Re: MSI interrupts

Liste des GroupesRevenir à c arch 
Sujet : Re: MSI interrupts
De : cross (at) *nospam* spitfire.i.gajendra.net (Dan Cross)
Groupes : comp.arch
Date : 25. Mar 2025, 18:54:42
Autres entêtes
Organisation : PANIX Public Access Internet and UNIX, NYC
Message-ID : <vruql2$cv2$3@reader1.panix.com>
References : 1 2 3 4
User-Agent : trn 4.0-test77 (Sep 1, 2010)
In article <0343529d63f68c76b5e0d227ef6e39dd@www.novabbs.org>,
MitchAlsup1 <mitchalsup@aol.com> wrote:
On Tue, 25 Mar 2025 15:10:19 +0000, Dan Cross wrote:
In article <b201b9212700f9359a831e03b1d60cef@www.novabbs.org>,
MitchAlsup1 <mitchalsup@aol.com> wrote:
On Tue, 25 Mar 2025 13:45:39 +0000, Scott Lurndal wrote:
mitchalsup@aol.com (MitchAlsup1) writes:
Nothing gets lost.
>
This isn't a hardware transaction.   If the thread has updated data
and you take away the lock, the system is now inconsistent
and nasal daemons will ensue.
>
In HW there is a little thing known as branch misprediction repair,
where all the work done is made to appear as if those instructions
were never executed to begin with. The size of the branch mispredict
window is on the order of 300 instructions now.
>
With Spectré and Meltdown even cache updates must be delayed until
after the causing instruction retires.
>
But again, that's irrelevant.  The "critical sections" in
question here are software abstractions created by system
software.
>
Say you have a critical section (of your description) that was
preceded by an if-statement that was predicted to enter CS, but
1,000 cycles later the branch is resolved to not enter the CS.
{{Branch was dependent on an MMI/O location down the PCIe tree
which is why it took so long to resolve.}}
>
HW Backs the core up so that the RF and all memory locations
contain their original values at the point of the mispredicted
branch.
>
Can thread running on core even tell if you entered CS or not !
>
No, your thread cannot.

Again, how is that relevant?  The branch predictor can only
handle speculative execution of so-many instructions.  Attempts
to perform more instructions than that will stall the pipeline.

A critical section as defined by software may contain more than
that number of instructions.  A critical section, as defined by
software, may touch more cache lines than your atomic event
proposal can handle.  Thus, for your architecture to be useful,
you must provide some mechanism for handling those
eventualities. 

Your stuff sounds fine for getting into the critsec; but it
doesn't help you once you're in it, and it seems like it may
hurt you if some higher-priority thing comes along and yanks
your lock out from under you.

- Dan C.


