From: Basavaraj Natikar <bnatikar@amd.com>
To: Mika Westerberg <mika.westerberg@linux.intel.com>,
Basavaraj Natikar <Basavaraj.Natikar@amd.com>
Cc: andreas.noever@gmail.com, westeri@kernel.org,
YehezkelShB@gmail.com, linux-usb@vger.kernel.org,
Mario.Limonciello@amd.com, Sanath S <Sanath.S@amd.com>
Subject: Re: [PATCH -next] thunderbolt: Add quirk to reset host interface on DMA path teardown for AMD USB4 routers
Date: Wed, 5 Aug 2026 16:40:43 +0530 [thread overview]
Message-ID: <7624adcb-77bb-4556-8e43-eeaa6735eb74@amd.com> (raw)
In-Reply-To: <20260805034908.GC235112@black.igk.intel.com>
On 8/5/2026 9:19 AM, Mika Westerberg wrote:
> Hi,
>
> On Tue, Aug 04, 2026 at 05:56:38PM +0530, Basavaraj Natikar wrote:
>> Some AMD USB4 host routers have a bug in the Host Interface where
>> DMA path setup and teardown cycles may cause the Tx ring to hang.
>>
>> Fix this by issuing a Host Interface Reset on every DMA path teardown
>> for affected routers. The Host Interface Reset brings the registers in
>> the memory BAR to their default state and clears the End-to-End Flow
>> Control state, preventing the hang condition.
>>
>> Co-developed-by: Sanath S <Sanath.S@amd.com>
>> Signed-off-by: Sanath S <Sanath.S@amd.com>
>> Signed-off-by: Basavaraj Natikar <Basavaraj.Natikar@amd.com>
>> ---
>> drivers/thunderbolt/nhi.c | 48 ++++++++++++++++++++++++++++++++++
>> drivers/thunderbolt/nhi_regs.h | 4 +++
>> drivers/thunderbolt/quirks.c | 16 ++++++++++++
>> drivers/thunderbolt/tb.c | 7 +++++
>> drivers/thunderbolt/tb.h | 3 +++
>> 5 files changed, 78 insertions(+)
>>
>> diff --git a/drivers/thunderbolt/nhi.c b/drivers/thunderbolt/nhi.c
>> index 383a36212f70..0865cab582e7 100644
>> --- a/drivers/thunderbolt/nhi.c
>> +++ b/drivers/thunderbolt/nhi.c
>> @@ -1160,6 +1160,54 @@ static void nhi_reset(struct tb_nhi *nhi)
>> dev_warn(nhi->dev, "timeout resetting host router\n");
>> }
>>
>> +/**
>> + * nhi_host_interface_reset() - Issue a host interface reset
>> + * @nhi: Pointer to the NHI structure
>> + *
>> + * Resets the Host Interface by setting the RST bit in the Host Interface
>> + * Reset register. This brings the registers in the memory BAR to their
>> + * default state and clears the End-to-End Flow Control state.
>> + *
>> + * The caller must ensure that the control channel (Ring 0) is stopped
>> + * before calling this function, since the reset clears ring state.
>> + * The caller is responsible for restarting Ring 0 afterward.
>> + *
>> + * After setting the RST bit, waits for tHIReset (10 ms) for the reset
>> + * to complete.
>> + */
>> +static void nhi_host_interface_reset(struct tb_nhi *nhi)
>> +{
>> + struct device *dev = nhi->dev;
>> + u32 val;
>> +
>> + val = ioread32(nhi->iobase + REG_CAPS);
>> + /* Host Interface Reset only applies to Ver. 1 routers */
>> + if (FIELD_GET(REG_CAPS_VERSION_MASK, val) >= REG_CAPS_VERSION_2)
>> + return;
>> +
>> + dev_dbg(dev, "issuing host interface reset\n");
>> +
>> + iowrite32(REG_HOST_INTERFACE_RESET_RST,
>> + nhi->iobase + REG_HOST_INTERFACE_RESET);
>> +
>> + /* Wait for tHIReset (10 ms) for the reset to complete */
>> + usleep_range(10000, 20000);
>> +}
>> +
>> +/**
>> + * tb_nhi_host_interface_reset() - Reset host interface with control channel
>> + * @tb: Pointer to the thunderbolt domain
>> + *
>> + * Stops the control channel, issues a Host Interface Reset, and restarts
>> + * the control channel.
