fix http boot

[jsetje]

This is a simple typo that is currently breaking http boot. When the interior loop goes looking for a second Content-Length field, will always match the first one, and if that's some other field things break.

[jsetje]

@dennis-tseng99 can you please confirm that I did not misunderstand your intent?

[rosslagerwall]

This makes sense to me, though using nested loops seems unnecessarily complicated. Surely something like this would be better...

	/* Check the length of the file */
	buf_size_set = false;
	for (i = 0; i < rx_message.HeaderCount; i++) {
		if (!strcasecmp(rx_message.Headers[i].FieldName,
				(CHAR8 *)"Content-Length")) {
			UINT64 new_buf_size = ascii_to_int(rx_message.Headers[i].FieldValue);
			if (buf_size_set && new_buf_size != *buf_size) {
				perror(L"Content-Length is invalid\n");
				goto error;
			}
			*buf_size = new_buf_size;
			buf_size_set = true;
		}
	}

[jsetje]

I suspect you mean unnecessarily computationally complex, and yes doing one loop for each didn't strike me as ideal either, although I'm not sure there are enough items here to really make a difference.

@vathpela let me know what you prefer, I'd probably drop buf_size_set from the above code and just test new_buf_size for not being 0.

[vathpela]

I think @rosslagerwall's way is fine.

[jsetje]

@rosslagerwall thanks for pushing to clean that up as well!

buf_size_set is kind of required since someone could spoof a 0 content length, and if they do, we probably want to catch that.

[marta-lewandowska]

Works for me on x86 and aarch64.