<br><br><div class="gmail_quote">2011/4/20 microcai <span dir="ltr"><<a href="mailto:microcai@fedoraproject.org">microcai@fedoraproject.org</a>></span><br><blockquote class="gmail_quote" style="margin: 0pt 0pt 0pt 0.8ex; border-left: 1px solid rgb(204, 204, 204); padding-left: 1ex;">
-----BEGIN PGP SIGNED MESSAGE-----<br>
Hash: SHA1<br>
<br>
于 2011年04月20日 20:39, Tom Cooksey 写道:<br>
<div><div></div><div class="h5">> Hi All,<br>
><br>
> Please excuse the spam... just trying to spread the word: Linaro is<br>
> currently developing a graphics memory manager for Linux which allows<br>
> buffers to be shared between different devices and different processes. On<br>
> ARM SoCs, many different hw devices need to access the same buffer (which<br>
> E.g. GEM doesn't support). As a simple example, a GPU needs to write to a<br>
> buffer which a display controller reads from. At least in ARM-Linux land,<br>
> this is a fairly hot-topic and a dedicated mailing list has been created to<br>
> discuss different requirements, figure out if TTM/GEM can be adapted, etc.<br>
><br>
><br>
> The mailing list is here:<br>
><br>
> <a href="http://lists.linaro.org/mailman/listinfo/linaro-mm-sig" target="_blank">http://lists.linaro.org/mailman/listinfo/linaro-mm-sig</a><br>
><br>
><br>
> There's also a wiki page gathering requirements here:<br>
><br>
> <a href="https://wiki.linaro.org/WorkingGroups/Middleware/Graphics/Projects/UnifiedMemoryManagement" target="_blank">https://wiki.linaro.org/WorkingGroups/Middleware/Graphics/Projects/UnifiedMemoryManagement</a><br>
><br>
><br>
> Hopefully there's people on the Wayland list which this will be interesting<br>
> to. Would be cool if a Wayland compositor used the mechanism Linaro comes up<br>
> with to share window buffers between different processes!<br>
><br>
><br>
><br>
> Cheers,<br>
><br>
> Tom<br>
><br>
><br>
><br>
<br>
</div></div>Not again, due!<br>
Does DRI has some mentally wrong that cann't be fixed other than re-create?<br></blockquote><div><br>Yes, possibly. The existing memory managers in Linux (GEM & TTM) don't:<br><br>a) Allow buffers to be shared between different device drivers<br>
b) Don't allow physically contiguous memory allocations (something simple hardware without MMUs need)<br></div><div><br>So while I agree re-inventing something just for the sake of it is bad, that's not what's happening. In fact, I think there's a good chance TTM will be ripped out, stuck into its own driver (with its own device node /dev/ttm) and extended to meet everyone's requirements. At least there has been some feasibility work done by Dave Airlie to partially address issue (a): <a href="http://airlied.livejournal.com/71734.html">http://airlied.livejournal.com/71734.html</a>.<br>
<br><br></div></div>Cheers,<br><br>Tom<br><br>