[Openvpn-devel] using openssl feature wherever possible

2021-03-09 Thread Илья Шипицин
Hello, if nobody minds, I can send several patches that eliminates comparison of OPENSSL_VERSION, for example diff --git a/src/openvpn/crypto_openssl.c b/src/openvpn/crypto_openssl.c index 49698e4b..316cca6f 100644 --- a/src/openvpn/crypto_openssl.c +++ b/src/openvpn/crypto_openssl.c @@ -51,7

Re: [Openvpn-devel] using openssl feature wherever possible

2021-03-09 Thread Gert Doering
Hi, On Tue, Mar 09, 2021 at 04:54:13PM +0500, ?? wrote: > if nobody minds, I can send several patches that eliminates comparison of > OPENSSL_VERSION, for example We do mind. They are coded this way on purpose - so when we drop support for OpenSSL before 1.1.0, it is clear

Re: [Openvpn-devel] using openssl feature wherever possible

2021-03-09 Thread Илья Шипицин
вт, 9 мар. 2021 г. в 17:47, Gert Doering : > Hi, > > On Tue, Mar 09, 2021 at 04:54:13PM +0500, ?? wrote: > > if nobody minds, I can send several patches that eliminates comparison of > > OPENSSL_VERSION, for example > > We do mind. They are coded this way on purpose - so

Re: [Openvpn-devel] using openssl feature wherever possible

2021-03-09 Thread Gert Doering
Hi, On Tue, Mar 09, 2021 at 05:52:12PM +0500, ?? wrote: > > On Tue, Mar 09, 2021 at 04:54:13PM +0500, ?? wrote: > > > if nobody minds, I can send several patches that eliminates comparison of > > > OPENSSL_VERSION, for example > > > > We do mind. They

Re: [Openvpn-devel] [PATCH] Avoid a crash in mbed TLS 2.25 with --verb < 8

2021-03-09 Thread Antonio Quartulli
Hi, On 08/03/2021 15:21, Arne Schwabe wrote: [cut] > diff --git a/src/openvpn/options.c b/src/openvpn/options.c > index 0eb049d8..6d908e15 100644 > --- a/src/openvpn/options.c > +++ b/src/openvpn/options.c > @@ -5883,6 +5883,12 @@ add_option(struct options *options, > { >

Re: [Openvpn-devel] using openssl feature wherever possible

2021-03-09 Thread Илья Шипицин
we may keep combo. both #ifdef EVP_PKEY_TLS1_PRF and comment related to supported openssl versions (to drop support if we decide) вт, 9 мар. 2021 г. в 17:56, Gert Doering : > Hi, > > On Tue, Mar 09, 2021 at 05:52:12PM +0500, ?? wrote: > > > On Tue, Mar 09, 2021 at 04:54:13PM

Re: [Openvpn-devel] using openssl feature wherever possible

2021-03-09 Thread Gert Doering
Hi, On Tue, Mar 09, 2021 at 06:26:08PM +0500, ?? wrote: > we may keep combo. > both #ifdef EVP_PKEY_TLS1_PRF and comment related to supported openssl > versions (to drop support if we decide) We could, but we won't. (I can see the benefits, but I'm not the one maintaining

Re: [Openvpn-devel] using openssl feature wherever possible

2021-03-09 Thread Arne Schwabe
Am 09.03.21 um 12:54 schrieb Илья Шипицин: > Hello, > > if nobody minds, I can send several patches that eliminates comparison > of OPENSSL_VERSION, for example > > > diff --git a/src/openvpn/crypto_openssl.c b/src/openvpn/crypto_openssl.c > index 49698e4b..316cca6f 100644 > ---

Re: [Openvpn-devel] using openssl feature wherever possible

2021-03-09 Thread David Sommerseth
On 09/03/2021 14:28, Gert Doering wrote: Hi, On Tue, Mar 09, 2021 at 06:26:08PM +0500, ?? wrote: we may keep combo. both #ifdef EVP_PKEY_TLS1_PRF and comment related to supported openssl versions (to drop support if we decide) We could, but we won't. (I can see the

Re: [Openvpn-devel] [PATCH] Require at least 20MB of mlock()-able memory if --mlock is used.

