ebpf/PSA: Checksum support for fields wider than 64 bits #3801
Conversation
This commit adds support for fields wider than 64 bits in implemented checksum algorithms.
|
@fruffy do you know why the test case |
| " bpf_trace_message(\"CRC%w%: data byte: %x\\n\", data[i-1]);\n" | ||
| " *reg ^= (u16) data[i-1];\n" | ||
| "void crc16_update(u16 * reg, const u8 * data, " | ||
| "u16 data_size, const u16 poly) {\n" |
There was a problem hiding this comment.
I am confused by this code, some documentation would help.
| if (addData) { | ||
| builder->appendFormat("%s = csum16_add(%s, %s)", stateVar.c_str(), | ||
| stateVar.c_str(), tmpVar.c_str()); | ||
| builder->appendFormat("csum16_add(%s, %s)", stateVar.c_str(), tmpVar.c_str()); |
There was a problem hiding this comment.
the result of csum16_add is ignored here?
There was a problem hiding this comment.
No, the result is not ignored. Assignment of the state variable is done 2 lines above, just before if statement. I moved this as a common expression for csum16_add and csum16_sub but I see it can be little confusing.
This is a flaky test. It can be fixed by rerunning. I do not know whether the fault is in bmv2's or testgen's checksum calculation yet. |
This PR adds support for fields wider than 64 bits in implemented checksum algorithms.
CC: @osinstom