Re: [PATCH] SCSI RDMA PROTOCOL (SRP) TARGET: Replace HTTP links with HTTPS ones

[Alexander A. Klimov]

Am 12.07.20 um 21:52 schrieb Bart Van Assche:
> On 2020-07-10 11:12, Alexander A. Klimov wrote:
> > Am 10.07.20 um 16:22 schrieb Bart Van Assche:
> > > On 2020-07-09 12:48, Alexander A. Klimov wrote:
> > > > diff --git a/drivers/infiniband/ulp/srpt/Kconfig b/drivers/infiniband/ulp/srpt/Kconfig
> > > > index 4b5d9b792cfa..f63b34d9ae32 100644
> > > > --- a/drivers/infiniband/ulp/srpt/Kconfig
> > > > +++ b/drivers/infiniband/ulp/srpt/Kconfig
> > > > @@ -10,4 +10,4 @@ config INFINIBAND_SRPT
> > > >  ÂÂÂÂÂÂÂ that supports the RDMA protocol. Currently the RDMA protocol is
> > > >  ÂÂÂÂÂÂÂ supported by InfiniBand and by iWarp network hardware. More
> > > >  ÂÂÂÂÂÂÂ information about the SRP protocol can be found on the website
> > > > -ÂÂÂÂÂ of the INCITS T10 technical committee (http://www.t10.org/).
> > > > +ÂÂÂÂÂ of the INCITS T10 technical committee (https://www.t10.org/).
> > >
> > > It is not clear to me how modifying an URL in a Kconfig file helps to
> > > reduce the attack surface on kernel devs?
> >
> > Not on all, just on the ones who open it.
>
> Is changing every single HTTP URL in the kernel into a HTTPS URL the best
> solution? Is this the only solution? Has it been considered to recommend
> kernel developers who are concerned about MITM attacks to install a browser
> extension like HTTPS Everywhere instead?
I've installed that addon myself.
But IMAO it's just a workaround which is (not available to all browsers, 
not installed by default in any of them and) not even 100% secure unless 
you tick a particular checkbox.

Anyway the majority of maintainers and Torvalds himself agree with my 
solution.

I mean, just look at
git log '--author=Alexander A. Klimov <grandmaster@xxxxxxxxxxxx>' \
--oneline v5.7..master

Or (better) wait for v5.9-rc1 (and all the yet just applied patches it 
will consist of) *and then* run the command.

> Thanks,
>
> Bart.