This is a note to let you know that I've just added the patch titled
thunderbolt: Workaround an IOMMU fault on certain systems with Intel Maple Ridge
to the 6.1-stable tree which can be found at:
http://www.kernel.org/git/?p=linux/kernel/git/stable/stable-queue.git;a=summary
The filename of the patch is:
thunderbolt-workaround-an-iommu-fault-on-certain-systems-with-intel-maple-ridge.patch
and it can be found in the queue-6.1 subdirectory.
If you, or anyone else, feels it should not be added to the stable tree,
please let <stable@xxxxxxxxxxxxxxx> know about it.
>From 582620d9f6b352552bc9a3316fe2b1c3acd8742d Mon Sep 17 00:00:00 2001
From: Mika Westerberg <mika.westerberg@xxxxxxxxxxxxxxx>
Date: Fri, 18 Aug 2023 15:27:46 +0300
Subject: thunderbolt: Workaround an IOMMU fault on certain systems with Intel Maple Ridge
From: Mika Westerberg <mika.westerberg@xxxxxxxxxxxxxxx>
commit 582620d9f6b352552bc9a3316fe2b1c3acd8742d upstream.
On some systems the IOMMU blocks the first couple of driver ready
messages to the connection manager firmware as can be seen in below
excerpts:
thunderbolt 0000:06:00.0: AMD-Vi: Event logged [IO_PAGE_FAULT domain=0x0010 address=0xbb0e3400 flags=0x0020]
or
DMAR: DRHD: handling fault status reg 2
DMAR: [DMA Write] Request device [04:00.0] PASID ffffffff fault addr 69974000 [fault reason 05] PTE Write access is not set
The reason is unknown and hard to debug because we were not able to
reproduce this locally. This only happens on certain systems with Intel
Maple Ridge Thunderbolt controller. If there is a device connected when
the driver is loaded the issue does not happen either. Only when there
is nothing connected (so typically when the system is booted up).
We can work this around by sending the driver ready several times. After
a couple of retries the message goes through and the controller works
just fine. For this reason make the number of retries a parameter for
icm_request() and then for Maple Ridge (and Titan Ridge as they us the
same function but this should not matter) increase number of retries
while shortening the timeout accordingly.
Reported-by: Werner Sembach <wse@xxxxxxxxxxxxxxxxxxx>
Reported-by: Konrad J Hambrick <kjhambrick@xxxxxxxxx>
Reported-by: Calvin Walton <calvin.walton@xxxxxxxxxx>
Closes: https://bugzilla.kernel.org/show_bug.cgi?id=214259
Cc: stable@xxxxxxxxxxxxxxx
Signed-off-by: Mika Westerberg <mika.westerberg@xxxxxxxxxxxxxxx>
Signed-off-by: Greg Kroah-Hartman <gregkh@xxxxxxxxxxxxxxxxxxx>
---
drivers/thunderbolt/icm.c | 40 ++++++++++++++++++++--------------------
1 file changed, 20 insertions(+), 20 deletions(-)
--- a/drivers/thunderbolt/icm.c
+++ b/drivers/thunderbolt/icm.c
@@ -41,6 +41,7 @@
#define PHY_PORT_CS1_LINK_STATE_SHIFT 26
#define ICM_TIMEOUT 5000 /* ms */
+#define ICM_RETRIES 3
#define ICM_APPROVE_TIMEOUT 10000 /* ms */
#define ICM_MAX_LINK 4
@@ -296,10 +297,9 @@ static bool icm_copy(struct tb_cfg_reque
static int icm_request(struct tb *tb, const void *request, size_t request_size,
void *response, size_t response_size, size_t npackets,
- unsigned int timeout_msec)
+ int retries, unsigned int timeout_msec)
{
struct icm *icm = tb_priv(tb);
- int retries = 3;
do {
struct tb_cfg_request *req;
@@ -410,7 +410,7 @@ static int icm_fr_get_route(struct tb *t
return -ENOMEM;
ret = icm_request(tb, &request, sizeof(request), switches,
- sizeof(*switches), npackets, ICM_TIMEOUT);
+ sizeof(*switches), npackets, ICM_RETRIES, ICM_TIMEOUT);
if (ret)
goto err_free;
@@ -463,7 +463,7 @@ icm_fr_driver_ready(struct tb *tb, enum
memset(&reply, 0, sizeof(reply));
ret = icm_request(tb, &request, sizeof(request), &reply, sizeof(reply),
- 1, ICM_TIMEOUT);
+ 1, ICM_RETRIES, ICM_TIMEOUT);
if (ret)
return ret;
@@ -488,7 +488,7 @@ static int icm_fr_approve_switch(struct
memset(&reply, 0, sizeof(reply));
/* Use larger timeout as establishing tunnels can take some time */
ret = icm_request(tb, &request, sizeof(request), &reply, sizeof(reply),
- 1, ICM_APPROVE_TIMEOUT);
