irqchip: riscv: Order normal writes and IPI writes
Subsystems (3)
PATCH RESEND 1/1 irqchip: riscv: Order normal writes and IPI writes
RISC-V distinguishes between normal memory accesses and device I/O and uses FENCE instruction to order them as viewed by othe RISC-V harts and external devices or coprocessors. The FENCE instruction can order any combination of device input(I), device output(O), memory reads(R) and memory writes(W). For example, 'fence w, o' can be used to ensure all memory writes from instructions preceding the FENCE instruction appear earlier in the global memory order than device output writes from instructions after the FENCE instruction.
RISC-V issues IPI by writing certain value to IMSIC/ACLINT MMIO registers, which is regarded as device output operation. However, the existing implementation of IMSIC/ACLINT driver issues IPI via writel_relaxed(), which does not guarantee the order of device output operation and preceding memory writes. Then the hart receiving IPI may not have seen the latest data yet.
This commit fixes this by replacing writel_relaxed() with writel() when issuing IPI, which will use 'fence w, o' to ensure all previous writes made by current hart are visible to other harts before they receive the IPI.
--- drivers/irqchip/irq-riscv-imsic-early.c | 2 +- drivers/irqchip/irq-thead-c900-aclint-sswi.c | 2 +- 2 files changed, 2 insertions(+), 2 deletions(-)
drivers/irqchip/irq-riscv-imsic-early.c +1 -1
@@ -27,7 +27,7 @@ static void imsic_ipi_send(unsigned int cpu) {struct imsic_local_config *local =per_cpu_ptr (imsic->global.local, cpu); -writel_relaxed (IMSIC_IPI_ID , local->msi_va); +writel (IMSIC_IPI_ID , local->msi_va); }static void imsic_ipi_starting_cpu (void )
drivers/irqchip/irq-thead-c900-aclint-sswi.c +1 -1
@@ -31,7 +31,7 @@ static DEFINE_PER_CPU(void __iomem *, sswi_cpu_regs);static void thead_aclint_sswi_ipi_send (unsigned int cpu) { -writel_relaxed (0x1 ,per_cpu (sswi_cpu_regs, cpu)); +writel (0x1 ,per_cpu (sswi_cpu_regs, cpu)); }static void thead_aclint_sswi_ipi_clear (void )
Signature · 1 line
2.20.1
[tip: irq/urgent] irqchip/riscv: Ensure ordering of memory writes and IPI writes
The following commit has been merged into the irq/urgent branch of tip:
The following commit has been merged into the irq/urgent branch of tip:
Commit-ID: 825c78e6a60c309a59d18d5ac5968aa79cef0bd6 Gitweb: https://git.kernel.org/tip/825c78e6a60c309a59d18d5ac5968aa79cef0bd6 Author: Xu Lu <[email protected]> AuthorDate: Mon, 27 Jan 2025 17:38:46 +08:00 Committer: Thomas Gleixner <[email protected]> CommitterDate: Mon, 27 Jan 2025 11:07:03 +01:00
irqchip/riscv: Ensure ordering of memory writes and IPI writes
RISC-V distinguishes between memory accesses and device I/O and uses FENCE instruction to order them as viewed by other RISC-V harts and external devices or coprocessors. The FENCE instruction can order any combination of device input(I), device output(O), memory reads(R) and memory writes(W). For example, 'fence w, o' is used to ensure all memory writes from instructions preceding the FENCE instruction appear earlier in the global memory order than device output writes from instructions after the FENCE instruction.
RISC-V issues IPIs by writing to the IMSIC/ACLINT MMIO registers, which is regarded as device output operation. However, the existing implementation of the IMSIC/ACLINT drivers issue the IPI via writel_relaxed(), which does not guarantee the order of device output operation and preceding memory writes. As a consequence the hart receiving the IPI might not observe the IPI related data.
Fix this by replacing writel_relaxed() with writel() when issuing IPIs, which uses 'fence w, o' to ensure all previous writes made by the current hart are visible to other harts before they receive the IPI.
