Can't renew ec-384 certificate after enabling FIPS mode.

On my Rocky Linux 9.5, we can no longer renew our certificates from letsencrypt after we enabled FIPS mode:

```
[Fri Apr 11 20:57:15 UTC 2025] EC key
[Fri Apr 11 20:57:15 UTC 2025] Let's try ASN1 OID
[Fri Apr 11 20:57:15 UTC 2025] ECC oid:
[Fri Apr 11 20:57:15 UTC 2025] Error creating new order.
[Fri Apr 11 20:57:15 UTC 2025] pid
[Fri Apr 11 20:57:15 UTC 2025] No need to restore nginx config, skipping.
[Fri Apr 11 20:57:15 UTC 2025] _clearupdns
[Fri Apr 11 20:57:15 UTC 2025] dns_entries
[Fri Apr 11 20:57:15 UTC 2025] Skipping dns.
[Fri Apr 11 20:57:15 UTC 2025] _on_issue_err
[Fri Apr 11 20:57:15 UTC 2025] Please add '--debug' or '--log' to see more information.
[Fri Apr 11 20:57:15 UTC 2025] See: https://github.com/acmesh-official/acme.sh/wiki/How-to-debug-acme.sh

```

This doesn't seem to be an issue with openssl ... was able to successfully generate a new key and csr locally without it barfing. Running this version:

```
# acme.sh version
https://github.com/acmesh-official/acme.sh
v3.1.1
```

I have certs that are going to expire in a couple of days... so this is urgent for me personally...

Reproduce on Rocky Linux with FIPS enabled using the same acme.sh version and EC-384 renewal command, then rerun with --debug or --log to capture full output. Start by reading the main acme.sh script and follow the renewal flow to the new-order creation path where the failure appears. Compare behavior in normal and FIPS modes, and consider it done when an order is created and renewal succeeds without the "Error creating new order" message.

Written by the indexing model from the issue text.