kernel.git/arch/loongarch, branch linux-6.15.y

LoongArch: vDSO: Remove -nostdlib complier flag

2025-08-20T16:35:45Z

commit d35ec48fa6c8fe0cfa4a03155109fec7677911d4 upstream. Since $(LD) is directly used, hence -nostdlib is unneeded, MIPS has removed this, we should remove it too. bdbf2038fbf4 ("MIPS: VDSO: remove -nostdlib compiler flag"). In fact, other architectures also use $(LD) now. fe00e50b2db8 ("ARM: 8858/1: vdso: use $(LD) instead of $(CC) to link VDSO") 691efbedc60d ("arm64: vdso: use $(LD) instead of $(CC) to link VDSO") 2ff906994b6c ("MIPS: VDSO: Use $(LD) instead of $(CC) to link VDSO") 2b2a25845d53 ("s390/vdso: Use $(LD) instead of $(CC) to link vDSO") Cc: stable@vger.kernel.org Reviewed-by: Yanteng Si Signed-off-by: Wentao Guan Signed-off-by: Huacai Chen Signed-off-by: Greg Kroah-Hartman

LoongArch: Avoid in-place string operation on FDT content

2025-08-20T16:35:45Z

commit 70a2365e18affc5ebdaab1ca6a0b3c4f3aac2ee8 upstream. In init_cpu_fullname(), a constant pointer to "model" property is retrieved. It's later modified by the strsep() function, which is illegal and corrupts kernel's FDT copy. This is shown by dmesg, OF: fdt: not creating '/sys/firmware/fdt': CRC check failed Create a mutable copy of the model property and do in-place operations on the mutable copy instead. loongson_sysconf.cpuname lives across the kernel lifetime, thus manually releasing isn't necessary. Also move the of_node_put() call for the root node after the usage of its property, since of_node_put() decreases the reference counter thus usage after the call is unsafe. Cc: stable@vger.kernel.org Fixes: 44a01f1f726a ("LoongArch: Parsing CPU-related information from DTS") Reviewed-by: Jiaxun Yang Signed-off-by: Yao Zi Signed-off-by: Huacai Chen Signed-off-by: Greg Kroah-Hartman

LoongArch: Make relocate_new_kernel_size be a .quad value

2025-08-20T16:35:45Z

commit a1a81b5477196ca1290b367404a461e046e647d5 upstream. Now relocate_new_kernel_size is a .long value, which means 32bit, so its high 32bit is undefined. This causes memcpy((void *)reboot_code_buffer, relocate_new_kernel, relocate_new_kernel_size) in machine_kexec_prepare() access out of range memories in some cases, and then end up with an ADE exception. So make relocate_new_kernel_size be a .quad value, which means 64bit, to avoid such errors. Cc: stable@vger.kernel.org Signed-off-by: Huacai Chen Signed-off-by: Greg Kroah-Hartman

LoongArch: Don't use %pK through printk() in unwinder

2025-08-20T16:35:45Z

commit 2362e8124ed21445c6886806e5deaee717629ddd upstream. In the past %pK was preferable to %p as it would not leak raw pointer values into the kernel log. Since commit ad67b74d2469 ("printk: hash addresses printed with %p") the regular %p has been improved to avoid this issue. Furthermore, restricted pointers ("%pK") were never meant to be used through printk(). They can still unintentionally leak raw pointers or acquire sleeping locks in atomic contexts. Switch to the regular pointer formatting which is safer and easier to reason about. Cc: stable@vger.kernel.org Signed-off-by: Thomas Weißschuh Signed-off-by: Huacai Chen Signed-off-by: Greg Kroah-Hartman

