Re: [PATCH v9 03/11] sandbox: Add a -N flag to control on-host behaviour
Hi Tom, On Thu, 31 Oct 2024 at 19:14, Tom Rini wrote: > > On Thu, Oct 31, 2024 at 06:52:33PM +0100, Simon Glass wrote: > > Hi Tom, > > > > On Tue, 29 Oct 2024 at 21:14, Tom Rini wrote: > > > > > > On Tue, Oct 29, 2024 at 08:22:11PM +0100, Simon Glass wrote: > > > > > > > Sandbox is its own architecture, but sometimes we want to mimic the host > > > > architecture, e.g. when running an EFI app not built by U-Boot. > > > > > > > > Add a -N/--native flag which tells sandbox to reflect the architecture > > > > of the host. > > > > > > > > Signed-off-by: Simon Glass > > > > --- > > > > > > > > (no changes since v8) > > > > > > I believe Heinrich asked that you invert the logic here and add a flag > > > for when you need the unexpected behavior. > > > > Yes, but that is not correct, unfortunately. The expected behaviour is > > for CI to use the same filename for its efi apps, no matter what host > > architecture it is happens to be running on. > > > > The use case that Heinrich mentions seems to be broken on x86_64 now, > > as I mentioned on the other thread, perhaps due to a recently > > introduced bug in UEFI or QEMU. Just to be clear, this is not a > > problem with sandbox. I suspect Heinrich will figure that out at some > > point and then I will be able to try it too. I would also like to run > > it on sandbox. > > > > I hope to get an ARM machine running CI eventually, but I need to get > > my lab running first. > > I talked with Heinrich about this in private and while neither of us > understand why you want this behavior, we'll stop objecting to it. OK, good. Heinrich mentioned that the RISC-V server did not work, for some reason. But eventually we might have one of those. I know from my fiddling with 32-bit rpi that even *running* sandbox on ARM will be a challenge, let alone passing tests, but we'll see. We should encourage ARM/Linux to set up an ARM runner at some point. Regards, Simon
Re: [PATCH v9 03/11] sandbox: Add a -N flag to control on-host behaviour
On Thu, Oct 31, 2024 at 06:52:33PM +0100, Simon Glass wrote: > Hi Tom, > > On Tue, 29 Oct 2024 at 21:14, Tom Rini wrote: > > > > On Tue, Oct 29, 2024 at 08:22:11PM +0100, Simon Glass wrote: > > > > > Sandbox is its own architecture, but sometimes we want to mimic the host > > > architecture, e.g. when running an EFI app not built by U-Boot. > > > > > > Add a -N/--native flag which tells sandbox to reflect the architecture > > > of the host. > > > > > > Signed-off-by: Simon Glass > > > --- > > > > > > (no changes since v8) > > > > I believe Heinrich asked that you invert the logic here and add a flag > > for when you need the unexpected behavior. > > Yes, but that is not correct, unfortunately. The expected behaviour is > for CI to use the same filename for its efi apps, no matter what host > architecture it is happens to be running on. > > The use case that Heinrich mentions seems to be broken on x86_64 now, > as I mentioned on the other thread, perhaps due to a recently > introduced bug in UEFI or QEMU. Just to be clear, this is not a > problem with sandbox. I suspect Heinrich will figure that out at some > point and then I will be able to try it too. I would also like to run > it on sandbox. > > I hope to get an ARM machine running CI eventually, but I need to get > my lab running first. I talked with Heinrich about this in private and while neither of us understand why you want this behavior, we'll stop objecting to it. -- Tom signature.asc Description: PGP signature
Re: [PATCH v9 03/11] sandbox: Add a -N flag to control on-host behaviour
Hi Tom, On Tue, 29 Oct 2024 at 21:14, Tom Rini wrote: > > On Tue, Oct 29, 2024 at 08:22:11PM +0100, Simon Glass wrote: > > > Sandbox is its own architecture, but sometimes we want to mimic the host > > architecture, e.g. when running an EFI app not built by U-Boot. > > > > Add a -N/--native flag which tells sandbox to reflect the architecture > > of the host. > > > > Signed-off-by: Simon Glass > > --- > > > > (no changes since v8) > > I believe Heinrich asked that you invert the logic here and add a flag > for when you need the unexpected behavior. Yes, but that is not correct, unfortunately. The expected behaviour is for CI to use the same filename for its efi apps, no matter what host architecture it is happens to be running on. The use case that Heinrich mentions seems to be broken on x86_64 now, as I mentioned on the other thread, perhaps due to a recently introduced bug in UEFI or QEMU. Just to be clear, this is not a problem with sandbox. I suspect Heinrich will figure that out at some point and then I will be able to try it too. I would also like to run it on sandbox. I hope to get an ARM machine running CI eventually, but I need to get my lab running first. Regards, Simon
Re: [PATCH v9 03/11] sandbox: Add a -N flag to control on-host behaviour
On Tue, Oct 29, 2024 at 08:22:11PM +0100, Simon Glass wrote: > Sandbox is its own architecture, but sometimes we want to mimic the host > architecture, e.g. when running an EFI app not built by U-Boot. > > Add a -N/--native flag which tells sandbox to reflect the architecture > of the host. > > Signed-off-by: Simon Glass > --- > > (no changes since v8) I believe Heinrich asked that you invert the logic here and add a flag for when you need the unexpected behavior. -- Tom signature.asc Description: PGP signature