>> + */
>> +void tb_nhi_host_interface_reset(struct tb *tb)
>> +{
>> + tb_ctl_stop(tb->ctl);
>> + nhi_host_interface_reset(tb->nhi);
>> + tb_ctl_start(tb->ctl);
>> +}
>> +
>> static struct tb *nhi_select_cm(struct tb_nhi *nhi)
>> {
>> struct tb *tb;
>> diff --git a/drivers/thunderbolt/nhi_regs.h b/drivers/thunderbolt/nhi_regs.h
>> index d6a197fabc74..99df60b6db36 100644
>> --- a/drivers/thunderbolt/nhi_regs.h
>> +++ b/drivers/thunderbolt/nhi_regs.h
>> @@ -115,6 +115,10 @@ struct ring_desc {
>> #define REG_CAPS_VERSION_MASK GENMASK(23, 16)
>> #define REG_CAPS_VERSION_2 0x40
>>
>> +/* Host Interface Reset - resets TX/RX rings and E2E flow control counters */
>> +#define REG_HOST_INTERFACE_RESET 0x39858
>> +#define REG_HOST_INTERFACE_RESET_RST BIT(0)
>> +
>> #define REG_DMA_MISC 0x39864
>> #define REG_DMA_MISC_INT_AUTO_CLEAR BIT(2)
>> #define REG_DMA_MISC_DISABLE_AUTO_CLEAR BIT(17)
>> diff --git a/drivers/thunderbolt/quirks.c b/drivers/thunderbolt/quirks.c
>> index 9f7914ac2f48..cc04d0873931 100644
>> --- a/drivers/thunderbolt/quirks.c
>> +++ b/drivers/thunderbolt/quirks.c
>> @@ -52,6 +52,12 @@ static void quirk_block_rpm_in_redrive(struct tb_switch *sw)
>> tb_sw_dbg(sw, "preventing runtime PM in DP redrive mode\n");
>> }
>>
>> +static void quirk_host_interface_reset(struct tb_switch *sw)
>> +{
>> + sw->quirks |= QUIRK_HOST_INTERFACE_RESET;
>> + tb_sw_dbg(sw, "enabling host interface reset on DMA path teardown\n");
>> +}
>> +
>> struct tb_quirk {
>> u16 hw_vendor_id;
>> u16 hw_device_id;
>> @@ -114,6 +120,16 @@ static const struct tb_quirk tb_quirks[] = {
>> { 0x0438, 0x0209, 0x0000, 0x0000, quirk_clx_disable },
>> { 0x0438, 0x020a, 0x0000, 0x0000, quirk_clx_disable },
>> { 0x0438, 0x020b, 0x0000, 0x0000, quirk_clx_disable },
>> + /*
>> + * AMD USB4 host routers may hang the Tx ring after repeated
>> + * DMA path teardowns. Issue a Host Interface Reset on each
>> + * teardown to prevent the hang.
>> + */
>> + { 0x0438, 0x020d, 0x0000, 0x0000, quirk_host_interface_reset },
>> + { 0x0438, 0x020e, 0x0000, 0x0000, quirk_host_interface_reset },
>> + { 0x0438, 0x020f, 0x0000, 0x0000, quirk_host_interface_reset },
>> + { 0x0438, 0x0210, 0x0000, 0x0000, quirk_host_interface_reset },
>> + { 0x0438, 0x0211, 0x0000, 0x0000, quirk_host_interface_reset },
> These quirks are for things inside USB4 fabric. Since this is the HI, we
> have quirks for those in pci.c/nhi.c like QUIRK_AUTO_CLEAR_INT. Can you
> make this one follow that convention instead?
