Received-SPF: pass (google.com: domain of linux-kernel+bounces-91779-linux.lists.archive=gmail.com@vger.kernel.org designates 2604:1380:40f1:3f00::1 as permitted sender) client-ip=2604:1380:40f1:3f00::1;
From: "bumyong.lee" <bumyong.lee@samsung.com>
To: "'karthikeyan'" <karthikeyan@linumiz.com>, <vkoul@kernel.org>
Cc: <dmaengine@vger.kernel.org>, <linux-kernel@vger.kernel.org>,
	<parthiban@linumiz.com>, <saravanan@linumiz.com>
In-Reply-To: <1553a526-6f28-4a68-88a8-f35bd22d9894@linumiz.com>
Subject: RE: dmaengine: CPU stalls while loading bluetooth module
Date: Tue, 5 Mar 2024 16:13:52 +0900
Message-ID: <000001da6ecc$adb25420$0916fc60$@samsung.com>
Precedence: bulk
MIME-Version: 1.0
Content-Transfer-Encoding: quoted-printable
Thread-Index: AQMwipOl/cAOi38P2p1VCUOGtvzp0QMWgtoHrmPNqBA=
Content-Language: ko
Content-Type: text/plain; charset="utf-8"
CMS-TYPE: 102P
DLP-Filter: Pass
References: <CGME20240305062038epcas2p143c5e1e725d8a934b0208266a2f78ccb@epcas2p1.samsung.com>
	<1553a526-6f28-4a68-88a8-f35bd22d9894@linumiz.com>

Hello.

> we have encountered CPU stalls in mainline kernel while loading the
> bluetooth module. We have custom board based on rockchip rv1109 soc and
> there is bluetooth chipset of relatek 8821cs. CPU is stalls  while realte=
k
> 8821cs module.
>=20
> Bug/Regression:
> In current mainline, we found CPU is stalls when we load bluetooth module=
.
> git bisect shows commit 22a9d9585812440211b0b34a6bc02ade62314be4
> as a bad, which produce CPU stalls.
>=20
> git show 22a9d9585812440211b0b34a6bc02ade62314be4
> commit 22a9d9585812440211b0b34a6bc02ade62314be4
> Author: Bumyong Lee <bumyong.lee=40samsung.com>
> Date:   Tue Dec 19 14:50:26 2023 +0900
>=20
>      dmaengine: pl330: issue_pending waits until WFP state
>=20
>      According to DMA-330 errata notice=5B1=5D 71930, DMAKILL
>      cannot clear internal signal, named pipeline_req_active.
>      it makes that pl330 would wait forever in WFP state
>      although dma already send dma request if pl330 gets
>      dma request before entering WFP state.
>=20
>      The errata suggests that polling until entering WFP state
>      as workaround and then peripherals allows to issue dma request.
>=20
>      =5B1=5D: https://developer.arm.com/documentation/genc008428/latest
>=20
>      Signed-off-by: Bumyong Lee <bumyong.lee=40samsung.com>
>      Link:
> https://lore.kernel.org/r/20231219055026.118695-1-bumyong.lee=40samsung.c=
om
>      Signed-off-by: Vinod Koul <vkoul=40kernel.org>
>=20
> diff --git a/drivers/dma/pl330.c b/drivers/dma/pl330.c index
> 3cf0b38387ae..c29744bfdf2c 100644
> --- a/drivers/dma/pl330.c
> +++ b/drivers/dma/pl330.c
> =40=40 -1053,6 +1053,9 =40=40 static bool _trigger(struct pl330_thread *t=
hrd)
>=20
>          thrd->req_running =3D idx;
>=20
> +       if (desc->rqtype =3D=3D DMA_MEM_TO_DEV =7C=7C desc->rqtype =3D=3D
> DMA_DEV_TO_MEM)
> +               UNTIL(thrd, PL330_STATE_WFP);
> +
>          return true;
>   =7D
>=20
> By reverting this commit, we have success in loading of bluetooth module.
>=20
>=20

> Output of CPU stalls:
> =23 modprobe hci_uart
> =5B   27.024749=5D Bluetooth: HCI UART driver ver 2.3
> =5B   27.025284=5D Bluetooth: HCI UART protocol Three-wire (H5) registere=
d
> =23 =5B   28.125338=5D dwmmc_rockchip ffc70000.mmc: Unexpected interrupt =
latency
> =5B   33.245339=5D dwmmc_rockchip ffc50000.mmc: Unexpected interrupt late=
ncy
> =5B  326.195321=5D rcu: INFO: rcu_sched detected stalls on CPUs/tasks:
> =5B  326.195880=5D rcu:     0-...0: (3 ticks this GP) idle=3De5f4/1/0x400=
00000
> softirq=3D551/552 fqs=3D420
> =5B  326.196621=5D rcu:              hardirqs   softirqs   csw/system
> =5B  326.197115=5D rcu:      number:        0          0            0
> =5B  326.197612=5D rcu:     cputime:        0          0            0   =
=3D=3D>
> 10500(ms)
> =5B  326.198231=5D rcu:     (detected by 1, t=3D2105 jiffies, g=3D-455, q=
=3D17
> ncpus=3D2)
> =5B  326.198823=5D Sending NMI from CPU 1 to CPUs 0:
>=20
> Expected Output:
> =23 modprobe hci_uart
> =5B   30.690321=5D Bluetooth: HCI UART driver ver 2.3
> =5B   30.690852=5D Bluetooth: HCI UART protocol Three-wire (H5) registere=
d
> =23 =5B   31.453586=5D Bluetooth: hci0: RTL: examining hci_ver=3D08 hci_r=
ev=3D000c
> lmp_ver=3D08 lmp_subver=3D8821
> =5B   31.458061=5D Bluetooth: hci0: RTL: rom_version status=3D0 version=
=3D1
> =5B   31.458608=5D Bluetooth: hci0: RTL: loading rtl_bt/rtl8821cs_fw.bin
> =5B   31.465029=5D Bluetooth: hci0: RTL: loading rtl_bt/rtl8821cs_config.=
bin
> =5B   31.483926=5D Bluetooth: hci0: RTL: cfg_sz 25, total sz 36953
> =5B   32.213105=5D Bluetooth: hci0: RTL: fw version 0x75b8f098
> =5B   32.274216=5D Bluetooth: MGMT ver 1.22
> =5B   32.285376=5D NET: Registered PF_ALG protocol family

I discussed this issue. Could you refer to this=5B1=5D?
I haven't received anymore reply from him after that.
If you have any more opinion, please let me know.

=5B1=5D: https://lore.kernel.org/lkml/000001da3869=24ca643fa0=245f2cbee0=24=
=40samsung.com/T/

Best Regards