arm64/sme: SME related fixes
Subsystems (2)
v2 cover arm64/sme: SME related fixes
This series collates together a number of SME related fixes that have come up in testing.
v2:
- Rebase onto v6.0-rc1.
- Fix allnoconfig build.
Mark Brown (4):
arm64/signal: Raise limit on stack frames
arm64/signal: Flush FPSIMD register state when disabling streaming
mode
arm64/sme: Don't flush SVE register state when allocating SME storage arm64/sme: Don't flush SVE register state when handling SME traps
arch/arm64/include/asm/fpsimd.h | 4 ++-- arch/arm64/kernel/fpsimd.c | 21 ++++++--------------- arch/arm64/kernel/ptrace.c | 6 +++--- arch/arm64/kernel/signal.c | 14 ++++++++++++-- 4 files changed, 23 insertions(+), 22 deletions(-)
base-commit: 568035b01cfb107af8d2e4bd2fb9aea22cf5b868
Signature · 7 lines
2.30.2 _______________________________________________ linux-arm-kernel mailing list [email protected] http://lists.infradead.org/mailman/listinfo/linux-arm-kernel
PATCH v2 1/4 arm64/signal: Raise limit on stack frames
The signal code has a limit of 64K on the size of a stack frame that it will generate, if this limit is exceeded then a process will be killed if it receives a signal. Unfortunately with the advent of SME this limit is too small - the maximum possible size of the ZA register alone is 64K. This is not an issue for practical systems at present but is easily seen using virtual platforms.
Raise the limit to 256K, this is substantially more than could be used by any current architecture extension.
--- arch/arm64/kernel/signal.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)
arch/arm64/kernel/signal.c +1 -1
@@ -91,7 +91,7 @@ static size_t sigframe_size(struct rt_sigframe_user_layout const *user)* not taken into account. This limit is not a guarantee and is * NOT ABI. */ -#define SIGFRAME_MAXSZ SZ_64K +#define SIGFRAME_MAXSZ (SZ_64K * 4)static int __sigframe_alloc (struct rt_sigframe_user_layout *user,unsigned long *offset,size_t size,bool extend)
Signature · 7 lines
2.30.2 _______________________________________________ linux-arm-kernel mailing list [email protected] http://lists.infradead.org/mailman/listinfo/linux-arm-kernel
Re: [PATCH v2 1/4] arm64/signal: Raise limit on stack frames
Quoted 20 lines · Mark Brown
The signal code has a limit of 64K on the size of a stack frame that it will generate, if this limit is exceeded then a process will be killed if it receives a signal. Unfortunately with the advent of SME this limit is too small - the maximum possible size of the ZA register alone is 64K. This is not an issue for practical systems at present but is easily seen using virtual platforms.
Raise the limit to 256K, this is substantially more than could be used by any current architecture extension.
Signed-off-by Mark Brown <[email protected]>--- arch/arm64/kernel/signal.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)arch/arm64/kernel/signal.c +0
-0
arch/arm64/kernel/signal.c static size_t sigframe_size(struct rt_sigframe_user_layout const *user) · line 94
Quoted 5 lines
arch/arm64/kernel/signal.c +1
-1@@ -91,7 +91,7 @@ static size_t sigframe_size(struct rt_sigframe_user_layout const *user)* not taken into account. This limit is not a guarantee and is * NOT ABI. */ -#define SIGFRAME_MAXSZ SZ_64K +#define SIGFRAME_MAXSZ (SZ_64K * 4)
Why not SZ_256K?
Otherwise it looks fine:
_______________________________________________
linux-arm-kernel mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel
Re: [PATCH v2 1/4] arm64/signal: Raise limit on stack frames
Quoted 6 lines · Catalin Marinas
arch/arm64/kernel/signal.c static size_t sigframe_size(struct rt_sigframe_user_layout const *user) · line 94
Quoted 2 lines · Mark Brown
arch/arm64/kernel/signal.c +1
-1-#define SIGFRAME_MAXSZ SZ_64K +#define SIGFRAME_MAXSZ (SZ_64K * 4)Why not
SZ_256K?
I think I'd initially picked a different number which didn't have a SZ_ macro then forgot to check if there was one when I dithered back to 256K.
_______________________________________________
linux-arm-kernel mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel
PATCH v2 2/4 arm64/signal: Flush FPSIMD register state when disabling streaming mode
In reply to Mark BrownWhen handling a signal delivered to a context with streaming mode enabled we will disable streaming mode for the signal handler, when doing so we should also flush the saved FPSIMD register state like exiting streaming mode in the hardware would do so that if that state is reloaded we get the same behaviour. Without this we will reload whatever the last FPSIMD state that was saved for the task was.
