public inbox for linuxppc-dev@ozlabs.org 
 help / color / mirror / Atom feed
From: Rusty Russell <rusty@rustcorp•com.au>
To: David Laight <David.Laight@ACULAB•COM>,
	'Madhavan Srinivasan' <maddy@linux•vnet.ibm.com>,
	"mpe\@ellerman.id.au" <mpe@ellerman•id.au>
Cc: "paulus@samba•org" <paulus@samba•org>,
	"linuxppc-dev@lists•ozlabs.org" <linuxppc-dev@lists•ozlabs.org>,
	"anton@samba•org" <anton@samba•org>
Subject: RE: [RFC PATCH 0/2] powerpc: CR based local atomic operation implementation
Date: Thu, 18 Dec 2014 21:23:56 +1030	[thread overview]
Message-ID: <87egrxteij.fsf@rustcorp.com.au> (raw)
In-Reply-To: <063D6719AE5E284EB5DD2968C1650D6D1CA0E993@AcuExch.aculab.com>

David Laight <David.Laight@ACULAB•COM> writes:
> From: Rusty Russell
>> David Laight <David.Laight@ACULAB•COM> writes:
>> > From: Madhavan Srinivasan [mailto:maddy@linux•vnet.ibm.com]
>> > ...
>> >> >>> I also wonder if it is possible to inspect the interrupted
>> >> >>> code to determine the start/end of the RAS block.
>> >> >>> (Easiest if you assume that there is a single 'write' instruction
>> >> >>> as the last entry in the block.)
>> >> >>>
>> >> >> So each local_* function also have code in the __ex_table section. IIUC,
>> >> >> __ex_table contains two address. So if the return address found in the
>> >> >> first column of the _ex_table, use the corresponding address in the
>> >> >> second column to continue from.
>> >> >
>> >> > That really doesn't scale.
>> >> > I don't know how many 1000 address pairs you table will have (and the
>> >> > ones in each loadable module), but the search isn't going to be cheap.
>> >> >
>> >> > If these sequences are restartable then they can only have one write
>> >> > to memory.
> ...
>> >> 2) resulting code with lot of condition and branch (for opcode decode)
>> >> will be lot messy and may be an issue incase of maintenance,
>> >
>> > You don't need to decode the instructions.
>> > Just look for the two specific instructions used as markers.
>> > This is only really possible with fixed-size instructions.
>> >
>> > It might also be that the 'interrupt entry' path is easier to
>> > modify than the 'interrupt exit' one (fewer code paths) and
>> > you just need to modify the 'pc' in the stack frame.
>> > You are only interested in interrupts from kernel space.
>> 
>> It's an overoptimization for case that statistically never happens.
>> You won't even be able to measure the difference.
>> 
>> The question of bloat remains, but that's also easily measured.  In
>> practice, I'd guess less than 1k.
>
> IIRC they were 'static inline' so the table of addresses is generated
> for every use site.
> (copyin/out generates a similarly enormous table of addresses on amd64)

There are about 20 callers in the entire kernel.

Cheers,
Rusty.

      reply	other threads:[~2014-12-18 23:39 UTC|newest]

Thread overview: 29+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2014-11-27 12:18 [RFC PATCH 0/2] powerpc: CR based local atomic operation implementation Madhavan Srinivasan
2014-11-27 12:18 ` [RFC PATCH 1/2]powerpc: foundation code to handle CR5 for local_t Madhavan Srinivasan
2014-11-27 16:56   ` Segher Boessenkool
2014-11-28  1:58     ` Benjamin Herrenschmidt
2014-11-28  3:00       ` Madhavan Srinivasan
2014-11-28 15:57       ` Segher Boessenkool
2014-11-28  2:57     ` Madhavan Srinivasan
2014-11-28 16:00       ` Segher Boessenkool
2014-11-28  0:56   ` Benjamin Herrenschmidt
2014-11-28  3:15     ` Madhavan Srinivasan
2014-11-28  3:21       ` Benjamin Herrenschmidt
2014-11-28 10:53         ` David Laight
2014-11-30  9:01           ` Benjamin Herrenschmidt
2014-12-01 21:35   ` Gabriel Paubert
2014-12-03 14:59     ` Madhavan Srinivasan
2014-12-03 17:07       ` Gabriel Paubert
2014-12-02  2:04   ` Scott Wood
2014-12-03 14:49     ` Madhavan Srinivasan
2014-11-27 12:18 ` [RFC PATCH 2/2]powerpc: rewrite local_* to use CR5 flag Madhavan Srinivasan
2014-12-01 18:01   ` Gabriel Paubert
2014-12-03 15:05     ` Madhavan Srinivasan
2014-11-27 14:05 ` [RFC PATCH 0/2] powerpc: CR based local atomic operation implementation David Laight
2014-11-28  8:27   ` Madhavan Srinivasan
2014-11-28 10:09     ` David Laight
2014-12-01 15:35       ` Madhavan Srinivasan
2014-12-01 15:53         ` David Laight
2014-12-18  4:18           ` Rusty Russell
2014-12-18  9:52             ` David Laight
2014-12-18 10:53               ` Rusty Russell [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=87egrxteij.fsf@rustcorp.com.au \
    --to=rusty@rustcorp$(echo .)com.au \
    --cc=David.Laight@ACULAB$(echo .)COM \
    --cc=anton@samba$(echo .)org \
    --cc=linuxppc-dev@lists$(echo .)ozlabs.org \
    --cc=maddy@linux$(echo .)vnet.ibm.com \
    --cc=mpe@ellerman$(echo .)id.au \
    --cc=paulus@samba$(echo .)org \
    /path/to/YOUR_REPLY

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