+ 1, ICM_RETRIES, ICM_APPROVE_TIMEOUT);
if (ret)
return ret;
@@ -515,7 +515,7 @@ static int icm_fr_add_switch_key(struct
memset(&reply, 0, sizeof(reply));
ret = icm_request(tb, &request, sizeof(request), &reply, sizeof(reply),
- 1, ICM_TIMEOUT);
+ 1, ICM_RETRIES, ICM_TIMEOUT);
if (ret)
return ret;
@@ -543,7 +543,7 @@ static int icm_fr_challenge_switch_key(s
memset(&reply, 0, sizeof(reply));
ret = icm_request(tb, &request, sizeof(request), &reply, sizeof(reply),
- 1, ICM_TIMEOUT);
+ 1, ICM_RETRIES, ICM_TIMEOUT);
if (ret)
return ret;
@@ -577,7 +577,7 @@ static int icm_fr_approve_xdomain_paths(
memset(&reply, 0, sizeof(reply));
ret = icm_request(tb, &request, sizeof(request), &reply, sizeof(reply),
- 1, ICM_TIMEOUT);
+ 1, ICM_RETRIES, ICM_TIMEOUT);
if (ret)
return ret;
@@ -1022,7 +1022,7 @@ icm_tr_driver_ready(struct tb *tb, enum
memset(&reply, 0, sizeof(reply));
ret = icm_request(tb, &request, sizeof(request), &reply, sizeof(reply),
- 1, 20000);
+ 1, 10, 2000);
if (ret)
return ret;
@@ -1055,7 +1055,7 @@ static int icm_tr_approve_switch(struct
memset(&reply, 0, sizeof(reply));
ret = icm_request(tb, &request, sizeof(request), &reply, sizeof(reply),
- 1, ICM_APPROVE_TIMEOUT);
+ 1, ICM_RETRIES, ICM_APPROVE_TIMEOUT);
if (ret)
return ret;
@@ -1083,7 +1083,7 @@ static int icm_tr_add_switch_key(struct
memset(&reply, 0, sizeof(reply));
ret = icm_request(tb, &request, sizeof(request), &reply, sizeof(reply),
- 1, ICM_TIMEOUT);
+ 1, ICM_RETRIES, ICM_TIMEOUT);
if (ret)
return ret;
@@ -1112,7 +1112,7 @@ static int icm_tr_challenge_switch_key(s
memset(&reply, 0, sizeof(reply));
ret = icm_request(tb, &request, sizeof(request), &reply, sizeof(reply),
- 1, ICM_TIMEOUT);
+ 1, ICM_RETRIES, ICM_TIMEOUT);
if (ret)
return ret;
@@ -1146,7 +1146,7 @@ static int icm_tr_approve_xdomain_paths(
memset(&reply, 0, sizeof(reply));
ret = icm_request(tb, &request, sizeof(request), &reply, sizeof(reply),
- 1, ICM_TIMEOUT);
+ 1, ICM_RETRIES, ICM_TIMEOUT);
if (ret)
return ret;
@@ -1172,7 +1172,7 @@ static int icm_tr_xdomain_tear_down(stru
memset(&reply, 0, sizeof(reply));
ret = icm_request(tb, &request, sizeof(request), &reply, sizeof(reply),
- 1, ICM_TIMEOUT);
+ 1, ICM_RETRIES, ICM_TIMEOUT);
if (ret)
return ret;
@@ -1498,7 +1498,7 @@ icm_ar_driver_ready(struct tb *tb, enum
memset(&reply, 0, sizeof(reply));
ret = icm_request(tb, &request, sizeof(request), &reply, sizeof(reply),
- 1, ICM_TIMEOUT);
+ 1, ICM_RETRIES, ICM_TIMEOUT);
if (ret)
return ret;
@@ -1524,7 +1524,7 @@ static int icm_ar_get_route(struct tb *t
memset(&reply, 0, sizeof(reply));
ret = icm_request(tb, &request, sizeof(request), &reply, sizeof(reply),
- 1, ICM_TIMEOUT);
+ 1, ICM_RETRIES, ICM_TIMEOUT);
if (ret)
return ret;
@@ -1545,7 +1545,7 @@ static int icm_ar_get_boot_acl(struct tb
memset(&reply, 0, sizeof(reply));
ret = icm_request(tb, &request, sizeof(request), &reply, sizeof(reply),
- 1, ICM_TIMEOUT);
+ 1, ICM_RETRIES, ICM_TIMEOUT);
if (ret)
return ret;
@@ -1606,7 +1606,7 @@ static int icm_ar_set_boot_acl(struct tb
memset(&reply, 0, sizeof(reply));
ret = icm_request(tb, &request, sizeof(request), &reply, sizeof(reply),
- 1, ICM_TIMEOUT);
+ 1, ICM_RETRIES, ICM_TIMEOUT);
if (ret)
return ret;
@@ -1628,7 +1628,7 @@ icm_icl_driver_ready(struct tb *tb, enum
memset(&reply, 0, sizeof(reply));
ret = icm_request(tb, &request, sizeof(request), &reply, sizeof(reply),
- 1, 20000);
+ 1, ICM_RETRIES, 20000);
if (ret)
return ret;
@@ -2300,7 +2300,7 @@ static int icm_usb4_switch_op(struct tb_
memset(&reply, 0, sizeof(reply));
ret = icm_request(tb, &request, sizeof(request), &reply, sizeof(reply),
- 1, ICM_TIMEOUT);
+ 1, ICM_RETRIES, ICM_TIMEOUT);
if (ret)
return ret;
Patches currently in stable-queue which might be from mika.westerberg@xxxxxxxxxxxxxxx are
queue-6.1/thunderbolt-workaround-an-iommu-fault-on-certain-systems-with-intel-maple-ridge.patch
queue-6.1/thunderbolt-restart-xdomain-discovery-handshake-after-failure.patch
queue-6.1/thunderbolt-check-that-lane-1-is-in-cl0-before-enabling-lane-bonding.patch