--- arch/arm64/kernel/signal.c | 10 ++++++++++ 1 file changed, 10 insertions(+)
arch/arm64/kernel/signal.c +10 -0
@@ -926,6 +926,16 @@ static void setup_return(struct pt_regs *regs, struct k_sigaction *ka,/* Signal handlers are invoked with ZA and streaming mode disabled */ if (system_supports_sme ()) { +/* +* If we were in streaming mode the saved register +* state was SVE but we will exit SM and use the +* FPSIMD register state - flush the saved FPSIMD +* register state in case it gets loaded. +*/ +if (current->thread.svcr &SVCR_SM_MASK ) +memset (¤t->thread.uw.fpsimd_state,0 , +sizeof (current->thread.uw.fpsimd_state)); + current->thread.svcr &= ~(SVCR_ZA_MASK |SVCR_SM_MASK );sme_smstop ();
Signature · 7 lines
2.30.2 _______________________________________________ linux-arm-kernel mailing list [email protected] http://lists.infradead.org/mailman/listinfo/linux-arm-kernel
Re: [PATCH v2 2/4] arm64/signal: Flush FPSIMD register state when disabling streaming mode
Quoted 9 lines · Mark Brown
When handling a signal delivered to a context with streaming mode enabled we will disable streaming mode for the signal handler, when doing so we should also flush the saved FPSIMD register state like exiting streaming mode in the hardware would do so that if that state is reloaded we get the same behaviour. Without this we will reload whatever the last FPSIMD state that was saved for the task was.
Fixes 40a8e87bb3285 ("arm64/sme: Disable ZA and streaming mode when handling signals")Signed-off-by Mark Brown <[email protected]>
_______________________________________________
linux-arm-kernel mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel
PATCH v2 3/4 arm64/sme: Don't flush SVE register state when allocating SME storage
In reply to Mark BrownCurrently when taking a SME access trap we allocate storage for the SVE register state in order to be able to handle storage of streaming mode SVE. Due to the original usage in a purely SVE context the SVE register state allocation this also flushes the register state for SVE if storage was already allocated but in the SME context this is not desirable. For a SME access trap to be taken the task must not be in streaming mode so either there already is SVE register state present for regular SVE mode which would be corrupted or the task does not have TIF_SVE and the flush is redundant.
Fix this by adding a flag to sve_alloc() indicating if we are in a SVE context and need to flush the state. Freshly allocated storage is always zeroed either way.
--- arch/arm64/include/asm/fpsimd.h | 4 ++-- arch/arm64/kernel/fpsimd.c | 10 ++++++---- arch/arm64/kernel/ptrace.c | 6 +++--- arch/arm64/kernel/signal.c | 2 +- 4 files changed, 12 insertions(+), 10 deletions(-)
arch/arm64/include/asm/fpsimd.h +2 -2
@@ -153,7 +153,7 @@ struct vl_info {#ifdef CONFIG_ARM64_SVE -extern void sve_alloc (struct task_struct *task); +extern void sve_alloc (struct task_struct *task,bool flush);extern void fpsimd_release_task (struct task_struct *task);extern void fpsimd_sync_to_sve (struct task_struct *task);extern void fpsimd_force_sync_to_sve (struct task_struct *task);
@@ -256,7 +256,7 @@ size_t sve_state_size(struct task_struct const *task);#else /* ! CONFIG_ARM64_SVE */ -static inline void sve_alloc (struct task_struct *task) { } +static inline void sve_alloc (struct task_struct *task,bool flush) { }static inline void fpsimd_release_task (struct task_struct *task) { }static inline void sve_sync_to_fpsimd (struct task_struct *task) { }static inline void sve_sync_from_fpsimd_zeropad (struct task_struct *task) { }
arch/arm64/kernel/fpsimd.c +6 -4
@@ -715,10 +715,12 @@ size_t sve_state_size(struct task_struct const *task)* do_sve_acc() case, there is no ABI requirement to hide stale data * written previously be task. */ -void sve_alloc (struct task_struct *task) +void sve_alloc (struct task_struct *task,bool flush) {if (task->thread.sve_state) { -memset (task->thread.sve_state,0 ,sve_state_size (task)); +if (flush) +memset (task->thread.sve_state,0 , +sve_state_size (task));return ; }
@@ -1388,7 +1390,7 @@ void do_sve_acc(unsigned long esr, struct pt_regs *regs)return ; } -sve_alloc (current); +sve_alloc (current, true);if (!current->thread.sve_state) {force_sig (SIGKILL );return ;
@@ -1439,7 +1441,7 @@ void do_sme_acc(unsigned long esr, struct pt_regs *regs)return ; } -sve_alloc (current); +sve_alloc (current, false);sme_alloc (current);if (!current->thread.sve_state || !current->thread.za_state) {force_sig (SIGKILL );
arch/arm64/kernel/ptrace.c +3 -3
