On Wed, 3 Dec 2025 12:12:49 +0000
Daniel P. Berrangé <[email protected]> wrote:
> On Wed, Dec 03, 2025 at 11:36:12AM +0000, Daniel P. Berrangé wrote:
> > On Wed, Dec 03, 2025 at 11:26:15AM +0000, Daniel P. Berrangé wrote:
> > > Normally bash in F42 runs with a prompt and we see
> > >
> > > 61385 ioctl(0,TCGETS,0x00007f34260697e8) = 0 ({c_iflag =
> > > ICRNL|IXON,c_oflag = OPOST|ONLCR,c_cflag = B38400,CS8,CREAD,c_lflag =
> > > ISIG|ICANON|ECHO|ECHOE|ECHOK|ECHOCTL|ECHOKE|IEXTEN,c_cc = "^C^\^?
> > > ^U^D",c_line = '^@'})
> > > 61385 ioctl(2,TCGETS,0x00007f34260697e8) = 0 ({c_iflag =
> > > ICRNL|IXON,c_oflag = OPOST|ONLCR,c_cflag = B38400,CS8,CREAD,c_lflag =
> > > ISIG|ICANON|ECHO|ECHOE|ECHOK|ECHOCTL|ECHOKE|IEXTEN,c_cc = "^C^\^?
> > > ^U^D",c_line = '^@'})
> > >
> > > When I redirect stderr to a file/pipe the latter ioctl
> > > obviously fails:
> > >
> > > 60841 ioctl(0,TCGETS,0x00007f19561fe7e8) = 0 ({c_iflag =
> > > ICRNL|IXON,c_oflag = OPOST|ONLCR,c_cflag = B38400,CS8,CREAD,c_lflag =
> > > ISIG|ICANON|ECHO|ECHOE|ECHOK|ECHOCTL|ECHOKE|IEXTEN,c_cc = "^C^\^?
> > > 60841 ioctl(2,TCGETS,0x00007f19561fe7e8) = -1 errno=25 (Inappropriate
> > > ioctl for device)
> > >
> > >
> > > Now looking at F43 bash I no longer see QEMU reporting TCGETS
> > > ioctl, instead I see
> > >
> > > 61809 ioctl(0,0x802c542a,0xd73fe8f4) = -1 errno=25 (Inappropriate ioctl
> > > for device)
> > >
> > > that suggests to me that bash is probably using a different (new?)
> > > ioctl that QEMU does not have emulation for currently.
> >
> > IIUC, 0x802c542a corresponds to TCGETS2, and indeed QEMU does not
> > currently emulate TCGETS2 which can be seen with QEMU's logging:
> >
> > # qemu-s390x -d unimp -L `pwd` ./usr/bin/bash 2>&1
> > Unsupported syscall: 383
> > Unsupported ioctl: cmd=0xffffffff802c542a
> >
> > IOW, F43 or later will be broken until this is impl in QEMU
>
> Already reported to QEMU upstream at
>
> https://gitlab.com/qemu-project/qemu/-/issues/3065
>
> And in Fedora context I see it affects aarch64 containers in addition
> to s390x, but interestly NOT the ppc64le containers (when tested from
> an x86_64 host).
>
> I'm unclear if x86_64 is also affected when run from from a non-x86_64
> host since I don't have a machine to test that scenario on
I think it is affected, it behaves the same as the s390x image on my
ppc64le host
Dan
--
_______________________________________________
devel mailing list -- [email protected]
To unsubscribe send an email to [email protected]
Fedora Code of Conduct:
https://docs.fedoraproject.org/en-US/project/code-of-conduct/
List Guidelines: https://fedoraproject.org/wiki/Mailing_list_guidelines
List Archives:
https://lists.fedoraproject.org/archives/list/[email protected]
Do not reply to spam, report it:
https://pagure.io/fedora-infrastructure/new_issue