--- drivers/irqchip/irq-riscv-imsic-early.c | 2 +- drivers/irqchip/irq-thead-c900-aclint-sswi.c | 2 +- 2 files changed, 2 insertions(+), 2 deletions(-)
drivers/irqchip/irq-riscv-imsic-early.c +1 -1
@@ -27,7 +27,7 @@ static void imsic_ipi_send(unsigned int cpu) {struct imsic_local_config *local =per_cpu_ptr (imsic->global.local, cpu); -writel_relaxed (IMSIC_IPI_ID , local->msi_va); +writel (IMSIC_IPI_ID , local->msi_va); }static void imsic_ipi_starting_cpu (void )
drivers/irqchip/irq-thead-c900-aclint-sswi.c +1 -1
@@ -31,7 +31,7 @@ static DEFINE_PER_CPU(void __iomem *, sswi_cpu_regs);static void thead_aclint_sswi_ipi_send (unsigned int cpu) { -writel_relaxed (0x1 ,per_cpu (sswi_cpu_regs, cpu)); +writel (0x1 ,per_cpu (sswi_cpu_regs, cpu)); }static void thead_aclint_sswi_ipi_clear (void )
On Mon, Jan 27 2025 at 17:38, Xu Lu wrote:
This is not a RESEND. The change log has been modified, no?
The prefix is incorrect. See
https://www.kernel.org/doc/html/latest/process/maintainer-tip.html
Quoted 1 line
RISC-V distinguishes between normal memory accesses and device I/O and
What is a normal memory write? Are there abnormal memory writes too?
Quoted 4 lines
uses FENCE instruction to order them as viewed by othe RISC-V harts and external devices or coprocessors. The FENCE instruction can order any combination of device input(I), device output(O), memory reads(R) and memory writes(W). For example, 'fence w, o' can be used to ensure all
Can be? It _is_ used, no?
Quoted 11 lines
memory writes from instructions preceding the FENCE instruction appear earlier in the global memory order than device output writes from instructions after the FENCE instruction.
RISC-V issues IPI by writing certain value to IMSIC/ACLINT MMIO registers, which is regarded as device output operation. However, the existing implementation of IMSIC/ACLINT driver issues IPI via
writel_relaxed(), which does not guarantee the order of device output operation and preceding memory writes. Then the hart receiving IPI may not have seen the latest data yet.
Quoted 2 lines
This commit fixes this by replacing
writel_relaxed()withwritel()when
'This commit' is equally wrong as 'This patch'. See Documentation/process/
Quoted 3 lines
issuing IPI, which will use 'fence w, o' to ensure all previous writes made by current hart are visible to other harts before they receive the IPI.
I've fixed it up for you.
Thanks,
tglx
Quoted 4 lines · Thomas Gleixner
On Mon, Jan 27 2025 at 17:38, Xu Lu wrote:
This is not a RESEND. The change log has been modified, no?
Sure, the change log has been modified. I will pay attention next time.
Quoted 8 lines
The prefix is incorrect. See
https://www.kernel.org/doc/html/latest/process/maintainer-tip.html
Quoted 1 line
RISC-V distinguishes between normal memory accesses and device I/O and
What is a normal memory write? Are there abnormal memory writes too?
Sorry for the misleading. By normal memory write, I mean memory writes and want to distinguish it from MMIO writes.
Quoted 7 lines
Quoted 5 lines
uses FENCE instruction to order them as viewed by othe RISC-V harts and external devices or coprocessors. The FENCE instruction can order any combination of device input(I), device output(O), memory reads(R) and memory writes(W). For example, 'fence w, o' can be used to ensure all
Can be? It _is_ used, no?
Yes, it _is_ used. 'Can be' is not accurate.
Quoted 12 lines
Quoted 12 lines
memory writes from instructions preceding the FENCE instruction appear earlier in the global memory order than device output writes from instructions after the FENCE instruction.
RISC-V issues IPI by writing certain value to IMSIC/ACLINT MMIO registers, which is regarded as device output operation. However, the existing implementation of IMSIC/ACLINT driver issues IPI via
writel_relaxed(), which does not guarantee the order of device output operation and preceding memory writes. Then the hart receiving IPI may not have seen the latest data yet.
Quoted 4 lines
Quoted 2 lines
This commit fixes this by replacing
writel_relaxed()withwritel()when'This commit' is equally wrong as 'This patch'. See
Documentation/process/
Thanks very much. I will check the documents.
Best Regards,
Xu Lu
Quoted 10 lines
Quoted 4 lines
issuing IPI, which will use 'fence w, o' to ensure all previous writes made by current hart are visible to other harts before they receive the IPI.
I've fixed it up for you.
