Re: [PATCH] ubifs-utils: Don't compile engine support if not available

From: Sebastian Andrzej Siewior <sebastian@breakpoint.cc>
To: Zhihao Cheng <chengzhihao1@huawei.com>
Cc: linux-mtd@lists.infradead.org,
	David Oberhollenzer <david.oberhollenzer@sigma-star.at>
Subject: Re: [PATCH] ubifs-utils: Don't compile engine support if not available
Date: Mon, 31 Aug 2026 08:25:38 +0200	[thread overview]
Message-ID: <20260831062538.0E_FwcMd@breakpoint.cc> (raw)
In-Reply-To: <481dc2f7-ba6f-743a-b841-a3bdb8807ff6@huawei.com>

On 2026-08-31 09:50:18 [+0800], Zhihao Cheng wrote:
> 在 2026/8/30 1:32, Sebastian Andrzej Siewior 写道:
> > @@ -139,6 +142,7 @@ static EVP_PKEY *read_private_key(const char *private_key_name, X509 **cert)
> >   	*cert = NULL;
> >   	if (!strncmp(private_key_name, "pkcs11:", 7)) {
> > +#ifndef OPENSSL_NO_ENGINE
> >   		ENGINE *e;
> >   		struct {
> >   			const char *url;
> > @@ -177,6 +181,10 @@ static EVP_PKEY *read_private_key(const char *private_key_name, X509 **cert)
> >   		}
> >   		*cert = parms.cert;
> >   		fprintf(stderr, "Using cert %p\n", *cert);
> > +#else
> > +		ssl_err_msg("PKCS#11 ENGINE support not available.");
> > +		return NULL;
> Hi Sebastian,
> Thanks for pointing that. I am wondering whether it would be better to
> implement PKCS11 encryption in #else branch by the OpenSSL 4.0 API provider.

Well. This is a the very least gets it compiled. The provider API is
supported since OpenSSL 3.0 so I wouldn't make it available only if
ENGINE isn't supported. You still need the actual provider.

> > +#endif
> >   	} else {
> >   		BIO *b;
> > 

Sebastian

______________________________________________________
Linux MTD discussion mailing list
http://lists.infradead.org/mailman/listinfo/linux-mtd/