Red Hat Bugzilla – Bug 1515418
RFE: Provide diagnostics for failed boot
Last modified: 2018-04-10 12:31:21 EDT
Description of problem: If you try and boot a non-EFI disk image on OVMF, it quite often just sits at the logo and doesn't tell the user anything about what went wrong; it looks like it just hung. We need some diagnostics so it's obvious to the user they screwed up and need to fix their disk image. This is especially important since we elected to remove both the EFI shell and the BIOS compatibility shim which 1) removes the easy way to debug and 2) makes it much more likely you'll hit it. Oh and 3) EFI disc images are much more complex so screwing them up is much easier. Version-Release number of selected component (if applicable): OVMF-20170228-5 How reproducible: 100% Steps to Reproduce: 1. Take a BIOS boot disk image 2. Boot it on an OVMF configured host 3. Actual results: an OVMF boot logo screen Expected results: Trying to boot ..... FAILED because .... Additional info:
Created attachment 1356689 [details] prototype on top of ovmf-20171011-1.git92d07e48907f.el7 This is just a quick PoC. The actual upstream solution should use status code routing / handling / reporting. We already have some REPORT_STATUS_CODE() invocations in EfiBootManagerBoot() and BdsEntry(); some new ones should be added, and the current ones should emit more information (e.g. with EFI_STATUS_CODE_DATA_TYPE_DEBUG_GUID).
Posted upstream series: [edk2] [PATCH 0/5] MdeModulePkg, OvmfPkg: more easily visible boot progress reporting http://mid.mail-archive.com/20171122235849.4177-1-lersek@redhat.com https://lists.01.org/pipermail/edk2-devel/2017-November/017850.html
Steps for virt-QE, from Dave: """ a) Install a normal BIOS VM (e.g. Fedora or RHEL) b) Use virt-manager to create a new VM, telling it to import the disk image from the VM in (a) but to customize the VM before install c) During the customize, change the machine type/bios to q35+ovmf d) Let it boot Current versions stick at the Tianocore logo This fix gives useful feedback and lets you get to the boot manager. """ Thanks.
Fix included in ovmf-20171011-4.git92d07e48907f.el7
Reproduced this bug with OVMF-20171011-1.git92d07e48907f.el7.noarch & qemu-kvm-rhev-2.10.0-11.el7.x86_64. Result: 1)stick at the Tianocore logo 2)OVMF log: Error: Image at 0007CEA3000 start failed: Unsupported Error: Image at 0007CDDF000 start failed: Unsupported Error: Image at 0007CD6D000 start failed: Unsupported Error: Image at 0007BE1A000 start failed: Unsupported Error: Image at 0007BD34000 start failed: Aborted [Bds] Unable to boot! 3) cann't get to the boot manager Verified this bug with OVMF-20171011-4.git92d07e48907f.el7.noarch & qemu-kvm-rhev-2.10.0-11.el7.x86_64. Result: 1)useful feedback as below: BdsDxe: No bootable option or device was found. BdsDxe: Press any key to enter the Boot Manager Menu. 2)type Enter, I can get to the boot manager. 3)part of ovmf log as below: ....... SataControllerStart error return status = Already started BlockSize : 512 LastBlock : 77FFFFF BlockSize : 512 LastBlock : 1FFFFF BlockSize : 512 LastBlock : 75FF7FF SataControllerStart START SataControllerStart error return status = Already started BlockSize : 512 LastBlock : 77FFFFF BlockSize : 512 LastBlock : 1FFFFF BlockSize : 512 LastBlock : 75FF7FF BlockSize : 512 LastBlock : 77FFFFF BlockSize : 512 LastBlock : 1FFFFF BlockSize : 512 LastBlock : 75FF7FF BlockSize : 512 LastBlock : 1FFFFF BlockSize : 512 LastBlock : 75FF7FF [Bds] Expand \EFI\BOOT\BOOTX64.EFI -> <null string>
According to comment18 & comment22, set this bug as verified.
Since the problem described in this bug report should be resolved in a recent advisory, it has been closed with a resolution of ERRATA. For information on the advisory, and where to find the updated files, follow the link below. If the solution does not work for you, open a new bug report. https://access.redhat.com/errata/RHBA-2018:0902