Thanks,
tglx
Quoted 8 lines · Xu Lu
RISC-V distinguishes between normal memory accesses and device I/O and uses FENCE instruction to order them as viewed by othe RISC-V harts and external devices or coprocessors. The FENCE instruction can order any combination of device input(I), device output(O), memory reads(R) and memory writes(W). For example, 'fence w, o' can be used to ensure all memory writes from instructions preceding the FENCE instruction appear earlier in the global memory order than device output writes from instructions after the FENCE instruction.
There is nothing risc-v specific in here really, it's just a bug in the driver: writel() means access the mmio register with appropriate barriers, while writel_relaxed() is a special case that should only ever be used if a particular function is sensitive to performance and never needs to be serialized.
Quoted 7 lines
drivers/irqchip/irq-thead-c900-aclint-sswi.c +0
-0
drivers/irqchip/irq-thead-c900-aclint-sswi.c static DEFINE_PER_CPU(void __iomem *, sswi_cpu_regs); · line 37
Quoted 8 lines
drivers/irqchip/irq-thead-c900-aclint-sswi.c +1
-1@@ -31,7 +31,7 @@ static DEFINE_PER_CPU(void __iomem *, sswi_cpu_regs);static void thead_aclint_sswi_ipi_send (unsigned int cpu) { -writel_relaxed (0x1 ,per_cpu (sswi_cpu_regs, cpu)); +writel (0x1 ,per_cpu (sswi_cpu_regs, cpu)); }static void thead_aclint_sswi_ipi_clear (void )
thead_aclint_sswi_ipi_clear() seems to have the same bug, it also uses the _relaxed() version for no apparent reason.
Arnd
Quoted 30 lines · Arnd Bergmann
Quoted 8 lines · Xu Lu
RISC-V distinguishes between normal memory accesses and device I/O and uses FENCE instruction to order them as viewed by othe RISC-V harts and external devices or coprocessors. The FENCE instruction can order any combination of device input(I), device output(O), memory reads(R) and memory writes(W). For example, 'fence w, o' can be used to ensure all memory writes from instructions preceding the FENCE instruction appear earlier in the global memory order than device output writes from instructions after the FENCE instruction.
There is nothing risc-v specific in here really, it's just a bug in the driver:
writel()means access the mmio register with appropriate barriers, whilewritel_relaxed()is a special case that should only ever be used if a particular function is sensitive to performance and never needs to be serialized.Quoted 13 lines
drivers/irqchip/irq-thead-c900-aclint-sswi.c +1
-1@@ -31,7 +31,7 @@ static DEFINE_PER_CPU(void __iomem *, sswi_cpu_regs);static void thead_aclint_sswi_ipi_send (unsigned int cpu) { -writel_relaxed (0x1 ,per_cpu (sswi_cpu_regs, cpu)); +writel (0x1 ,per_cpu (sswi_cpu_regs, cpu)); }
Quoted 5 lines
drivers/irqchip/irq-thead-c900-aclint-sswi.c static DEFINE_PER_CPU(void __iomem *, sswi_cpu_regs);
Quoted 2 lines
drivers/irqchip/irq-thead-c900-aclint-sswi.c +0
-1static void thead_aclint_sswi_ipi_clear (void ) --
thead_aclint_sswi_ipi_clear()seems to have the same bug, it also uses the_relaxed()version for no apparent reason.
Hi Arnd,
There seems no need to modify thead_aclint_sswi_ipi_clear() as it only clears pending IPI on current hart. No other harts require to see strict order between preceding memory writes and this ACLINT MMIO write. Please correct me if I missed anything.
Thanks,
Xu Lu
Quoted 2 lines
Arnd
Quoted 11 lines · Xu Lu
Quoted 4 lines · Arnd Bergmann
On Mon, Jan 27, 2025, at 10:38, Xu Lu wrote:
thead_aclint_sswi_ipi_clear()seems to have the same bug, it also uses the_relaxed()version for no apparent reason.There seems no need to modify
thead_aclint_sswi_ipi_clear()as it only clears pending IPI on current hart. No other harts require to see strict order between preceding memory writes and this ACLINT MMIO write. Please correct me if I missed anything.
My point was that you should always default to the normal operations, since the relaxed ones keep causing problems. If you still think it's important to use writel_relaxed() in thead_aclint_sswi_ipi_clear(), please add a comment explaining how you have shown that it's correct and why it matters, otherwise convert them both.
Arnd