[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <YF8K0k1bc3e381qS@kroah.com>
Date: Sat, 27 Mar 2021 11:37:06 +0100
From: Greg Kroah-Hartman <gregkh@...uxfoundation.org>
To: Alexander Lochmann <info@...xander-lochmann.de>
Cc: Dmitry Vyukov <dvyukov@...gle.com>,
Andrey Konovalov <andreyknvl@...gle.com>,
Jonathan Corbet <corbet@....net>,
Randy Dunlap <rdunlap@...radead.org>,
Andrew Klychkov <andrew.a.klychkov@...il.com>,
Miguel Ojeda <ojeda@...nel.org>,
Andrew Morton <akpm@...ux-foundation.org>,
Jakub Kicinski <kuba@...nel.org>,
Aleksandr Nogikh <nogikh@...gle.com>,
Wei Yongjun <weiyongjun1@...wei.com>,
Maciej Grochowski <maciej.grochowski@...me>,
kasan-dev@...glegroups.com, linux-doc@...r.kernel.org,
linux-kernel@...r.kernel.org
Subject: Re: [PATCHv3] Introduced new tracing mode KCOV_MODE_UNIQUE.
On Fri, Mar 26, 2021 at 09:51:28PM +0100, Alexander Lochmann wrote:
> It simply stores the executed PCs.
> The execution order is discarded.
> Each bit in the shared buffer represents every fourth
> byte of the text segment.
> Since a call instruction on every supported
> architecture is at least four bytes, it is safe
> to just store every fourth byte of the text segment.
> In contrast to KCOV_MODE_TRACE_PC, the shared buffer
> cannot overflow. Thus, all executed PCs are recorded.
Odd line-wrapping :(
Anyway, this describes _what_ this does, but I have no idea _why_ we
want this at all. What does this do that you can not do today? Why is
this needed? Who will use this? What tools have been modified to work
with it to prove it works properly?
thanks,
greg k-h
Powered by blists - more mailing lists