Date Sujet#  Auteur
13 Mar 25 * MSI interrupts163Robert Finch
13 Mar 25 `* Re: MSI interrupts162MitchAlsup1
13 Mar 25  +* Re: MSI interrupts5Robert Finch
13 Mar 25  i+- Re: MSI interrupts1MitchAlsup1
13 Mar 25  i`* Re: MSI interrupts3Robert Finch
13 Mar 25  i +- Re: MSI interrupts1MitchAlsup1
13 Mar 25  i `- Re: MSI interrupts1Stefan Monnier
13 Mar 25  `* Re: MSI interrupts156MitchAlsup1
13 Mar 25   `* Re: MSI interrupts155MitchAlsup1
14 Mar 25    `* Re: MSI interrupts154MitchAlsup1
14 Mar 25     `* Re: MSI interrupts153MitchAlsup1
14 Mar 25      `* Re: MSI interrupts152MitchAlsup1
15 Mar 25       `* Re: MSI interrupts151Robert Finch
15 Mar 25        `* Re: MSI interrupts150MitchAlsup1
15 Mar 25         `* Re: MSI interrupts149Robert Finch
15 Mar 25          `* Re: MSI interrupts148MitchAlsup1
16 Mar 25           `* Re: MSI interrupts147Robert Finch
16 Mar 25            +- Re: MSI interrupts1MitchAlsup1
17 Mar 25            +* Re: MSI interrupts142Michael S
17 Mar 25            i+- Re: MSI interrupts1Robert Finch
17 Mar 25            i+* Re: MSI interrupts133Robert Finch
18 Mar 25            ii+* Re: MSI interrupts127Robert Finch
19 Mar 25            iii+* Re: MSI interrupts124MitchAlsup1
19 Mar 25            iiii+* Re: MSI interrupts121Dan Cross
19 Mar 25            iiiii+* Re: MSI interrupts112MitchAlsup1
20 Mar 25            iiiiii`* Re: MSI interrupts111Dan Cross
20 Mar 25            iiiiii `* Re: MSI interrupts110MitchAlsup1
20 Mar 25            iiiiii  `* Re: MSI interrupts109Dan Cross
20 Mar 25            iiiiii   +* Re: MSI interrupts31MitchAlsup1
24 Mar 25            iiiiii   i`* Re: MSI interrupts30Dan Cross
24 Mar 25            iiiiii   i +* Re: MSI interrupts20MitchAlsup1
24 Mar 25            iiiiii   i i+* Re: MSI interrupts18Stefan Monnier
24 Mar 25            iiiiii   i ii`* Re: MSI interrupts17MitchAlsup1
24 Mar 25            iiiiii   i ii `* Re: MSI interrupts16Dan Cross
24 Mar 25            iiiiii   i ii  +* Re: MSI interrupts8MitchAlsup1
25 Mar 25            iiiiii   i ii  i`* Re: MSI interrupts7Dan Cross
25 Mar 25            iiiiii   i ii  i `* Re: MSI interrupts6Dan Cross
25 Mar 25            iiiiii   i ii  i  `* Re: MSI interrupts5Stefan Monnier
25 Mar 25            iiiiii   i ii  i   +- Re: MSI interrupts1Stefan Monnier
25 Mar 25            iiiiii   i ii  i   +- Re: MSI interrupts1Dan Cross
27 Mar 25            iiiiii   i ii  i   `* Re: MSI interrupts2Terje Mathisen
27 Mar 25            iiiiii   i ii  i    `- Re: MSI interrupts1MitchAlsup1
25 Mar 25            iiiiii   i ii  `* Re: MSI interrupts7Chris M. Thomasson
25 Mar 25            iiiiii   i ii   `* Re: MSI interrupts6Dan Cross
25 Mar 25            iiiiii   i ii    `* Re: MSI interrupts5Chris M. Thomasson
25 Mar 25            iiiiii   i ii     `* Re: MSI interrupts4Dan Cross
26 Mar 25            iiiiii   i ii      `* Re: MSI interrupts3Chris M. Thomasson
26 Mar 25            iiiiii   i ii       `* Re: MSI interrupts2Dan Cross
26 Mar 25            iiiiii   i ii        `- Re: MSI interrupts1Chris M. Thomasson
24 Mar 25            iiiiii   i i`- Re: MSI interrupts1Dan Cross
24 Mar 25            iiiiii   i +- Re: MSI interrupts1MitchAlsup1
24 Mar 25            iiiiii   i `* Re: MSI interrupts8Dan Cross
24 Mar 25            iiiiii   i  `* Re: MSI interrupts7Chris M. Thomasson
24 Mar 25            iiiiii   i   `* Re: MSI interrupts6Dan Cross
25 Mar 25            iiiiii   i    `* Re: MSI interrupts5Chris M. Thomasson
25 Mar 25            iiiiii   i     +* Re: MSI interrupts2Chris M. Thomasson
25 Mar 25            iiiiii   i     i`- Re: MSI interrupts1Dan Cross
25 Mar 25            iiiiii   i     `* Re: MSI interrupts2Dan Cross
25 Mar 25            iiiiii   i      `- Re: MSI interrupts1Chris M. Thomasson
24 Mar 25            iiiiii   +- Re: MSI interrupts1Chris M. Thomasson
