Hi,
This series of two patches fixes the issue introduced in
cf586021642d80 ("drm/i915/gt: Pipelined page migration") where,
as reported by Matt, in a chain of requests an error is reported
only if happens in the last request.
However Chris noticed that without ensuring exclusivity in the
locking we might end up in some deadlock. That's why patch 1
throttles for the ringspace in order to make sure that no one is
holding it.
Version 1 of this patch has been reviewed by matt and this
version is adding Chris exclusive locking.
Thanks Chris for this work.
Andi
Changelog
=========
v1 -> v2
- Add patch 1 for ensuring exclusive locking of the timeline
- Reword git commit of patch 2.
Andi Shyti (1):
drm/i915/gt: Make sure that errors are propagated through request
chains
Chris Wilson (1):
drm/i915: Throttle for ringspace prior to taking the timeline mutex
drivers/gpu/drm/i915/gt/intel_context.c | 41 +++++++++++++++++++++++++
drivers/gpu/drm/i915/gt/intel_context.h | 2 ++
drivers/gpu/drm/i915/gt/intel_migrate.c | 31 +++++++++++++------
drivers/gpu/drm/i915/i915_request.c | 3 ++
4 files changed, 67 insertions(+), 10 deletions(-)
--
2.39.1
Hi,
I have received two reports [0][1] about WARN_ON_ONCE() in
sk_stream_kill_queues().
This is because the stable tree backported ca43ccf41224 ("dccp/tcp:
Avoid negative sk_forward_alloc by ipv6_pinfo.pktoptions.") without
62ec33b44e0f ("net: Remove WARN_ON_ONCE(sk->sk_forward_alloc) from
sk_stream_kill_queues().").
The reports are about 6.1.14 and 5.15.95 though, 62ec33b44e0f can be
applied cleanly on 6.1.y only, and 4.14 ~ 5.10 will have the same
issue.
So, please backport 62ec33b44e0f to 6.1.y. I will post patches for
other trees later.
[0]: https://lore.kernel.org/netdev/eb5ad452-0abe-8ea6-7e9e-1dd16852e8db@hauke-m…
[1]: https://bodhi.fedoraproject.org/updates/FEDORA-2023-3b67299c42
Thanks,
Kuniyuki
With the introduction of KERNEL_IBRS, STIBP is no longer needed
to prevent cross thread training in the kernel space. When KERNEL_IBRS
was added, it also disabled the user-mode protections for spectre_v2.
KERNEL_IBRS does not mitigate cross thread training in the userspace.
In order to demonstrate the issue, one needs to avoid syscalls in the
victim as syscalls can shorten the window size due to
a user -> kernel -> user transition which sets the
IBRS bit when entering kernel space and clearing any training the
attacker may have done.
Allow users to select a spectre_v2_user mitigation (STIBP always on,
opt-in via prctl) when KERNEL_IBRS is enabled.
Reported-by: José Oliveira <joseloliveira11(a)gmail.com>
Reported-by: Rodrigo Branco <rodrigo(a)kernelhacking.com>
Reviewed-by: Alexandra Sandulescu <aesa(a)google.com>
Reviewed-by: Jim Mattson <jmattson(a)google.com>
Fixes: 7c693f54c873 ("x86/speculation: Add spectre_v2=ibrs option to support Kernel IBRS")
Cc: stable(a)vger.kernel.org
Signed-off-by: KP Singh <kpsingh(a)kernel.org>
---
arch/x86/kernel/cpu/bugs.c | 25 +++++++++++++++++--------
1 file changed, 17 insertions(+), 8 deletions(-)
diff --git a/arch/x86/kernel/cpu/bugs.c b/arch/x86/kernel/cpu/bugs.c
index bca0bd8f4846..b05ca1575d81 100644
--- a/arch/x86/kernel/cpu/bugs.c
+++ b/arch/x86/kernel/cpu/bugs.c
@@ -1132,6 +1132,19 @@ static inline bool spectre_v2_in_ibrs_mode(enum spectre_v2_mitigation mode)
mode == SPECTRE_V2_EIBRS_LFENCE;
}
+static inline bool spectre_v2_user_no_stibp(enum spectre_v2_mitigation mode)
+{
+ /* When IBRS or enhanced IBRS is enabled, STIBP is not needed.
+ *
+ * However, With KERNEL_IBRS, the IBRS bit is cleared on return
+ * to user and the user-mode code needs to be able to enable protection
+ * from cross-thread training, either by always enabling STIBP or
+ * by enabling it via prctl.
+ */
+ return (spectre_v2_in_ibrs_mode(mode) &&
+ !cpu_feature_enabled(X86_FEATURE_KERNEL_IBRS));
+}
+
static void __init
spectre_v2_user_select_mitigation(void)
{
@@ -1193,13 +1206,8 @@ spectre_v2_user_select_mitigation(void)
"always-on" : "conditional");
}
- /*
- * If no STIBP, IBRS or enhanced IBRS is enabled, or SMT impossible,
- * STIBP is not required.
- */
- if (!boot_cpu_has(X86_FEATURE_STIBP) ||
- !smt_possible ||
- spectre_v2_in_ibrs_mode(spectre_v2_enabled))
+ if (!boot_cpu_has(X86_FEATURE_STIBP) || !smt_possible ||
+ spectre_v2_user_no_stibp(spectre_v2_enabled))
return;
/*
@@ -1496,6 +1504,7 @@ static void __init spectre_v2_select_mitigation(void)
break;
case SPECTRE_V2_IBRS:
+ pr_err("enabling KERNEL_IBRS");
setup_force_cpu_cap(X86_FEATURE_KERNEL_IBRS);
if (boot_cpu_has(X86_FEATURE_IBRS_ENHANCED))
pr_warn(SPECTRE_V2_IBRS_PERF_MSG);
@@ -2327,7 +2336,7 @@ static ssize_t mmio_stale_data_show_state(char *buf)
static char *stibp_state(void)
{
- if (spectre_v2_in_ibrs_mode(spectre_v2_enabled))
+ if (spectre_v2_user_no_stibp(spectre_v2_enabled))
return "";
switch (spectre_v2_user_stibp) {
--
2.39.2.637.g21b0678d19-goog