_terminate_all(chan); usbhsf_dma_unmap(pkt); + } else { + if (usbhs_pipe_is_dir_in(pipe)) + usbhsf_rx_irq_ctrl(pipe, 0); + else + usbhsf_tx_irq_ctrl(pipe, 0); } usbhs_pipe_clear_without_sequence(pipe, 0, 0);
This driver has a potential issue which this driver is possible to cause superfluous irqs after usb_pkt_pop() is called. So, after the commit 3af32605289e ("usb: renesas_usbhs: fix error return code of usbhsf_pkt_handler()") had been applied, we could observe the following error happened when we used g_audio. renesas_usbhs e6590000.usb: irq_ready run_error 1 : -22 To fix the issue, disable the tx or rx interrupt in usb_pkt_pop(). Fixes: 2743e7f90dc0 ("usb: renesas_usbhs: fix the usb_pkt_pop()") Cc: <stable@vger.kernel.org> # v4.4+ Signed-off-by: Yoshihiro Shimoda <yoshihiro.shimoda.uh@renesas.com> --- I'm sorry for causing a trouble before [1], so I resend a suitable patch. [1] https://lore.kernel.org/linux-renesas-soc/TY2PR01MB3692555C6EAC8F02BC8B3D63D8329@TY2PR01MB3692.jpnprd01.prod.outlook.com/ drivers/usb/renesas_usbhs/fifo.c | 7 +++++++ 1 file changed, 7 insertions(+)