<!DOCTYPE html><html><head> <meta http-equiv="Content-Type" content="text/html; charset=utf-8"> </head> <body> Am 31.01.24 um 11:20 schrieb Zhang, Julia: <blockquote type="cite" cite="mid:IA1PR12MB653270F7FD75C1B4DF8B90E8F27C2@IA1PR12MB6532.namprd12.prod.outlook.com"> <pre class="moz-quote-pre" wrap="">On 2024/1/30 22:23, Christian König wrote: </pre> <blockquote type="cite"> <pre class="moz-quote-pre" wrap="">Am 30.01.24 um 12:16 schrieb Daniel Vetter: </pre> <blockquote type="cite"> <pre class="moz-quote-pre" wrap="">On Tue, Jan 30, 2024 at 12:10:31PM +0100, Daniel Vetter wrote: </pre> <blockquote type="cite">[SNIP] <pre class="moz-quote-pre" wrap=""></pre> </blockquote> </blockquote> <pre class="moz-quote-pre" wrap=""> </pre> </blockquote> <pre class="moz-quote-pre" wrap="">Hi Sima, Christian, </pre> <blockquote type="cite"> <pre class="moz-quote-pre" wrap="">Yeah, that is really just speculative. All importers need to set the peer2peer flag just in case. </pre> </blockquote> <pre class="moz-quote-pre" wrap=""> I see, I will modify this. </pre> <blockquote type="cite"> <pre class="moz-quote-pre" wrap=""> What happens under the hood is that IOMMU redirects the "VRAM" memory access to whatever address the DMA-buf on the host is pointing to (system, VRAM, doorbell, IOMMU, whatever). I'm also not 100% sure if all the cache snooping is done correctly in all cases, but for now it seems to work. </pre> <blockquote type="cite"> <blockquote type="cite"> <blockquote type="cite"> <pre class="moz-quote-pre" wrap=""> </pre> </blockquote> <pre class="moz-quote-pre" wrap="">Frankly the more I look at the original patch that added vram export support the more this just looks like a "pls revert, this is just too broken". </pre> </blockquote> <pre class="moz-quote-pre" wrap="">The commit I mean is this one: ea5ea3d8a117 ("drm/virtio: support mapping exported vram"). The commit message definitely needs to cite that one, and also needs a cc: stable because not rejecting invalid imports is a pretty big deal. </pre> </blockquote> <pre class="moz-quote-pre" wrap=""> Yeah, I've pointed out that commit in an internal discussion as well. I was just not aware that it's that severely broken. </pre> </blockquote> <pre class="moz-quote-pre" wrap=""> Yeah we have mentioned this patch before, but I don't totally understand why this is too broken. Without exporting vram objects, dGPU prime feature would not be realized. Would you mind to explain more about it. Thanks!</pre> </blockquote> One reason is that using sg tables without struct pages is actually a hack we came up with because we couldn't hope to clean up the sg table structure any time soon to not include struct page pointers. Another reason is that using this with devices which don't expect a DMA address pointing into a virtual PCI BAR. So doing this without checking the peer2peer flag can most likely cause quite a bit of trouble. Regards, Christian. <blockquote type="cite" cite="mid:IA1PR12MB653270F7FD75C1B4DF8B90E8F27C2@IA1PR12MB6532.namprd12.prod.outlook.com"> <pre class="moz-quote-pre" wrap=""> Best regards, Julia </pre> <blockquote type="cite"> <pre class="moz-quote-pre" wrap="">Regards, Christian. </pre> </blockquote> </blockquote> </body> </html>