2021-03-09 Thread David Sommerseth
On 08/03/2021 14:45, tincanteksup wrote: On 08/03/2021 08:06, Arne Schwabe wrote: Looking at this feature  from today's perspective, it feels like one of OpenVPN's boutique features. Was probably useful at some point but doesn't really make much sense today anymore. Esepcially with what is

[Openvpn-devel] [PATCH applied] Re: Implement server side of AUTH_PENDING with extending timeout

2021-03-09 Thread Gert Doering
I won't claim to understand what this stuff is doing in the grand scheme of things (as far as I can see, all it does is "it introduces a timeout variable all over the place, which then gets sent to the client") but the changes look safe enough :-) I have subjected this to client-side and server

Re: [Openvpn-devel] [PATCH] Require at least 20MB of mlock()-able memory if --mlock is used.

2021-03-09 Thread tincanteksup
On 09/03/2021 18:52, David Sommerseth wrote: On 08/03/2021 14:45, tincanteksup wrote: On 08/03/2021 08:06, Arne Schwabe wrote: Looking at this feature  from today's perspective, it feels like one of OpenVPN's boutique features. Was probably useful at some point but doesn't really make

Re: [Openvpn-devel] [PATCH] Require at least 20MB of mlock()-able memory if --mlock is used.

2021-03-09 Thread David Sommerseth
On 07/03/2021 19:44, Gert Doering wrote: Hi, On Sun, Mar 07, 2021 at 01:36:03PM -0500, Selva Nair wrote: "I'm not sure", TBH. rlimit handling in unix is a bit of an unknown territory for me. What I understand is that root can *increment* the rlimit at will, but I'd assume that the rlimit

Re: [Openvpn-devel] [PATCH] Require at least 20MB of mlock()-able memory if --mlock is used.

2021-03-09 Thread tincanteksup
On 09/03/2021 20:53, David Sommerseth wrote: On 09/03/2021 21:04, tincanteksup wrote: > I have swapping issues all the time and I can't add more RAM. I don't want system wide disk encryption. And I don't want an SSD either. I do not have the money to keep up with modern hardware. Having

Re: [Openvpn-devel] [PATCH] Require at least 20MB of mlock()-able memory if --mlock is used.

2021-03-09 Thread Gert Doering
Hi, On Mon, Mar 08, 2021 at 09:06:10AM +0100, Arne Schwabe wrote: > That being said I am against your patch, I am just wondering if that is > a feature we need to keep at all. Just for clarification: is there a "not" missing in that sentence? (Code-wise, mlock / *rlimit is just ~40-50 lines of

Re: [Openvpn-devel] [PATCH v2] Require at least 100MB of mlock()-able memory if --mlock is used.

2021-03-09 Thread Selva Nair
Hi, On Mon, Mar 8, 2021 at 2:11 AM Gert Doering wrote: > If --mlock is used, the amount of memory OpenVPN can use is guarded > by the RLIMIT_MEMLOCK value (see mlockall(2)). The OS default for this > is usually 64 Kbyte, which is enough for OpenVPN to initialize, but > as soon as the first TLS

Re: [Openvpn-devel] [PATCH] Require at least 20MB of mlock()-able memory if --mlock is used.

2021-03-09 Thread David Sommerseth
On 09/03/2021 21:04, tincanteksup wrote: > I have swapping issues all the time and I can't add more RAM. I don't want system wide disk encryption. And I don't want an SSD either. I do not have the money to keep up with modern hardware. Having openvpn --mlock is exactly the right choice for my

Re: [Openvpn-devel] [PATCH] Avoid a crash in mbed TLS 2.25 with --verb < 8

2021-03-09 Thread Arne Schwabe
Am 09.03.21 um 10:09 schrieb Antonio Quartulli: > Hi, > > On 08/03/2021 15:21, Arne Schwabe wrote: > [cut] > >> diff --git a/src/openvpn/options.c b/src/openvpn/options.c >> index 0eb049d8..6d908e15 100644 >> --- a/src/openvpn/options.c >> +++ b/src/openvpn/options.c >> @@ -5883,6 +5883,12 @@