RE: [PATCH] Adds a new ioctl32 syscall for backwards compatibility layers
From: Christoph Hellwig > Sent: 16 January 2021 09:07 ... > > I personally would like to see in_compat_syscall() go away, > > but some other people (Hi, Christoph!) disagree, and usage seems to be > > increasing, not decreasing. > > I'm absolutely against it going away. in_compat_syscall helped to > remove so much crap compared to the explicit compat syscalls. The only other real option is to pass the 'syscall type' explicitly through all the layers into every piece of code that might need it. So passing it as a 'parameter' that is (probably) current->syscall_type does make sense. It might even make sense have separate bits for the required emulations. So you'd have separate bits for '32bit pointers' and '64bit items 32bit aligned' (etc). David - Registered Address Lakeside, Bramley Road, Mount Farm, Milton Keynes, MK1 1PT, UK Registration No: 1397386 (Wales)
Re: [PATCH] Adds a new ioctl32 syscall for backwards compatibility layers
On Fri, Jan 15, 2021 at 04:07:46PM -0800, Andy Lutomirski wrote: > Finally, I'm not convinced that this patch works correctly. We have > in_compat_syscall(), and code that uses it may well be reachable from > ioctl. ioctls are the prime user of in_compat_syscall(). > I personally would like to see in_compat_syscall() go away, > but some other people (Hi, Christoph!) disagree, and usage seems to be > increasing, not decreasing. I'm absolutely against it going away. in_compat_syscall helped to remove so much crap compared to the explicit compat syscalls.
Re: [PATCH] Adds a new ioctl32 syscall for backwards compatibility layers
On Fri, Jan 15, 2021 at 1:03 AM Arnd Bergmann wrote: > > On Fri, Jan 15, 2021 at 3:06 AM Ryan Houdek wrote: > > On Wed, Jan 6, 2021 at 12:49 AM Arnd Bergmann wrote: > >> On Wed, Jan 6, 2021 at 7:48 AM wrote: > >> > From: Ryan Houdek > >> ... > >> > >> For x86, this has another complication, as some ioctls also need to > >> check whether they are in an ia32 task (with packed u64 and 32-bit > >> __kernel_old_time_t) or an x32 task (with aligned u64 and 64-bit > >> __kernel_old_time_t). If the new syscall gets wired up on x86 as well, > >> you'd need to decide which of the two behaviors you want. > > > > > > I can have a follow-up patch that makes this do ni-syscall on x86_64 since > > we can go through the int 0x80 handler, or x32 handler path and choose > > whichever one there. > > I'd say for consistency > We need to make it crystal clear on x86 what this ioctl does. We have a silly selection of options: - ioctl32() via SYSCALL, x32 bit clear -- presumably does an i386 ioctl? - ioctl32() via SYSCALL, x32 bit set -- this needs to do something clearly documented. - ioctl32() via int80 -- presumably you're not wiring this up In any case, the compat alloc thing should just go away. It's a hack and serves no real purpose. Finally, I'm not convinced that this patch works correctly. We have in_compat_syscall(), and code that uses it may well be reachable from ioctl. I personally would like to see in_compat_syscall() go away, but some other people (Hi, Christoph!) disagree, and usage seems to be increasing, not decreasing.
Re: [PATCH] Adds a new ioctl32 syscall for backwards compatibility layers
On Fri, Jan 15, 2021 at 11:17:09PM +0100, Arnd Bergmann wrote: > On Fri, Jan 15, 2021 at 9:01 PM David Laight wrote: > > > > From: sonicadvan...@gmail.com > > > Sent: 15 January 2021 07:03 > > > Problem presented: > > > A backwards compatibility layer that allows running x86-64 and x86 > > > processes inside of an AArch64 process. > > > - CPU is emulated > > > - Syscall interface is mostly passthrough > > > - Some syscalls require patching or emulation depending on behaviour > > > - Not viable from the emulator design to use an AArch32 host process > > > > > > > You are going to need to add all the x86 compatibility code into > > your arm64 kernel. > > This is likely to be different from the 32bit arm compatibility > > because 64bit items are only aligned on 32bit boundaries. > > The x86 x32 compatibility will be more like the 32bit arm 'compat' > > code - I'm pretty sure arm32 64bit aligned 64bit data. > > All other architectures that have both 32-bit and 64-bit variants > use the same alignment for all types, except for x86. > > There are additional differences though, especially if one > were to try to generalize the interface to all architectures. > A subset of the issues includes > > - x32 has 64-bit types in places of some types that are > 32 bit everywhere else (time_t, ino_t, off_t, clock_t, ...) > > - m68k aligns struct members to at most 16 bits > > - uid_t/gid_t/ino_t/dev_t/... are > > > You'll then need to remember how the process entered the kernel > > to work out which compatibility code to invoke. > > This is what x86 does. > > It allows a single process to do all three types of system call. > > > > Trying to 'patch up' structures outside the kernel, or in the > > syscall interface code will always cause grief somewhere. > > The only sane place is in the code that uses the structures. > > Which, for ioctls, means inside the driver that parses them. > > He's already doing the system call emulation for all the system > calls other than ioctl in user space though. In my experience, > there are actually fairly few ioctl commands that are different > between architectures -- most of them have no misaligned > or architecture-defined struct members at all. > > Once you have conversion functions to deal with the 32/64-bit > interface differences and architecture specifics of sockets, > sysvipc, signals, stat, and input_event, handling the > x86-32 specific ioctl commands is comparably easy. Indeed, all of this should just be done in userspace. Note (as you of course know, but others on CC probably don't) that we did this in musl libc for the sake of being able to run a time64 userspace on a pre-time64 kernel, with translation from the new time64 ioctl structures to the versions needed by the old ioctls and back using a fairly simple table: https://git.musl-libc.org/cgit/musl/tree/src/misc/ioctl.c?id=v1.2.2 I imagine there's a fair bit more to be done for 32-/64-bit mismatch in size/long/pointer types and different alignments, but the problem is almost certainly tractable, and much easier than what they already have to be doing for syscalls. Rich
RE: [PATCH] Adds a new ioctl32 syscall for backwards compatibility layers
... > He's already doing the system call emulation for all the system > calls other than ioctl in user space though. In my experience, > there are actually fairly few ioctl commands that are different > between architectures -- most of them have no misaligned > or architecture-defined struct members at all. Aren't there also some intractable issues with socket options? IIRC the kernel code that tried to change them to 64bit was horribly broken in some obscure cases. Pushing the conversion down the stack not only identified the issues, it also made them easier to fix. If you change the kernel so a 64bit process can execute 32bit system calls then a lot of the problems do go away. This is probably easiest done by setting a high bit on the system call number - as x86_64 does for x32 calls. You still have to solve the different alignment of 64bit data on i386. Of course the system call numbers are different - but that is just a lookup. David - Registered Address Lakeside, Bramley Road, Mount Farm, Milton Keynes, MK1 1PT, UK Registration No: 1397386 (Wales)
Re: [PATCH] Adds a new ioctl32 syscall for backwards compatibility layers
On Fri, Jan 15, 2021 at 9:01 PM David Laight wrote: > > From: sonicadvan...@gmail.com > > Sent: 15 January 2021 07:03 > > Problem presented: > > A backwards compatibility layer that allows running x86-64 and x86 > > processes inside of an AArch64 process. > > - CPU is emulated > > - Syscall interface is mostly passthrough > > - Some syscalls require patching or emulation depending on behaviour > > - Not viable from the emulator design to use an AArch32 host process > > > > You are going to need to add all the x86 compatibility code into > your arm64 kernel. > This is likely to be different from the 32bit arm compatibility > because 64bit items are only aligned on 32bit boundaries. > The x86 x32 compatibility will be more like the 32bit arm 'compat' > code - I'm pretty sure arm32 64bit aligned 64bit data. All other architectures that have both 32-bit and 64-bit variants use the same alignment for all types, except for x86. There are additional differences though, especially if one were to try to generalize the interface to all architectures. A subset of the issues includes - x32 has 64-bit types in places of some types that are 32 bit everywhere else (time_t, ino_t, off_t, clock_t, ...) - m68k aligns struct members to at most 16 bits - uid_t/gid_t/ino_t/dev_t/... are > You'll then need to remember how the process entered the kernel > to work out which compatibility code to invoke. > This is what x86 does. > It allows a single process to do all three types of system call. > > Trying to 'patch up' structures outside the kernel, or in the > syscall interface code will always cause grief somewhere. > The only sane place is in the code that uses the structures. > Which, for ioctls, means inside the driver that parses them. He's already doing the system call emulation for all the system calls other than ioctl in user space though. In my experience, there are actually fairly few ioctl commands that are different between architectures -- most of them have no misaligned or architecture-defined struct members at all. Once you have conversion functions to deal with the 32/64-bit interface differences and architecture specifics of sockets, sysvipc, signals, stat, and input_event, handling the x86-32 specific ioctl commands is comparably easy. Arnd
RE: [PATCH] Adds a new ioctl32 syscall for backwards compatibility layers
From: sonicadvan...@gmail.com > Sent: 15 January 2021 07:03 > Problem presented: > A backwards compatibility layer that allows running x86-64 and x86 > processes inside of an AArch64 process. > - CPU is emulated > - Syscall interface is mostly passthrough > - Some syscalls require patching or emulation depending on behaviour > - Not viable from the emulator design to use an AArch32 host process > You are going to need to add all the x86 compatibility code into your arm64 kernel. This is likely to be different from the 32bit arm compatibility because 64bit items are only aligned on 32bit boundaries. The x86 x32 compatibility will be more like the 32bit arm 'compat' code - I'm pretty sure arm32 64bit aligned 64bit data. You'll then need to remember how the process entered the kernel to work out which compatibility code to invoke. This is what x86 does. It allows a single process to do all three types of system call. Trying to 'patch up' structures outside the kernel, or in the syscall interface code will always cause grief somewhere. The only sane place is in the code that uses the structures. Which, for ioctls, means inside the driver that parses them. David - Registered Address Lakeside, Bramley Road, Mount Farm, Milton Keynes, MK1 1PT, UK Registration No: 1397386 (Wales)
Re: [PATCH] Adds a new ioctl32 syscall for backwards compatibility layers
On Fri, Jan 15, 2021 at 3:06 AM Ryan Houdek wrote: > On Wed, Jan 6, 2021 at 12:49 AM Arnd Bergmann wrote: >> On Wed, Jan 6, 2021 at 7:48 AM wrote: >> > From: Ryan Houdek >> ... >> >> For x86, this has another complication, as some ioctls also need to >> check whether they are in an ia32 task (with packed u64 and 32-bit >> __kernel_old_time_t) or an x32 task (with aligned u64 and 64-bit >> __kernel_old_time_t). If the new syscall gets wired up on x86 as well, >> you'd need to decide which of the two behaviors you want. > > > I can have a follow-up patch that makes this do ni-syscall on x86_64 since > we can go through the int 0x80 handler, or x32 handler path and choose > whichever one there. I'd say for consistency >> > 3a) Userspace consumes all VA space above 32bit. Forcing allocations to >> > occur in lower 32bits >> > - This is the current implementation >> > 3b) Ensure any allocation in the ioctl handles ioctl entrypoint rather >> > than just allow generic memory allocations in full VA space >> > - This is hard to guarantee >> >> What kind of allocation do you mean here? Can you give an example of >> an ioctl that does this? >> > My concern here would be something like DRM allocating memory and > returning a pointer to userspace that ends up in 64bit space. > I can see something like `drm_get_unmapped_area` calls in to > `current->mm->get_unmapped_area` which I believe only ends up falling > down TASK_SIZE checks. I see. > Which could potentially return pointers in the 64bit address space range > in this case. Theoretically can be resolved either by thieving the full 64bit > VA range, or doing something like the Tango layer patches that on > syscall entry changes the syscall to a "compat" syscall. > compat syscall flag like Tango might be nicer here? Not sure how that flag is best encoded, but yes, it would have to be somewhere that arch_get_unmapped_area() and arch_get_mmap_end() can find. Clearly we want a solution that works for both tango and for your work, as well as being portable to any architecture. >> > } >> > + >> > +COMPAT_SYSCALL_DEFINE3(ioctl, unsigned int, fd, unsigned int, cmd, >> > + compat_ulong_t, arg) >> > +{ >> > + return do_ioctl32(fd, cmd, arg); >> > +} >> > + >> > +SYSCALL_DEFINE3(ioctl32, unsigned int, fd, unsigned int, cmd, >> > + compat_ulong_t, arg) >> > +{ >> > + return do_ioctl32(fd, cmd, arg); >> > +} >> >> These two look identical to me, I don't think you need to add a wrapper >> here at all, but can just use the normal compat_sys_ioctl entry point >> unless you want to add a 'flags' argument to control the struct padding. > > > I tried having the dispatch table call directly in to the COMPAT one and > the way things were lining up weren't allowing me to do this. > Since this is a bit unique in how it operates, I'm not quite sure if there is > another example I could pull from for this. For the asm-generic/unistd.h, you should be able to write htis as #if __BITS_PER_LONG == 64 __SC_COMP(__NR_ioctl32, compat_sys_ioctl, sys_ni_syscall) #endif Which means that the native syscall in a 64-bit process always points to compat_sys_ioctl, while a 32-bit process always gets -ENOSYS. Similarly, the syscall_64.tbl file on x86 and the other 64-bit architectures would use 44264 ioctl32 compat_sys_ioctl FWIW, I suppose you can rename compat_sys_ioctl to sys_ioctl32 treewide, if that name makes more sense. Arnd
[PATCH] Adds a new ioctl32 syscall for backwards compatibility layers
From: Ryan Houdek Problem presented: A backwards compatibility layer that allows running x86-64 and x86 processes inside of an AArch64 process. - CPU is emulated - Syscall interface is mostly passthrough - Some syscalls require patching or emulation depending on behaviour - Not viable from the emulator design to use an AArch32 host process x86-64 and x86 userspace emulator source: https://github.com/FEX-Emu/FEX Usage of ioctl32 is currently in a downstream fork. This will be the first user of the syscall. Cross documentation: https://github.com/FEX-Emu/FEX/wiki/32Bit-x86-Woes#ioctl---54 ioctls are opaque from the emulator perspective and the data wants to be passed through a syscall as unimpeded as possible. Sadly due to ioctl struct differences between x86 and x86-64, we need a syscall that exposes the compatibility ioctl handler to userspace in a 64bit process. This is necessary behaves of the behaviour differences that occur between an x86 process doing an ioctl and an x86-64 process doing an ioctl. Both of which are captured and passed through the AArch64 ioctl space. This is implementing a new ioctl32 syscall that allows us to pass 32bit x86 ioctls through to the kernel with zero or minimal manipulation. The only supported hosts where we care about this currently is AArch64 and x86-64 (For testing purposes). PPC64LE, MIPS64LE, and RISC-V64 might be interesting to support in the future; But I don't have any platforms that get anywhere near Cortex-A77 performance in those architectures. Nor do I have the time to bring up the emulator on them. x86-64 can get to the compatibility ioctl through the int $0x80 handler. This does not solve the following problems: 1) compat_alloc_user_space inside ioctl 2) ioctls that check task mode instead of entry point for behaviour 3) ioctls allocating memory 4) struct packing problems between architectures Workarounds for the problems presented: 1a) Do a stack pivot to the lower 32bits from userspace - Forces host 64bit process to have its thread stacks to live in 32bit space. Not ideal. - Only do a stack pivot on ioctl to save previous 32bit VA space 1b) Teach kernel that compat_alloc_userspace can return a 64bit pointer - x86-64 truncates stack from this function - AArch64 returns the full stack pointer - Only ~29 users. Validating all of them support a 64bit stack is trivial? 2a) Any application using these can be checked for compatibility in userspace and put on a block list. 2b) Fix any ioctls doing broken behaviour based on task mode rather than ioctl entry point 3a) Userspace consumes all VA space above 32bit. Forcing allocations to occur in lower 32bits - This is the current implementation 3b) Ensure any allocation in the ioctl handles ioctl entrypoint rather than just allow generic memory allocations in full VA space - This is hard to guarantee 4a) Blocklist any application using ioctls that have different struct packing across the boundary - Can happen when struct packing of 32bit x86 application goes down the aarch64 compat_ioctl path - Userspace is a AArch64 process passing 32bit x86 ioctl structures through the compat_ioctl path which is typically for AArch32 processes - None currently identified 4b) Work with upstream kernel and userspace projects to evaluate and fix - Identify the problem ioctls - Implement a new ioctl with more sane struct packing that matches cross-arch - Implement new ioctl while maintaining backwards compatibility with previous ioctl handler - Change upstream project to use the new compatibility ioctl - ioctl deprecation will be case by case per device and project 4b) Userspace implements a full ioctl emulation layer - Parses the full ioctl tree - Either passes through ioctls that it doesn't understand or transforms ioctls that it knows are trouble - Has the downside that it can still run in to edge cases that will fail - Performance of additional tracking is a concern - Prone to failure keeping the kernel ioctl and userspace ioctl handling in sync - Really want to have it in the kernel space as much as possible Signed-off-by: Ryan Houdek --- arch/alpha/kernel/syscalls/syscall.tbl | 1 + arch/arm/tools/syscall.tbl | 1 + arch/arm64/include/asm/unistd.h | 2 +- arch/arm64/include/asm/unistd32.h | 2 ++ arch/ia64/kernel/syscalls/syscall.tbl | 1 + arch/m68k/kernel/syscalls/syscall.tbl | 1 + arch/microblaze/kernel/syscalls/syscall.tbl | 1 + arch/mips/kernel/syscalls/syscall_n32.tbl | 1 + arch/mips/kernel/syscalls/syscall_n64.tbl | 2 ++ arch/mips/kernel/syscalls/syscall_o32.tbl | 1 + arch/parisc/kernel/syscalls/syscall.tbl | 1 + arch/powerpc/kernel/syscalls/syscall.tbl| 1 + arch/s390/kernel/syscalls/syscall.tbl | 1 + arch/sh/kernel/syscalls/syscall.tbl | 1 + arch/sparc/kernel/syscalls/syscall.tbl | 1 + arch/x86/entry/syscalls/syscall_32.tbl
Re: [PATCH] Adds a new ioctl32 syscall for backwards compatibility layers
On Wed, Jan 6, 2021 at 12:49 AM Arnd Bergmann wrote: > > On Wed, Jan 6, 2021 at 7:48 AM wrote: > > From: Ryan Houdek > ... > > This does not solve the following problems: > > 1) compat_alloc_user_space inside ioctl > > 2) ioctls that check task mode instead of entry point for behaviour > > 3) ioctls allocating memory > > 4) struct packing problems between architectures > > > > Workarounds for the problems presented: > > 1a) Do a stack pivot to the lower 32bits from userspace > > - Forces host 64bit process to have its thread stacks to live in 32bit > > space. Not ideal. > > - Only do a stack pivot on ioctl to save previous 32bit VA space > > 1b) Teach kernel that compat_alloc_userspace can return a 64bit pointer > > - x86-64 truncates stack from this function > > - AArch64 returns the full stack pointer > > - Only ~29 users. Validating all of them support a 64bit stack is > > trivial? > > I've almost completed the removal of compat_alloc_user_space(), > that should no longer be a concern when the syscall gets added. > > > 2a) Any application using these can be checked for compatibility in > > userspace and put on a block list. > > 2b) Fix any ioctls doing broken behaviour based on task mode rather than > > ioctl entry point > > What the ioctls() actually check is 'in_compat_syscall()', which is not > the mode of the task but the type of syscall. There is actually a general > trend to use this helper more rather than less, and I think the only > way forward here is to ensure that this returns true when entering > through the new syscall number. > > For x86, this has another complication, as some ioctls also need to > check whether they are in an ia32 task (with packed u64 and 32-bit > __kernel_old_time_t) or an x32 task (with aligned u64 and 64-bit > __kernel_old_time_t). If the new syscall gets wired up on x86 as well, > you'd need to decide which of the two behaviors you want. I can have a follow-up patch that makes this do ni_syscall on x86_64 since we can go through the int 0x80 handler, or x32 handler path and choose whichever one there. > > > 3a) Userspace consumes all VA space above 32bit. Forcing allocations to > > occur in lower 32bits > > - This is the current implementation > > 3b) Ensure any allocation in the ioctl handles ioctl entrypoint rather > > than just allow generic memory allocations in full VA space > > - This is hard to guarantee > > What kind of allocation do you mean here? Can you give an example of > an ioctl that does this? My concern here would be something like DRM allocating memory and returning a pointer to userspace that ends up in 64bit space. I can see something like `drm_get_unmapped_area` calls in to `current->mm->get_unmapped_area` which I believe only ends up falling down TASK_SIZE checks. Which could potentially return pointers in the 64bit address space range in this case. Theoretically can be resolved either by thieving the full 64bit VA range, or doing something like the Tango layer patches that on syscall entry changes the syscall to a "compat" syscall. compat syscall flag like Tango might be nicer here? > > > 4a) Blocklist any application using ioctls that have different struct > > packing across the boundary > > - Can happen when struct packing of 32bit x86 application goes down > > the aarch64 compat_ioctl path > > - Userspace is a AArch64 process passing 32bit x86 ioctl structures > > through the compat_ioctl path which is typically for AArch32 processes > > - None currently identified > > 4b) Work with upstream kernel and userspace projects to evaluate and fix > > - Identify the problem ioctls > > - Implement a new ioctl with more sane struct packing that matches > > cross-arch > > - Implement new ioctl while maintaining backwards compatibility with > > previous ioctl handler > > - Change upstream project to use the new compatibility ioctl > > - ioctl deprecation will be case by case per device and project > > 4b) Userspace implements a full ioctl emulation layer > > - Parses the full ioctl tree > > - Either passes through ioctls that it doesn't understand or > > transforms ioctls that it knows are trouble > > - Has the downside that it can still run in to edge cases that will > > fail > > - Performance of additional tracking is a concern > > - Prone to failure keeping the kernel ioctl and userspace ioctl > > handling in sync > > - Really want to have it in the kernel space as much as possible > > I think there are only a few ioctls that are affected, and you can > probably get a list from qemu, which emulates them in user space > already. Doing that transformation should not be all that hard > in the end. > > If we want to do this in the kernel, this probably requires changes > to the syscall calling convention. Adding a flag to pick a particular > style of ioctl arguments would work, and this could enable the > case of emulating arm32 ioctls on x86-64 hosts. > > > diff --git a/arch/arm64/incl
Re: [PATCH] Adds a new ioctl32 syscall for backwards compatibility layers
I encountered a similar problem when writing the Tango binary translator (https://www.amanieusystems.com/). Tango allows AArch32 programs to run on AArch64 CPUs that don't support AArch32 (e.g. ThunderX). The technology has been licensed to several customers who are primarily using it to run Android instances on cloud servers. Doing compat system call translation entirely in user doesn't work out for several reasons: - As Ryan stated, it is impractical and error-prone to manually translate all ioctls in user space. It would be much better to reuse the existing compat layer in the kernel. This is even worse if you take out-of-tree drivers (which have their own ioctls) into account. - There are quite a few system calls which create VM mappings: the usual suspects (mmap, etc) but also io_setup and some ioctls (e.g. in the out-of-tree Mali GPU drivers). - Seccomp filters simply don't work. They could be emulated in user space but this is insecure since the translator is not designed to be a secure sandbox. They also don't propagate across execve. - Some syscalls rely on information that is only known to the kernel. For example, a hugetlbfs fd can only be mapped at a huge page boundary, but the translator cannot know this when selecting a virtual address in the low 4GB for the mapping. The solution that we ended up with was to allow AArch64 processes to issue AArch32 syscalls by using a special system call number. This has several effects for the duration of the syscall: - The compat syscall table is used instead of the primary one. - is_compat_task/in_compat_syscall return true. - get_unmapped_area returns addresses below 4G and uses a separate mmap_base. The separate mmap_base is needed to allow 32-bit applications to benefit from address space randomization. Tango still uses normal mmap syscalls for private memory allocations that are not visible to the translated 32-bit process. - KSTK_EIP and KSTK_ESP return the values of x13/x15 instead of pc/sp. This is necessary for correct functioning seccomp filters and SELinux checks respectively. Tango will set x13 and x15 to the correct values when issuing a 32-bit syscall. - System call restart after a signal sets things up so that restart_syscall is executed as a 32-bit syscall after the signal. I feel that a solution along these lines would also solve Ryan's problem since the vast majority of the syscall translation work is from the difference between 32-bit and 64-bit data structures in memory. The remaining few differences in the syscall ABI between AArch32 and x86-32 can be handled in user mode by the translator program. The kernel patch (based on the v5.4 LTS kernel) can be found at https://github.com/Amanieu/linux/tree/tango-v5.4. On Wed, Jan 6, 2021 at 6:49 AM wrote: > > From: Ryan Houdek > > Problem presented: > A backwards compatibility layer that allows running x86-64 and x86 > processes inside of an AArch64 process. > - CPU is emulated > - Syscall interface is mostly passthrough > - Some syscalls require patching or emulation depending on behaviour > - Not viable from the emulator design to use an AArch32 host process > > x86-64 and x86 userspace emulator source: > https://github.com/FEX-Emu/FEX > Usage of ioctl32 is currently in a downstream fork. This will be the > first user of the syscall. > > Cross documentation: > https://github.com/FEX-Emu/FEX/wiki/32Bit-x86-Woes#ioctl---54 > > ioctls are opaque from the emulator perspective and the data wants to be > passed through a syscall as unimpeded as possible. > Sadly due to ioctl struct differences between x86 and x86-64, we need a > syscall that exposes the compatibility ioctl handler to userspace in a > 64bit process. > > This is necessary behaves of the behaviour differences that occur > between an x86 process doing an ioctl and an x86-64 process doing an > ioctl. > > Both of which are captured and passed through the AArch64 ioctl space. > This is implementing a new ioctl32 syscall that allows us to pass 32bit > x86 ioctls through to the kernel with zero or minimal manipulation. > > The only supported hosts where we care about this currently is AArch64 > and x86-64 (For testing purposes). > PPC64LE, MIPS64LE, and RISC-V64 might be interesting to support in the > future; But I don't have any platforms that get anywhere near Cortex-A77 > performance in those architectures. Nor do I have the time to bring up > the emulator on them. > x86-64 can get to the compatibility ioctl through the int $0x80 handler. > > This does not solve the following problems: > 1) compat_alloc_user_space inside ioctl > 2) ioctls that check task mode instead of entry point for behaviour > 3) ioctls allocating memory > 4) struct packing problems between architectures > > Workarounds for the problems presented: > 1a) Do a stack pivot to the lower 32bits from userspace > - Forces host 64bit process to have its thread stacks to live in 32bit > space. Not ideal. > - Only do a stack pivot on ioctl to save pre
Re: [PATCH] Adds a new ioctl32 syscall for backwards compatibility layers
On Wed, Jan 6, 2021 at 7:48 AM wrote: > From: Ryan Houdek ... > This does not solve the following problems: > 1) compat_alloc_user_space inside ioctl > 2) ioctls that check task mode instead of entry point for behaviour > 3) ioctls allocating memory > 4) struct packing problems between architectures > > Workarounds for the problems presented: > 1a) Do a stack pivot to the lower 32bits from userspace > - Forces host 64bit process to have its thread stacks to live in 32bit > space. Not ideal. > - Only do a stack pivot on ioctl to save previous 32bit VA space > 1b) Teach kernel that compat_alloc_userspace can return a 64bit pointer > - x86-64 truncates stack from this function > - AArch64 returns the full stack pointer > - Only ~29 users. Validating all of them support a 64bit stack is > trivial? I've almost completed the removal of compat_alloc_user_space(), that should no longer be a concern when the syscall gets added. > 2a) Any application using these can be checked for compatibility in > userspace and put on a block list. > 2b) Fix any ioctls doing broken behaviour based on task mode rather than > ioctl entry point What the ioctls() actually check is 'in_compat_syscall()', which is not the mode of the task but the type of syscall. There is actually a general trend to use this helper more rather than less, and I think the only way forward here is to ensure that this returns true when entering through the new syscall number. For x86, this has another complication, as some ioctls also need to check whether they are in an ia32 task (with packed u64 and 32-bit __kernel_old_time_t) or an x32 task (with aligned u64 and 64-bit __kernel_old_time_t). If the new syscall gets wired up on x86 as well, you'd need to decide which of the two behaviors you want. > 3a) Userspace consumes all VA space above 32bit. Forcing allocations to > occur in lower 32bits > - This is the current implementation > 3b) Ensure any allocation in the ioctl handles ioctl entrypoint rather > than just allow generic memory allocations in full VA space > - This is hard to guarantee What kind of allocation do you mean here? Can you give an example of an ioctl that does this? > 4a) Blocklist any application using ioctls that have different struct > packing across the boundary > - Can happen when struct packing of 32bit x86 application goes down > the aarch64 compat_ioctl path > - Userspace is a AArch64 process passing 32bit x86 ioctl structures > through the compat_ioctl path which is typically for AArch32 processes > - None currently identified > 4b) Work with upstream kernel and userspace projects to evaluate and fix > - Identify the problem ioctls > - Implement a new ioctl with more sane struct packing that matches > cross-arch > - Implement new ioctl while maintaining backwards compatibility with > previous ioctl handler > - Change upstream project to use the new compatibility ioctl > - ioctl deprecation will be case by case per device and project > 4b) Userspace implements a full ioctl emulation layer > - Parses the full ioctl tree > - Either passes through ioctls that it doesn't understand or > transforms ioctls that it knows are trouble > - Has the downside that it can still run in to edge cases that will > fail > - Performance of additional tracking is a concern > - Prone to failure keeping the kernel ioctl and userspace ioctl > handling in sync > - Really want to have it in the kernel space as much as possible I think there are only a few ioctls that are affected, and you can probably get a list from qemu, which emulates them in user space already. Doing that transformation should not be all that hard in the end. If we want to do this in the kernel, this probably requires changes to the syscall calling convention. Adding a flag to pick a particular style of ioctl arguments would work, and this could enable the case of emulating arm32 ioctls on x86-64 hosts. > diff --git a/arch/arm64/include/asm/unistd.h b/arch/arm64/include/asm/unistd.h > index 86a9d7b3eabe..949788f5ba40 100644 > --- a/arch/arm64/include/asm/unistd.h > +++ b/arch/arm64/include/asm/unistd.h > @@ -38,7 +38,7 @@ > #define __ARM_NR_compat_set_tls(__ARM_NR_COMPAT_BASE + 5) > #define __ARM_NR_COMPAT_END(__ARM_NR_COMPAT_BASE + 0x800) > > -#define __NR_compat_syscalls 442 > +#define __NR_compat_syscalls 443 > #endif > > #define __ARCH_WANT_SYS_CLONE > diff --git a/arch/arm64/include/asm/unistd32.h > b/arch/arm64/include/asm/unistd32.h > index cccfbbefbf95..35e3bc83dbdc 100644 > --- a/arch/arm64/include/asm/unistd32.h > +++ b/arch/arm64/include/asm/unistd32.h > @@ -891,6 +891,8 @@ __SYSCALL(__NR_faccessat2, sys_faccessat2) > __SYSCALL(__NR_process_madvise, sys_process_madvise) > #define __NR_epoll_pwait2 441 > __SYSCALL(__NR_epoll_pwait2, compat_sys_epoll_pwait2) > +#define __NR_ioctl32 442 > +__SYSCALL(__NR_ioctl32, compat_sys_ioctl) > I'm not sure why
[PATCH] Adds a new ioctl32 syscall for backwards compatibility layers
From: Ryan Houdek Problem presented: A backwards compatibility layer that allows running x86-64 and x86 processes inside of an AArch64 process. - CPU is emulated - Syscall interface is mostly passthrough - Some syscalls require patching or emulation depending on behaviour - Not viable from the emulator design to use an AArch32 host process x86-64 and x86 userspace emulator source: https://github.com/FEX-Emu/FEX Usage of ioctl32 is currently in a downstream fork. This will be the first user of the syscall. Cross documentation: https://github.com/FEX-Emu/FEX/wiki/32Bit-x86-Woes#ioctl---54 ioctls are opaque from the emulator perspective and the data wants to be passed through a syscall as unimpeded as possible. Sadly due to ioctl struct differences between x86 and x86-64, we need a syscall that exposes the compatibility ioctl handler to userspace in a 64bit process. This is necessary behaves of the behaviour differences that occur between an x86 process doing an ioctl and an x86-64 process doing an ioctl. Both of which are captured and passed through the AArch64 ioctl space. This is implementing a new ioctl32 syscall that allows us to pass 32bit x86 ioctls through to the kernel with zero or minimal manipulation. The only supported hosts where we care about this currently is AArch64 and x86-64 (For testing purposes). PPC64LE, MIPS64LE, and RISC-V64 might be interesting to support in the future; But I don't have any platforms that get anywhere near Cortex-A77 performance in those architectures. Nor do I have the time to bring up the emulator on them. x86-64 can get to the compatibility ioctl through the int $0x80 handler. This does not solve the following problems: 1) compat_alloc_user_space inside ioctl 2) ioctls that check task mode instead of entry point for behaviour 3) ioctls allocating memory 4) struct packing problems between architectures Workarounds for the problems presented: 1a) Do a stack pivot to the lower 32bits from userspace - Forces host 64bit process to have its thread stacks to live in 32bit space. Not ideal. - Only do a stack pivot on ioctl to save previous 32bit VA space 1b) Teach kernel that compat_alloc_userspace can return a 64bit pointer - x86-64 truncates stack from this function - AArch64 returns the full stack pointer - Only ~29 users. Validating all of them support a 64bit stack is trivial? 2a) Any application using these can be checked for compatibility in userspace and put on a block list. 2b) Fix any ioctls doing broken behaviour based on task mode rather than ioctl entry point 3a) Userspace consumes all VA space above 32bit. Forcing allocations to occur in lower 32bits - This is the current implementation 3b) Ensure any allocation in the ioctl handles ioctl entrypoint rather than just allow generic memory allocations in full VA space - This is hard to guarantee 4a) Blocklist any application using ioctls that have different struct packing across the boundary - Can happen when struct packing of 32bit x86 application goes down the aarch64 compat_ioctl path - Userspace is a AArch64 process passing 32bit x86 ioctl structures through the compat_ioctl path which is typically for AArch32 processes - None currently identified 4b) Work with upstream kernel and userspace projects to evaluate and fix - Identify the problem ioctls - Implement a new ioctl with more sane struct packing that matches cross-arch - Implement new ioctl while maintaining backwards compatibility with previous ioctl handler - Change upstream project to use the new compatibility ioctl - ioctl deprecation will be case by case per device and project 4b) Userspace implements a full ioctl emulation layer - Parses the full ioctl tree - Either passes through ioctls that it doesn't understand or transforms ioctls that it knows are trouble - Has the downside that it can still run in to edge cases that will fail - Performance of additional tracking is a concern - Prone to failure keeping the kernel ioctl and userspace ioctl handling in sync - Really want to have it in the kernel space as much as possible Signed-off-by: Ryan Houdek --- arch/arm64/include/asm/unistd.h | 2 +- arch/arm64/include/asm/unistd32.h | 2 ++ fs/ioctl.c | 16 ++-- include/linux/syscalls.h| 2 ++ include/uapi/asm-generic/unistd.h | 9 - kernel/sys_ni.c | 3 +++ tools/include/uapi/asm-generic/unistd.h | 9 - 7 files changed, 38 insertions(+), 5 deletions(-) diff --git a/arch/arm64/include/asm/unistd.h b/arch/arm64/include/asm/unistd.h index 86a9d7b3eabe..949788f5ba40 100644 --- a/arch/arm64/include/asm/unistd.h +++ b/arch/arm64/include/asm/unistd.h @@ -38,7 +38,7 @@ #define __ARM_NR_compat_set_tls(__ARM_NR_COMPAT_BASE + 5) #define __ARM_NR_COMPAT_END(__ARM_NR_COMPAT_BASE + 0x800) -#define __NR_compat_syscalls 44