1.6.5 to 1.6.6 Errors
Opened this issue · 2 comments
Running 1.6.6 [git] produces the following warnings:
[WRN] Unexpected instruction at 0xfffffff00766d6d4
[WRN] Unexpected instruction at 0xfffffff00766d6d4
[WRN] Multiple vtab candidates for OSNumber
[WRN] Multiple vtab candidates for IOService
[WRN] Skipping class OSNumber because vtable is missing.
[WRN] Skipping class IOService because vtable is missing.
[WRN] Skipping class IOResources because parent class was skipped.
[WRN] Skipping class IOPMPowerSource because parent class was skipped.
[WRN] Skipping class IOMapper because parent class was skipped.
[WRN] Skipping class IOPlatformExpert because parent class was skipped.
[WRN] Skipping class IODTPlatformExpert because parent class was skipped.
[WRN] Skipping class IOPlatformExpertDevice because parent class was skipped.
[WRN] Skipping class IOPlatformDevice because parent class was skipped.
[WRN] Skipping class IOPanicPlatform because parent class was skipped.
[WRN] Skipping class IOCPU because parent class was skipped.
[WRN] Skipping class IOInterruptController because parent class was skipped.
[WRN] Skipping class IOCPUInterruptController because parent class was skipped.
[WRN] Skipping class IODTNVRAM because parent class was skipped.
[WRN] Skipping class IODMAController because parent class was skipped.
[WRN] Skipping class IOSharedInterruptController because parent class was skipped.
[WRN] Symmap entry for IOUserIterator has 0 methods, vtab has 1.
[WRN] Skipping class IOUserClient because parent class was skipped.
[WRN] Skipping class IOKitDiagnosticsClient because parent class was skipped.
[WRN] Skipping class IORTC because parent class was skipped.
............
and not much output is produced
1.6.5 produces no errors. The "Unexpected instruction" seems to be a CMP
The above output was produced after running iometa on an iPhone7,1 16G201 kernel
I think that comes from me adding handling for more code paths that were previously skipped.
I have local commits that add support for adds/subs (which includes cmp), but now I have the issue of multiple vtab candidates on the kernel you mention. :|
But I'll look into this.