Date: Sat, 25 Apr 2020 14:14:46 +0000 From: greg@unrelenting.technology To: "Mark Murray" <markm@freebsd.org>, "Marcel Flores" <marcel@brickporch.com> Cc: freebsd-arm@freebsd.org Subject: Re: Bootable image for Macchatobin Double Shot? Message-ID: <092721df6b1de3e75820acd32ba1b0e7@unrelenting.technology> In-Reply-To: <9F7FB8F8-A29C-4F95-B0BB-CFEFEAFDDD5E@FreeBSD.org> References: <9F7FB8F8-A29C-4F95-B0BB-CFEFEAFDDD5E@FreeBSD.org> <1039B382-2CA4-4F49-9F95-08BD1386A447@FreeBSD.org> <CA07A9C8-267B-4657-B9D1-1F53B62A969B@brickporch.com>
index | next in thread | previous in thread | raw e-mail
April 25, 2020 3:21 PM, "Mark Murray" <markm@freebsd.org> wrote: > I tried a FreeBSD-12 image - boots with ACPI, but PCI not present. With UEFI DTB > I get the same hang. It is present: > pcib0: <Generic PCI host controller> on acpi0 > pci0: <PCI bus> on pcib0 But not the device. The thing is, the PCIe controller has a bug (in ECAM mode). It doesn't properly filter something, so a device could appear replicated multiple times. So the Semihalf engineers decided to do a workaround: move the memory address in ACPI so that the OS only sees the last device. I don't know which devices they tried, but my Radeon RX 480 *does not* get replicated. Same with an Intel and Mellanox NIC. A Radeon HD 7970 was replicated but only twice, not 32 times. You can try my EDK2 builds, at least one of them should work. I reverted that workaround. https://unrelentingtech.s3.dualstack.eu-west-1.amazonaws.com/flash-image.bin https://unrelentingtech.s3.dualstack.eu-west-1.amazonaws.com/flash-image2.bin Relevant discussion links: https://github.com/MarvellEmbeddedProcessors/edk2-open-platform/issues/11 https://github.com/andreiw/MacchiatoBin-edk2/commit/2c49a74d3c0438b656e56ef133a4699ab3f8b22f#commitcomment-34139610home | help
Want to link to this message? Use this
URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?092721df6b1de3e75820acd32ba1b0e7>