References: <20210810061644.20353-1-maximilian.fillinger@...2507...> <364dc95f-d198-397b-3720-3c9dac9890ae@...1227...> <78db12b9f4824871aecf26c93b0daf89@...2507...> From: Antonio Quartulli Message-ID: Date: Tue, 10 Aug 2021 14:55:59 +0200 MIME-Version: 1.0 In-Reply-To: <78db12b9f4824871aecf26c93b0daf89@...2507...> Content-Language: en-US Subject: Re: [Openvpn-devel] [PATCH] Replace deprecated mbedtls DRBG update function List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Content-Type: text/plain; charset=utf-8 Content-Transfer-Encoding: 7bit To: Maximilian Fillinger , Arne Schwabe , "openvpn-devel@lists.sourceforge.net" On 10/08/2021 14:51, Maximilian Fillinger wrote: >> Normally we have patch v2 here and also a patch v2 in the subject (use >> -v 2 when doing git format-patch) but for this small patch it is not a >> problem. > > I'll keep it in mind for next time! > >> Apart from the fact that we might want to abort (M_FATAL) if this fails >> instead basically ignoring the error and just log it, the change is >> fine. Considering the return status was ignored before, this patch is >> otherwise good. But failing also does not have any really bad impact... > > I was thinking about making it fatal, but there's another place where > personalizing random can fail that also only gives a warning. So that's what > I went with. (I wouldn't be opposed to making them both fatal, though.) > IMHO it's ok to keep WARN for now because, like pointed out, other parts of the code path just warn in case of issues. Just a few line above Max's change there is another failure treated with a WARN only. Therefore I'd stick to WARN unless a complete revamp is performed on the entire codepath. Regards, -- Antonio Quartulli