linux-kernel - Re: [PATCH v5] arch/x86: port I/O tracing on x86

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [day] [month] [year] [list]

Date:   Tue, 24 Oct 2023 11:32:41 +0200
From:   Greg KH <gregkh@...uxfoundation.org>
To:     Dan Raymond <raymod2@...il.com>
Cc:     linux-kernel@...r.kernel.org, x86@...nel.org,
        linux-serial <linux-serial@...r.kernel.org>, tglx@...utronix.de,
        mingo@...hat.com, bp@...en8.de, dave.hansen@...ux.intel.com,
        hpa@...or.com, peterz@...radead.org,
        andriy.shevchenko@...ux.intel.com, quic_saipraka@...cinc.com,
        Steven Rostedt <rostedt@...dmis.org>
Subject: Re: [PATCH v5] arch/x86: port I/O tracing on x86

On Mon, Oct 23, 2023 at 02:28:04PM -0600, Dan Raymond wrote:
> > Anyway, it's up to the x86 maintainers now, good luck!
> > 
> > But personally, I don't see the real need for this at all.  It's a
> > debugging thing for what exactly?  Who needs this?  Who will use it?
> > When will they use it?  And why?
> 
> This comment confuses me.  As you know I originally submitted a patch
> that added I/O tracing just to the 8250 serial driver.  The patch was
> titled "create debugfs interface for UART register tracing".  You said
> this at the time:
> 
>    "Anyway, again, cool feature, I like it, but if you can tie it into
>    the existing trace framework better (either by using that entirely
>    which might be best), or at the least, putting your hook into the
>    data path with it, that would be best."

Remember some of us, like myself, get on average 1000+ emails a day that
they need to file/delete/review, so what I wrote yesterday I usually
can't remember, let alone weeks ago :)

> My original patch went through a few revisions before Andy Shevchenko
> suggested I should add portio tracing instead in a manner similar to
> how CONFIG_TRACE_MMIO_ACCESS works.  You agreed.  Hence I created this
> patch.

That's great, but it turns out that the x86 maintainers don't like this,
so perhaps that's not going to work out well.

I still think the original idea of using tracepoints for serial data
would be best, but hey, I don't need this feature :)

thanks,

greg k-h