24 Mar 25            iiiiii   `* Re: MSI interrupts76Dan Cross
24 Mar 25            iiiiii    +* Re: MSI interrupts57MitchAlsup1
25 Mar 25            iiiiii    i`* Re: MSI interrupts56Dan Cross
25 Mar 25            iiiiii    i `* Re: MSI interrupts55MitchAlsup1
25 Mar 25            iiiiii    i  +* Re: MSI interrupts2Stefan Monnier
25 Mar 25            iiiiii    i  i`- Re: MSI interrupts1Chris M. Thomasson
25 Mar 25            iiiiii    i  +- Re: MSI interrupts1Dan Cross
25 Mar 25            iiiiii    i  `* Re: MSI interrupts51MitchAlsup1
25 Mar 25            iiiiii    i   `* Re: MSI interrupts50Dan Cross
25 Mar 25            iiiiii    i    `* Re: MSI interrupts49MitchAlsup1
25 Mar 25            iiiiii    i     `* Re: MSI interrupts48Dan Cross
25 Mar 25            iiiiii    i      `* Re: MSI interrupts47MitchAlsup1
25 Mar 25            iiiiii    i       `* Re: MSI interrupts46Dan Cross
25 Mar 25            iiiiii    i        +* Re: MSI interrupts8Stefan Monnier
26 Mar 25            iiiiii    i        i+* Re: MSI interrupts5Dan Cross
26 Mar 25            iiiiii    i        ii`* Re: MSI interrupts4Stefan Monnier
26 Mar 25            iiiiii    i        ii +- Re: MSI interrupts1Dan Cross
26 Mar 25            iiiiii    i        ii `* Re: MSI interrupts2MitchAlsup1
27 Mar 25            iiiiii    i        ii  `- Re: MSI interrupts1Stefan Monnier
26 Mar 25            iiiiii    i        i+- Re: MSI interrupts1Chris M. Thomasson
26 Mar 25            iiiiii    i        i`- Re: MSI interrupts1MitchAlsup1
26 Mar 25            iiiiii    i        `* Re: MSI interrupts37MitchAlsup1
26 Mar 25            iiiiii    i         `* Re: MSI interrupts36Dan Cross
26 Mar 25            iiiiii    i          +* Re: MSI interrupts4Stefan Monnier
26 Mar 25            iiiiii    i          i`* Re: MSI interrupts3Dan Cross
26 Mar 25            iiiiii    i          i `* Re: MSI interrupts2Chris M. Thomasson
4 Apr 25            iiiiii    i          i  `- Re: MSI interrupts1Chris M. Thomasson
26 Mar 25            iiiiii    i          `* Re: MSI interrupts31MitchAlsup1
26 Mar 25            iiiiii    i           +- Re: MSI interrupts1Stefan Monnier
26 Mar 25            iiiiii    i           +* Re: MSI interrupts2Stefan Monnier
27 Mar 25            iiiiii    i           i`- Re: MSI interrupts1MitchAlsup1
27 Mar 25            iiiiii    i           +* Re: MSI interrupts3MitchAlsup1
27 Mar 25            iiiiii    i           i`* Re: MSI interrupts2MitchAlsup1
27 Mar 25            iiiiii    i           i `- Re: MSI interrupts1Dan Cross
27 Mar 25            iiiiii    i           `* Re: MSI interrupts24Dan Cross
27 Mar 25            iiiiii    i            +* Re: MSI interrupts2Stefan Monnier
27 Mar 25            iiiiii    i            i`- Re: MSI interrupts1Dan Cross
27 Mar 25            iiiiii    i            +* Re: MSI interrupts12MitchAlsup1
28 Mar 25            iiiiii    i            i`* Re: MSI interrupts11Dan Cross
28 Mar 25            iiiiii    i            i +- Re: MSI interrupts1MitchAlsup1
28 Mar 25            iiiiii    i            i +* Re: MSI interrupts5MitchAlsup1
28 Mar 25            iiiiii    i            i +* Re: MSI interrupts2Stefan Monnier
28 Mar 25            iiiiii    i            i `* Re: MSI interrupts2Chris M. Thomasson
27 Mar 25            iiiiii    i            `* Re: MSI interrupts9MitchAlsup1
25 Mar 25            iiiiii    `* Re: MSI interrupts18Dan Cross
20 Mar 25            iiiii`* Re: MSI interrupts8MitchAlsup1
19 Mar 25            iiii`* Re: MSI interrupts2MitchAlsup1
19 Mar 25            iii`* Re: MSI interrupts2Robert Finch
18 Mar 25            ii`* Re: MSI interrupts5MitchAlsup1
17 Mar 25            i`* Re: MSI interrupts7MitchAlsup1
17 Mar 25            `* Re: MSI interrupts3Robert Finch

Haut de la page

Les messages affichés proviennent d'usenet.

NewsPortal