From: Sergey Matyukevich Date: Thu, 27 Jun 2024 15:04:09 +0300 Subject: [PATCH v2 1/1] lib: sbi: check incoming dbtr shmem address Message-ID: <20240627121228.67874-1-geomatsi@gmail.com> List-Id: To: opensbi@lists.infradead.org MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit Current Debug Trigger SBI extension proposal suggests to activate shmem area and obtain its physical address from S-mode software in the following way: : If both `shmem_phys_lo` and `shmem_phys_hi` parameters are not : all-ones bitwise then `shmem_phys_lo` specifies the lower XLEN : bits and `shmem_phys_hi` specifies the upper XLEN bits of the : shared memory physical base address. The `shmem_phys_lo` MUST : be `(XLEN / 8)` byte aligned and the size of shared memory is : assumed to be `trig_max * (XLEN / 2)` bytes. For more details see the current version of the proposal: - https://lists.riscv.org/g/tech-debug/message/1302 On the other hand, on RV32, the M-mode can only access the first 4GB of the physical address space because M-mode does not have MMU to access full 34-bit physical address space. Similarly, on RV64, the M-mode can only access memory addressed by 64 bits. This commit checks shmem address in function sbi_dbtr_setup_shmem to make sure that shmem_phys_hi part of the valid address is zero. Besides, the macro DBTR_SHMEM_MAKE_PHYS is updated to take into account only low XLEN part. Signed-off-by: Sergey Matyukevich --- Tested RV32 and RV64 using QEMU and patches from Himanshu Chauhan kernel tree https://github.com/hschauhan/riscv-linux/blob/linux-6.8-rc5-sdtrig with the following patch on top of it: : diff --git a/arch/riscv/kernel/hw_breakpoint.c b/arch/riscv/kernel/hw_breakpoint.c : index fe8215868530..adab63a80336 100644 : --- a/arch/riscv/kernel/hw_breakpoint.c : +++ b/arch/riscv/kernel/hw_breakpoint.c : @@ -27,16 +27,6 @@ static int dbtr_total_num __ro_after_init; : static int dbtr_type __ro_after_init; : static int dbtr_init __ro_after_init; : : -#if __riscv_xlen == 64 : -#define MEM_HI(_m) ((u64)_m >> 32) : -#define MEM_LO(_m) ((u64)_m & 0xFFFFFFFFUL) : -#elif __riscv_xlen == 32 : -#define MEM_HI(_m) (((u64)_m >> 32) & 0x3) : -#define MEM_LO(_m) ((u64)_m & 0xFFFFFFFFUL) : -#else : -#error "Unknown __riscv_xlen" : -#endif : - : static int arch_smp_setup_sbi_shmem(unsigned int cpu) : { : struct sbi_dbtr_shmem_entry *dbtr_shmem; : @@ -52,10 +42,14 @@ static int arch_smp_setup_sbi_shmem(unsigned int cpu) : : shmem_pa = __pa(dbtr_shmem); : : - ret = sbi_ecall(SBI_EXT_DBTR, SBI_EXT_DBTR_SETUP_SHMEM, : - (!MEM_LO(shmem_pa) ? 0xFFFFFFFFUL : MEM_LO(shmem_pa)), : - (!MEM_HI(shmem_pa) ? 0xFFFFFFFFUL : MEM_HI(shmem_pa)), : - 0, 0, 0, 0); : + if (IS_ENABLED(CONFIG_32BIT)) : + ret = sbi_ecall(SBI_EXT_DBTR, SBI_EXT_DBTR_SETUP_SHMEM, : + lower_32_bits(shmem_pa), : + upper_32_bits(shmem_pa), : + 0, 0, 0, 0); : + else : + ret = sbi_ecall(SBI_EXT_DBTR, SBI_EXT_DBTR_SETUP_SHMEM, : + shmem_pa, 0, 0, 0, 0, 0); : : if (ret.error) { : switch(ret.error) { Changes RFC (v1) -> v2 - align with DBTR SBI extension proposal lib/sbi/sbi_dbtr.c | 20 ++++++++++++++------ 1 file changed, 14 insertions(+), 6 deletions(-) diff --git a/lib/sbi/sbi_dbtr.c b/lib/sbi/sbi_dbtr.c index 9c92af6..6e2083e 100644 --- a/lib/sbi/sbi_dbtr.c +++ b/lib/sbi/sbi_dbtr.c @@ -47,13 +47,7 @@ static unsigned long hart_state_ptr_offset; _idx < _max; \ _idx++, _entry = ((_etype *)_base + _idx)) -#if __riscv_xlen == 64 #define DBTR_SHMEM_MAKE_PHYS(_p_hi, _p_lo) (_p_lo) -#elif __riscv_xlen == 32 -#define DBTR_SHMEM_MAKE_PHYS(_p_hi, _p_lo) (((u64)(_p_hi) << 32) | (_p_lo)) -#else -#error "Undefined XLEN" -#endif /* must call with hs != NULL */ static inline bool sbi_dbtr_shmem_disabled( @@ -277,6 +271,20 @@ int sbi_dbtr_setup_shmem(const struct sbi_domain *dom, unsigned long smode, if (shmem_phys_lo & SBI_DBTR_SHMEM_ALIGN_MASK) return SBI_ERR_INVALID_PARAM; + /* + * On RV32, the M-mode can only access the first 4GB of + * the physical address space because M-mode does not have + * MMU to access full 34-bit physical address space. + * So fail if the upper 32 bits of the physical address + * is non-zero on RV32. + * + * On RV64, kernel sets upper 64bit address part to zero. + * So fail if the upper 64bit of the physical address + * is non-zero on RV64. + */ + if (shmem_phys_hi) + return SBI_EINVALID_ADDR; + if (dom && !sbi_domain_check_addr(dom, DBTR_SHMEM_MAKE_PHYS(shmem_phys_hi, shmem_phys_lo), smode, SBI_DOMAIN_READ | SBI_DOMAIN_WRITE)) -- 2.45.2