LoongArch: BPF: Fix jump offset calculation in tailcall

2025-08-20T16:35:44Z

commit cd39d9e6b7e4c58fa77783e7aedf7ada51d02ea3 upstream. The extra pass of bpf_int_jit_compile() skips JIT context initialization which essentially skips offset calculation leaving out_offset = -1, so the jmp_offset in emit_bpf_tail_call is calculated by "#define jmp_offset (out_offset - (cur_offset))" is a negative number, which is wrong. The final generated assembly are as follow. 54: bgeu $a2, $t1, -8 # 0x0000004c 58: addi.d $a6, $s5, -1 5c: bltz $a6, -16 # 0x0000004c 60: alsl.d $t2, $a2, $a1, 0x3 64: ld.d $t2, $t2, 264 68: beq $t2, $zero, -28 # 0x0000004c Before apply this patch, the follow test case will reveal soft lock issues. cd tools/testing/selftests/bpf/ ./test_progs --allow=tailcalls/tailcall_bpf2bpf_1 dmesg: watchdog: BUG: soft lockup - CPU#2 stuck for 26s! [test_progs:25056] Cc: stable@vger.kernel.org Fixes: 5dc615520c4d ("LoongArch: Add BPF JIT support") Reviewed-by: Hengqi Chen Signed-off-by: Haoran Jiang Signed-off-by: Huacai Chen Signed-off-by: Greg Kroah-Hartman

LoongArch: KVM: Check interrupt route from physical CPU

2025-07-06T09:04:12Z

commit 45515c643d0abb75c2cc760a6bc6b235eadafd66 upstream. With EIOINTC interrupt controller, physical CPU ID is set for irq route. However the function kvm_get_vcpu() is used to get destination vCPU when delivering irq. With API kvm_get_vcpu(), the logical CPU ID is used. With API kvm_get_vcpu_by_cpuid(), vCPU ID can be searched from physical CPU ID. Cc: stable@vger.kernel.org Fixes: 3956a52bc05b ("LoongArch: KVM: Add EIOINTC read and write functions") Signed-off-by: Bibo Mao Signed-off-by: Huacai Chen Signed-off-by: Greg Kroah-Hartman

LoongArch: KVM: Fix interrupt route update with EIOINTC

2025-07-06T09:04:12Z

commit c34bbc2c990700ba07b271fc7c8113b0bc3e4093 upstream. With function eiointc_update_sw_coremap(), there is forced assignment like val = *(u64 *)pvalue. Parameter pvalue may be pointer to char type or others, there is problem with forced assignment with u64 type. Here the detailed value is passed rather address pointer. Cc: stable@vger.kernel.org Fixes: 3956a52bc05b ("LoongArch: KVM: Add EIOINTC read and write functions") Signed-off-by: Bibo Mao Signed-off-by: Huacai Chen Signed-off-by: Greg Kroah-Hartman

LoongArch: KVM: Add address alignment check for IOCSR emulation

2025-07-06T09:04:11Z

commit 9159c5e733cfa35ec863fa81960a3e7435f831fb upstream. IOCSR instruction supports 1/2/4/8 bytes access, the address should be naturally aligned with its access size. Here address alignment check is added in the EIOINTC kernel emulation. Cc: stable@vger.kernel.org Fixes: 3956a52bc05b ("LoongArch: KVM: Add EIOINTC read and write functions") Signed-off-by: Bibo Mao Signed-off-by: Huacai Chen Signed-off-by: Greg Kroah-Hartman

LoongArch: KVM: Disable updating of "num_cpu" and "feature"

2025-07-06T09:04:11Z

commit 955853cf83657faa58572ef3f08b44f0f88885c1 upstream. Property "num_cpu" and "feature" are read-only once eiointc is created, which are set with KVM_DEV_LOONGARCH_EXTIOI_GRP_CTRL attr group before device creation. Attr group KVM_DEV_LOONGARCH_EXTIOI_GRP_SW_STATUS is to update register and software state for migration and reset usage, property "num_cpu" and "feature" can not be update again if it is created already. Here discard write operation with property "num_cpu" and "feature" in attr group KVM_DEV_LOONGARCH_EXTIOI_GRP_CTRL. Cc: stable@vger.kernel.org Fixes: 1ad7efa552fd ("LoongArch: KVM: Add EIOINTC user mode read and write functions") Signed-off-by: Bibo Mao Signed-off-by: Huacai Chen Signed-off-by: Greg Kroah-Hartman

LoongArch: KVM: Check validity of "num_cpu" from user space

2025-07-06T09:04:11Z

commit cc8d5b209e09d3b52bca1ffe00045876842d96ae upstream. The maximum supported cpu number is EIOINTC_ROUTE_MAX_VCPUS about irqchip EIOINTC, here add validation about cpu number to avoid array pointer overflow. Cc: stable@vger.kernel.org Fixes: 1ad7efa552fd ("LoongArch: KVM: Add EIOINTC user mode read and write functions") Signed-off-by: Bibo Mao Signed-off-by: Huacai Chen Signed-off-by: Greg Kroah-Hartman