Message ID | afafa68eb3a74a414848c116382e95811f924905.1536298766.git.chunfeng.yun@mediatek.com (mailing list archive) |
---|---|
State | New, archived |
Headers | show |
Series | usb: xhci: fix interrupt transfer error happened on MTK platforms | expand |
Hi, Should add Fixes: tag and Cc: <stable@vger.kernel.org> also to apply the patch to the stable tree. Otherwise, the problem still is kept in the stable kernel. The patch really can solve the problem by testing with the ASIX AX88179 u3 Ethernet can't work since v4.17.x. So, Tested-by: Sean Wang <sean.wang@mediatek.com> and attach the log testing with AX88179 [ 126.369574] usb 2-1: new SuperSpeed Gen 1 USB device number 2 using xhci-mtk [ 126.730806] ax88179_178a 2-1:1.0 eth2: register 'ax88179_178a' at usb-1a0c0000.usb-1, ASIX AX88179 USB 3.0 Gigabit Ethernet, 00:11:6b:68:4c:9e [ 126.759932] ax88179_178a 2-1:1.0 enx00116b684c9e: renamed from eth2 ~# ifconfig enx00116b684c9e 10.10.10.1 ~# ping 10[ 147.613266] ax88179_178a 2-1:1.0 enx00116b684c9e: ax88179 - Link status is: 1 .10.10.3 PING 10.10.10.3 (10.10.10.3) 56(84) bytes of data. 64 bytes from 10.10.10.3: icmp_seq=1 ttl=64 time=0.414 ms 64 bytes from 10.10.10.3: icmp_seq=2 ttl=64 time=0.422 ms 64 bytes from 10.10.10.3: icmp_seq=3 ttl=64 time=0.376 ms 64 bytes from 10.10.10.3: icmp_seq=4 ttl=64 time=0.309 ms 64 bytes from 10.10.10.3: icmp_seq=5 ttl=64 time=0.392 ms 64 bytes from 10.10.10.3: icmp_seq=6 ttl=64 time=0.355 ms 64 bytes from 10.10.10.3: icmp_seq=7 ttl=64 time=0.407 ms 64 bytes from 10.10.10.3: icmp_seq=8 ttl=64 time=0.325 ms On Fri, 2018-09-07 at 13:51 +0800, Chunfeng Yun wrote: > The MTK xHCI controller use some reserved bytes in endpoint context for > bandwidth scheduling, so need keep them in xhci_endpoint_copy(); > > The issue is introduced by: > commit f5249461b504 ("xhci: Clear the host side toggle manually when > endpoint is soft reset") > It resets endpoints and will drop bandwidth scheduling parameters used > by interrupt or isochronous endpoints on MTK xHCI controller. > > Signed-off-by: Chunfeng Yun <chunfeng.yun@mediatek.com> > --- > drivers/usb/host/xhci-mem.c | 4 ++++ > 1 file changed, 4 insertions(+) > > diff --git a/drivers/usb/host/xhci-mem.c b/drivers/usb/host/xhci-mem.c > index ef350c3..b1f27aa 100644 > --- a/drivers/usb/host/xhci-mem.c > +++ b/drivers/usb/host/xhci-mem.c > @@ -1613,6 +1613,10 @@ void xhci_endpoint_copy(struct xhci_hcd *xhci, > in_ep_ctx->ep_info2 = out_ep_ctx->ep_info2; > in_ep_ctx->deq = out_ep_ctx->deq; > in_ep_ctx->tx_info = out_ep_ctx->tx_info; > + if (xhci->quirks & XHCI_MTK_HOST) { > + in_ep_ctx->reserved[0] = out_ep_ctx->reserved[0]; > + in_ep_ctx->reserved[1] = out_ep_ctx->reserved[1]; > + } > } > > /* Copy output xhci_slot_ctx to the input xhci_slot_ctx.
