I haven't followed the associativity discussion closely, but perhaps it needs to be folden in there instead. Ard. > On 8 okt. 2015, at 03:25, Roy Franz <roy.franz@xxxxxxxxxx> wrote: > > On Wed, Oct 7, 2015 at 1:35 AM, Ard Biesheuvel > <ard.biesheuvel@xxxxxxxxxx> wrote: >> This is an alternative approach to some of the issues addressed by my series >> 'arm64 UEFI early FDT handling', of which the latest version (v3) can be found >> here: http://thread.gmane.org/gmane.linux.kernel.efi/6334 >> >> The issues that are addressed by the original series are: >> - when booting via UEFI, memreserve entries are removed from the device tree but >> the /reserved-memory node is not >> - memory nodes are removed from the device tree in a way that is not officially >> supported by the libfdt API (i.e., you cannot delete nodes while traversing >> the tree) >> - removal of memory nodes may discard annotations (such as NUMA topology) that >> should ideally be retained, or may corrupt the tree by discarding nodes >> referenced by phandles. >> >> After having discussed this offline between Grant, Leif and myself, we have come >> to the [preliminary] conclusion that discarding the /reserved-memory is not the >> way to go, and we will be far better off honoring those reservations to the >> extent possible. At the same time, the /memory node binding does not contain any >> annotations, so discarding those prevents more issues than it creates as long as >> we do it cautiously. > So if/when NUMA annotations are added to the memory node, we will need > to revisit this? Yes. But now, I am not so convinced anymore that /memory is where these annotations should go in the first place. Preserving but ignoring /memory means you are preserving info that we don't know is accurate (i.e., in sync with the UEFI memory map) > >> >> This series addresses both issues: it it sanity check the /reserved-memory node >> to ensure that it does not contain any reservations we will not be able to >> honor (patch #1, #2) and updates the /memory removal loop to restart from the >> root of the tree each time a /memory node is found and removed (patch #3) >> >> Ard Biesheuvel (3): >> efi/libstub: move FDT sanity check out of allocation loop >> efi/libstub: sanity check the /reserved-memory DT node >> efi/libstub: fix deletion of FDT memory nodes >> >> drivers/firmware/efi/libstub/fdt.c | 187 ++++++++++++++++++-- >> 1 file changed, 169 insertions(+), 18 deletions(-) >> >> -- >> 1.9.1 >> >> -- >> To unsubscribe from this list: send the line "unsubscribe linux-efi" in >> the body of a message to majordomo@xxxxxxxxxxxxxxx >> More majordomo info at http://vger.kernel.org/majordomo-info.html -- To unsubscribe from this list: send the line "unsubscribe linux-efi" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