@@ -882,7 +882,7 @@ static int sve_set_common(struct task_struct *target,* state and ensure there's storage. */ if (target->thread.svcr != old_svcr) -sve_alloc (target); +sve_alloc (target, true); }/* Registers: FPSIMD-only case */
@@ -912,7 +912,7 @@ static int sve_set_common(struct task_struct *target, goto out; } -sve_alloc (target); +sve_alloc (target, true);if (!target->thread.sve_state) { ret = -ENOMEM ;clear_tsk_thread_flag (target,TIF_SVE );
@@ -1082,7 +1082,7 @@ static int za_set(struct task_struct *target,/* Ensure there is some SVE storage for streaming mode */ if (!target->thread.sve_state) { -sve_alloc (target); +sve_alloc (target, false);if (!target->thread.sve_state) {clear_thread_flag (TIF_SME ); ret = -ENOMEM ;
arch/arm64/kernel/signal.c +1 -1
@@ -310,7 +310,7 @@ static int restore_sve_fpsimd_context(struct user_ctxs *user)fpsimd_flush_task_state (current);/* From now, fpsimd_thread_switch() won't touch thread.sve_state */ -sve_alloc (current); +sve_alloc (current, true);if (!current->thread.sve_state) {clear_thread_flag (TIF_SVE );return -ENOMEM ;
Signature · 7 lines
2.30.2 _______________________________________________ linux-arm-kernel mailing list [email protected] http://lists.infradead.org/mailman/listinfo/linux-arm-kernel
Re: [PATCH v2 3/4] arm64/sme: Don't flush SVE register state when allocating SME storage
Quoted 9 lines · Mark Brown
Currently when taking a SME access trap we allocate storage for the SVE register state in order to be able to handle storage of streaming mode SVE. Due to the original usage in a purely SVE context the SVE register state allocation this also flushes the register state for SVE if storage was already allocated but in the SME context this is not desirable. For a SME access trap to be taken the task must not be in streaming mode so either there already is SVE register state present for regular SVE mode which would be corrupted or the task does not have
TIF_SVEand the flush is redundant.
Quoted 6 lines
Fix this by adding a flag to
sve_alloc()indicating if we are in a SVE context and need to flush the state. Freshly allocated storage is always zeroed either way.Fixes 8bd7f91c03d88 ("arm64/sme: Implement traps and syscall handling for SME")Signed-off-by Mark Brown <[email protected]>
_______________________________________________
linux-arm-kernel mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel
PATCH v2 4/4 arm64/sme: Don't flush SVE register state when handling SME traps
In reply to Mark BrownCurrently as part of handling a SME access trap we flush the SVE register state. This is not needed and would corrupt register state if the task has access to the SVE registers already. For non-streaming mode accesses the required flushing will be done in the SVE access trap. For streaming mode SVE register accesses the architecture guarantees that the register state will be flushed when streaming mode is entered or exited so there is no need for us to do so. Simply remove the register initialisation.
--- arch/arm64/kernel/fpsimd.c | 11 ----------- 1 file changed, 11 deletions(-)
arch/arm64/kernel/fpsimd.c +0 -11
@@ -1462,17 +1462,6 @@ void do_sme_acc(unsigned long esr, struct pt_regs *regs)fpsimd_bind_task_to_cpu (); } -/* -* If SVE was not already active initialise the SVE registers, -* any non-shared state between the streaming and regular SVE -* registers is architecturally guaranteed to be zeroed when -* we enter streaming mode. We do not need to initialize ZA -* since ZA must be disabled at this point and enabling ZA is -* architecturally defined to zero ZA. -*/ -if (system_supports_sve () && !test_thread_flag (TIF_SVE )) -sve_init_regs (); -put_cpu_fpsimd_context (); }
Signature · 7 lines
2.30.2 _______________________________________________ linux-arm-kernel mailing list [email protected] http://lists.infradead.org/mailman/listinfo/linux-arm-kernel
Re: [PATCH v2 4/4] arm64/sme: Don't flush SVE register state when handling SME traps
Quoted 10 lines · Mark Brown
Currently as part of handling a SME access trap we flush the SVE register state. This is not needed and would corrupt register state if the task has access to the SVE registers already. For non-streaming mode accesses the required flushing will be done in the SVE access trap. For streaming mode SVE register accesses the architecture guarantees that the register state will be flushed when streaming mode is entered or exited so there is no need for us to do so. Simply remove the register initialisation.
Fixes 8bd7f91c03d88 ("arm64/sme: Implement traps and syscall handling for SME")Signed-off-by Mark Brown <[email protected]>
_______________________________________________
linux-arm-kernel mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel