Please apply commit 2347961b11d4 ("binfmt_misc: pass binfmt_misc flags to the interpreter") to 5.10.y and later
Hi
2347961b11d4 ("binfmt_misc: pass binfmt_misc flags to the interpreter") was applied in mainline and included in 5.12-rc1.
Probably you could argue here on both a bugfix or feature addition.
My intention is the following: In the Debian bugreport
https://bugs.debian.org/970460 an issue was raised with qemu-user which needs to know if it has to preserve the argv[0]. As shown there it is an issue with multi-call binaries.
So again, not sure if you want to consider it, but defintively Yunqiang Su and others would appreicate. If it gets backported we will pick it up automatically.
Regards,
Salvatore
Sasha Levin
maintainerQuoted 7 lines · Salvatore Bonaccorso
Hi
2347961b11d4("binfmt_misc: passbinfmt_miscflags to the interpreter") was applied in mainline and included in 5.12-rc1.Probably you could argue here on both a bugfix or feature addition.
Quoted 8 lines
My intention is the following: In the Debian bugreport
https://bugs.debian.org/970460 an issue was raised with qemu-user which needs to know if it has to preserve the argv[0]. As shown there it is an issue with multi-call binaries.So again, not sure if you want to consider it, but defintively Yunqiang Su and others would appreicate. If it gets backported we will pick it up automatically.
Given it needs changes in userspace too I'm not sure it qulifies as a fix per-se. Though as you point out, it's really borderline.
Generally I would just take it, but it affects interactions with userspace, so I feel less comfortable with this.
Signature · 2 lines
Thanks, Sasha
Greg Kroah-Hartman
maintainerQuoted 17 lines · Sasha Levin
Quoted 7 lines · Salvatore Bonaccorso
Hi
2347961b11d4("binfmt_misc: passbinfmt_miscflags to the interpreter") was applied in mainline and included in 5.12-rc1.Probably you could argue here on both a bugfix or feature addition.
Quoted 8 lines
My intention is the following: In the Debian bugreport
https://bugs.debian.org/970460 an issue was raised with qemu-user which needs to know if it has to preserve the argv[0]. As shown there it is an issue with multi-call binaries.So again, not sure if you want to consider it, but defintively Yunqiang Su and others would appreicate. If it gets backported we will pick it up automatically.
Quoted 5 lines
Given it needs changes in userspace too I'm not sure it qulifies as a fix per-se. Though as you point out, it's really borderline.
Generally I would just take it, but it affects interactions with userspace, so I feel less comfortable with this.
Yes, I'm not ok with taking this either at this point in time, as it requires userspace changes.
thanks,
greg k-h