RE: [EXTERNAL] Re: SVSM Development Call July 2nd, 2025
Jon Lange <[email protected]> Thu, 10 Jul 2025 04:32:58 +0000
| Newsgroups | dev.linux.lists.coconut-svsm,dev.linux.lists.linux-coco |
|---|---|
| Message-ID | <LV5PR21MB5210A051A344852EC30C089CCA48A@LV5PR21MB5210.namprd21.prod.outlook.com> |
> i.e. support for SVSM_CORE_DEPOSIT_MEM ? No, the SVSM knows the full memory map at boot time and therefore can know the amount of memory required for all bitmaps as part of its boot flow. The simplest approach will be for the SVSM to carve out whatever memory it requires, and to tell the guest (OVMF) what memory it has claimed for itself. There is no need for additional bitmap memory unless the guest support memory ballooning, which I think we can consider out of scope for now. So the question is what is the best way for the SVSM to advertise its carve-out to OVMF. It is the case that the guest can execute SVSM_CORE_PVALIDATE on every 4 KB page of available guest memory, and the SVSM must be prepared to handle such a case, so there must be sufficient space to track one bit per 4 KB page of available guest memory. It is also safe for the SVSM to return an error on any SVSM_CORE_PVALIDATE request that does not correspond to known guest memory, since there is no expectation for such a PVALIDATE request to succeed if there is no evidence that there is physical memory backing such an address. -Jon -----Original Message----- From: Gerd Hoffmann <[email protected]> Sent: Wednesday, July 9, 2025 2:44 AM To: Relph, Richard <[email protected]> Cc: Jörg Rödel <[email protected]>; [email protected]; [email protected] Subject: [EXTERNAL] Re: SVSM Development Call July 2nd, 2025 > > I assume you mean dynamic at boot time? > > Jon Lange should weigh in here... he's the expert on this. > But I believe it would need to be dynamic at run-time. i.e. support for SVSM_CORE_DEPOSIT_MEM ? > Being able start with a "reasonable" amount for page state while > allowing for expansion later if needed is the preferred solution. What events would trigger the need for expansion? Would OVMF (which typically would run for a few seconds at boot) need to support that? Or is that something only the linux kernel (or other guest kernels) would have to worry about? take care, Gerd