qemu-devel.nongnu.org archive mirror
 help / color / mirror / Atom feed
From: "Alex Bennée" <alex.bennee@linaro.org>
To: Mahmoud Mandour <ma.mandourr@gmail.com>
Cc: cota@braap.org, qemu-devel@nongnu.org
Subject: Re: [RFC PATCH v3 0/4] Cache TCG plugin & symbol-resolution API
Date: Tue, 22 Jun 2021 15:57:35 +0100	[thread overview]
Message-ID: <87pmwerlnb.fsf@linaro.org> (raw)
In-Reply-To: <20210608040532.56449-1-ma.mandourr@gmail.com>


Mahmoud Mandour <ma.mandourr@gmail.com> writes:

> This RFC series introduces a new cache TCG plugin that models separate
> L1 data cache and L1 instruction cache and uses one shared cache for
> all the cores.
>
> It also includes a commit by Alex that adds an API call that resolves
> the symbol of an insn.
>
> The original RFC patch posted by Alex Bennée included incorporating
> symbol resolution into the cache plugin that caused conflicts, so I
> dropped the plugin additions from that and introduced them afterwards.

Queued patches 1-2 to plugins/next, will queue the rest from next
revision. Could you please also add a section to
docs/devel/tcg-plugins.rst to detail the argument options.

>
> v2 -> v3:
>     Precomputed the value of block size shift once and stored in the
>     cache.
>
>     Removed tag shifting since it's okay to leave the tag in the
>     high-order bits and mask out set index and block offset.
>
>     Used one hashtable to store InsnData structs and made the structs
>     have separate counters for data misses and instruction misses.
>
>     Used a boolean to indicate whether an access resulted in a hit or a
>     miss.
>
>     Inserted an InsnData struct into the hashtable on translation-time
>     and made sure we do so once so that we don't rewrite the struct if
>     an instruction is translated multiple times.
>
>     Made the output format for most-missing instructions more
>     machine-readable.
>
>     Removed trace-generation.
>
>     Freed tokenized strings after argument parsing.
>
>     Returned null from cache_init() if argument cache config is bad.
>
>     Used one enum to indicate the chosen eviction policy.
>
>     Added function pointers for cache update and metadata initialization
>     and destroying. Those pointers are assigned to policy-specific
>     functions.
>
>     Remade LRU. Assigned a generation number that is incremented on each
>     set access to the currently-accessed block's priority. On miss, 
>     evicted the block with the least generation number.
>
>     Allowed to give multiple "evict" arguments and sticked to the last
>     one.
>
> Alex Bennée (1):
>   plugins/api: expose symbol lookup to plugins
>
> Mahmoud Mandour (3):
>   plugins: Added a new cache modelling plugin.
>   plugins/cache: Enabled cache parameterization
>   plugins/cache: Added FIFO and LRU eviction policies.
>
>  contrib/plugins/Makefile   |   1 +
>  contrib/plugins/cache.c    | 642 +++++++++++++++++++++++++++++++++++++
>  include/qemu/qemu-plugin.h |   9 +
>  plugins/api.c              |   6 +
>  4 files changed, 658 insertions(+)
>  create mode 100644 contrib/plugins/cache.c


-- 
Alex Bennée


      parent reply	other threads:[~2021-06-22 14:59 UTC|newest]

Thread overview: 8+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2021-06-08  4:05 [RFC PATCH v3 0/4] Cache TCG plugin & symbol-resolution API Mahmoud Mandour
2021-06-08  4:05 ` [RFC PATCH v3 1/4] plugins/api: expose symbol lookup to plugins Mahmoud Mandour
2021-06-08  4:05 ` [RFC PATCH v3 2/4] plugins: Added a new cache modelling plugin Mahmoud Mandour
2021-06-08  4:05 ` [RFC PATCH v3 3/4] plugins/cache: Enabled cache parameterization Mahmoud Mandour
2021-06-22 14:37   ` Alex Bennée
2021-06-08  4:05 ` [RFC PATCH v3 4/4] plugins/cache: Added FIFO and LRU eviction policies Mahmoud Mandour
2021-06-22 14:57   ` Alex Bennée
2021-06-22 14:57 ` Alex Bennée [this message]

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=87pmwerlnb.fsf@linaro.org \
    --to=alex.bennee@linaro.org \
    --cc=cota@braap.org \
    --cc=ma.mandourr@gmail.com \
    --cc=qemu-devel@nongnu.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).