On Thu, 14 Mar 2024 at 16:00, Puranjay Mohan <puranjay12@xxxxxxxxx> wrote: > > Add support for [LDX | STX | ST], PROBE_MEM32, [B | H | W | DW] > instructions. They are similar to PROBE_MEM instructions with the > following differences: > - PROBE_MEM32 supports store. > - PROBE_MEM32 relies on the verifier to clear upper 32-bit of the > src/dst register > - PROBE_MEM32 adds 64-bit kern_vm_start address (which is stored in R28 > in the prologue). Due to bpf_arena constructions such R28 + reg + > off16 access is guaranteed to be within arena virtual range, so no > address check at run-time. > - PROBE_MEM32 allows STX and ST. If they fault the store is a nop. When > LDX faults the destination register is zeroed. > > To support these on arm64, we do tmp2 = R28 + src/dst reg and then use > tmp2 as the new src/dst register. This allows us to reuse most of the > code for normal [LDX | STX | ST]. > > Signed-off-by: Puranjay Mohan <puranjay12@xxxxxxxxx> > --- Hi Alexei, Puranjay and I were discussing this stuff off list and noticed that atomic instructions are not handled. It turns out that will cause a kernel crash right now because the 32-bit offset into arena will be dereferenced directly. e.g. something like this: @@ -55,6 +56,7 @@ int arena_list_add(void *ctx) test_val++; n->value = i; arena_sum += i; + __sync_fetch_and_add(&arena_sum, 0); list_add_head(&n->node, list_head); } #else I will try to prepare a fix for the x86 JIT. Puranjay will do the same for his set.