Hi Sean, I'll send a new version, thanks On Fri, 2018-09-07 at 14:31 +0800, Sean Wang wrote: > Hi, > > Should add Fixes: tag and Cc: <stable@vger.kernel.org> also to apply the patch to the stable tree. Otherwise, the problem still is kept in the stable kernel. > > The patch really can solve the problem by testing with the ASIX AX88179 u3 Ethernet can't work since v4.17.x. > > So, Tested-by: Sean Wang <sean.wang@mediatek.com> > > and attach the log testing with AX88179 > > [ 126.369574] usb 2-1: new SuperSpeed Gen 1 USB device number 2 using xhci-mtk > [ 126.730806] ax88179_178a 2-1:1.0 eth2: register 'ax88179_178a' at usb-1a0c0000.usb-1, ASIX AX88179 USB 3.0 Gigabit Ethernet, 00:11:6b:68:4c:9e > [ 126.759932] ax88179_178a 2-1:1.0 enx00116b684c9e: renamed from eth2 > > ~# ifconfig enx00116b684c9e 10.10.10.1 > ~# ping 10[ 147.613266] ax88179_178a 2-1:1.0 enx00116b684c9e: ax88179 - Link status is: 1 > .10.10.3 > PING 10.10.10.3 (10.10.10.3) 56(84) bytes of data. > 64 bytes from 10.10.10.3: icmp_seq=1 ttl=64 time=0.414 ms > 64 bytes from 10.10.10.3: icmp_seq=2 ttl=64 time=0.422 ms > 64 bytes from 10.10.10.3: icmp_seq=3 ttl=64 time=0.376 ms > 64 bytes from 10.10.10.3: icmp_seq=4 ttl=64 time=0.309 ms > 64 bytes from 10.10.10.3: icmp_seq=5 ttl=64 time=0.392 ms > 64 bytes from 10.10.10.3: icmp_seq=6 ttl=64 time=0.355 ms > 64 bytes from 10.10.10.3: icmp_seq=7 ttl=64 time=0.407 ms > 64 bytes from 10.10.10.3: icmp_seq=8 ttl=64 time=0.325 ms > > On Fri, 2018-09-07 at 13:51 +0800, Chunfeng Yun wrote: > > The MTK xHCI controller use some reserved bytes in endpoint context for > > bandwidth scheduling, so need keep them in xhci_endpoint_copy(); > > > > The issue is introduced by: > > commit f5249461b504 ("xhci: Clear the host side toggle manually when > > endpoint is soft reset") > > It resets endpoints and will drop bandwidth scheduling parameters used > > by interrupt or isochronous endpoints on MTK xHCI controller. > > > > Signed-off-by: Chunfeng Yun <chunfeng.yun@mediatek.com> > > --- > > drivers/usb/host/xhci-mem.c | 4 ++++ > > 1 file changed, 4 insertions(+) > > > > diff --git a/drivers/usb/host/xhci-mem.c b/drivers/usb/host/xhci-mem.c > > index ef350c3..b1f27aa 100644 > > --- a/drivers/usb/host/xhci-mem.c > > +++ b/drivers/usb/host/xhci-mem.c > > @@ -1613,6 +1613,10 @@ void xhci_endpoint_copy(struct xhci_hcd *xhci, > > in_ep_ctx->ep_info2 = out_ep_ctx->ep_info2; > > in_ep_ctx->deq = out_ep_ctx->deq; > > in_ep_ctx->tx_info = out_ep_ctx->tx_info; > > + if (xhci->quirks & XHCI_MTK_HOST) { > > + in_ep_ctx->reserved[0] = out_ep_ctx->reserved[0]; > > + in_ep_ctx->reserved[1] = out_ep_ctx->reserved[1]; > > + } > > } > > > > /* Copy output xhci_slot_ctx to the input xhci_slot_ctx. > >
diff --git a/drivers/usb/host/xhci-mem.c b/drivers/usb/host/xhci-mem.c index ef350c3..b1f27aa 100644 --- a/drivers/usb/host/xhci-mem.c +++ b/drivers/usb/host/xhci-mem.c @@ -1613,6 +1613,10 @@ void xhci_endpoint_copy(struct xhci_hcd *xhci, in_ep_ctx->ep_info2 = out_ep_ctx->ep_info2; in_ep_ctx->deq = out_ep_ctx->deq; in_ep_ctx->tx_info = out_ep_ctx->tx_info; + if (xhci->quirks & XHCI_MTK_HOST) { + in_ep_ctx->reserved[0] = out_ep_ctx->reserved[0]; + in_ep_ctx->reserved[1] = out_ep_ctx->reserved[1]; + } } /* Copy output xhci_slot_ctx to the input xhci_slot_ctx.
The MTK xHCI controller use some reserved bytes in endpoint context for bandwidth scheduling, so need keep them in xhci_endpoint_copy(); The issue is introduced by: commit f5249461b504 ("xhci: Clear the host side toggle manually when endpoint is soft reset") It resets endpoints and will drop bandwidth scheduling parameters used by interrupt or isochronous endpoints on MTK xHCI controller. Signed-off-by: Chunfeng Yun <chunfeng.yun@mediatek.com> --- drivers/usb/host/xhci-mem.c | 4 ++++ 1 file changed, 4 insertions(+)