>
>> };
>>
>> /**
>> diff --git a/drivers/thunderbolt/tb.c b/drivers/thunderbolt/tb.c
>> index 47753a5c0f2e..d40cc9e57364 100644
>> --- a/drivers/thunderbolt/tb.c
>> +++ b/drivers/thunderbolt/tb.c
>> @@ -2395,6 +2395,13 @@ static void __tb_disconnect_xdomain_paths(struct tb *tb, struct tb_xdomain *xd,
>> * the same host router USB4 downstream port.
>> */
>> tb_enable_clx(sw);
>> +
>> + /*
>> + * Some host routers may hang the Tx ring after DMA path teardowns.
>> + * Issue a Host Interface Reset to prevent it.
>> + */
>> + if (tb->root_switch->quirks & QUIRK_HOST_INTERFACE_RESET)
>> + tb_nhi_host_interface_reset(tb);
>> }
>>
>> static int tb_disconnect_xdomain_paths(struct tb *tb, struct tb_xdomain *xd,
>> diff --git a/drivers/thunderbolt/tb.h b/drivers/thunderbolt/tb.h
>> index 4373336d9425..d21feb631f3e 100644
>> --- a/drivers/thunderbolt/tb.h
>> +++ b/drivers/thunderbolt/tb.h
>> @@ -26,6 +26,8 @@
>> #define QUIRK_NO_CLX BIT(1)
>> /* Need to keep power on while USB4 port is in redrive mode */
>> #define QUIRK_KEEP_POWER_IN_DP_REDRIVE BIT(2)
>> +/* Reset Host Interface on DMA path teardown to prevent Tx ring hang */
>> +#define QUIRK_HOST_INTERFACE_RESET BIT(3)
>>
>> /**
>> * struct tb_nvm - Structure holding NVM information
>> @@ -1507,6 +1509,7 @@ static inline bool usb4_port_device_is_offline(const struct usb4_port *usb4)
>> }
>>
>> void tb_check_quirks(struct tb_switch *sw);
>> +void tb_nhi_host_interface_reset(struct tb *tb);
> Also we don't need to expose this then (and if it needs to be exposed it
> belongs to nhi.h not here.).
Sure Mika, I'll address both in v2:
- move the quirk to the NHI - set nhi->quirks in nhi_pci_check_quirks()
by PCI ID, following the QUIRK_AUTO_CLEAR_INT convention.
- move the helper declaration to nhi.h (out of tb.h).
Will send v2 shortly.
Thanks,
--
Basavaraj
>
>>
>> #ifdef CONFIG_ACPI
>> bool tb_acpi_add_links(struct tb_nhi *nhi);
>> --
>> 2.34.1
prev parent reply other threads:[~2026-08-05 11:10 UTC|newest]
Thread overview: 3+ messages / expand[flat|nested] mbox.gz Atom feed top
2026-08-04 12:26 [PATCH -next] thunderbolt: Add quirk to reset host interface on DMA path teardown for AMD USB4 routers Basavaraj Natikar
2026-08-05 3:49 ` Mika Westerberg
2026-08-05 11:10 ` Basavaraj Natikar [this message]
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=7624adcb-77bb-4556-8e43-eeaa6735eb74@amd.com \
--to=bnatikar@amd.com \
--cc=Basavaraj.Natikar@amd.com \
--cc=Mario.Limonciello@amd.com \
--cc=Sanath.S@amd.com \
--cc=YehezkelShB@gmail.com \
--cc=andreas.noever@gmail.com \
--cc=linux-usb@vger.kernel.org \
--cc=mika.westerberg@linux.intel.com \
--cc=westeri@kernel.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox