linux.git/arch/riscv/kernel, branch v6.9

riscv: process: Fix kernel gp leakage

2024-04-04T19:35:05+00:00

childregs represents the registers which are active for the new thread
in user context. For a kernel thread, childregs->gp is never used since
the kernel gp is not touched by switch_to. For a user mode helper, the
gp value can be observed in user space after execve or possibly by other
means.

[From the email thread]

The /* Kernel thread */ comment is somewhat inaccurate in that it is also used
for user_mode_helper threads, which exec a user process, e.g. /sbin/init or
when /proc/sys/kernel/core_pattern is a pipe. Such threads do not have
PF_KTHREAD set and are valid targets for ptrace etc. even before they exec.

childregs is the *user* context during syscall execution and it is observable
from userspace in at least five ways:

1. kernel_execve does not currently clear integer registers, so the starting
   register state for PID 1 and other user processes started by the kernel has
   sp = user stack, gp = kernel __global_pointer$, all other integer registers
   zeroed by the memset in the patch comment.

   This is a bug in its own right, but I'm unwilling to bet that it is the only
   way to exploit the issue addressed by this patch.

2. ptrace(PTRACE_GETREGSET): you can PTRACE_ATTACH to a user_mode_helper thread
   before it execs, but ptrace requires SIGSTOP to be delivered which can only
   happen at user/kernel boundaries.

3. /proc/*/task/*/syscall: this is perfectly happy to read pt_regs for
   user_mode_helpers before the exec completes, but gp is not one of the
   registers it returns.

4. PERF_SAMPLE_REGS_USER: LOCKDOWN_PERF normally prevents access to kernel
   addresses via PERF_SAMPLE_REGS_INTR, but due to this bug kernel addresses
   are also exposed via PERF_SAMPLE_REGS_USER which is permitted under
   LOCKDOWN_PERF. I have not attempted to write exploit code.

5. Much of the tracing infrastructure allows access to user registers. I have
   not attempted to determine which forms of tracing allow access to user
   registers without already allowing access to kernel registers.

Fixes: 7db91e57a0ac ("RISC-V: Task implementation")
Cc: stable@vger.kernel.org
Signed-off-by: Stefan O'Rear 
Reviewed-by: Alexandre Ghiti 
Link: https://lore.kernel.org/r/20240327061258.2370291-1-sorear@fastmail.com
Signed-off-by: Palmer Dabbelt

riscv: Disable preemption when using patch_map()

2024-04-04T19:33:38+00:00

patch_map() uses fixmap mappings to circumvent the non-writability of
the kernel text mapping.

The __set_fixmap() function only flushes the current cpu tlb, it does
not emit an IPI so we must make sure that while we use a fixmap mapping,
the current task is not migrated on another cpu which could miss the
newly introduced fixmap mapping.

So in order to avoid any task migration, disable the preemption.

Reported-by: Andrea Parri 
Closes: https://lore.kernel.org/all/ZcS+GAaM25LXsBOl@andrea/
Reported-by: Andy Chiu 
Closes: https://lore.kernel.org/linux-riscv/CABgGipUMz3Sffu-CkmeUB1dKVwVQ73+7=sgC45-m0AE9RCjOZg@mail.gmail.com/
Fixes: cad539baa48f ("riscv: implement a memset like function for text")
Fixes: 0ff7c3b33127 ("riscv: Use text_mutex instead of patch_lock")
Co-developed-by: Andy Chiu 
Signed-off-by: Andy Chiu 
Signed-off-by: Alexandre Ghiti 
Acked-by: Puranjay Mohan 
Link: https://lore.kernel.org/r/20240326203017.310422-3-alexghiti@rivosinc.com
Signed-off-by: Palmer Dabbelt

riscv: Fix warning by declaring arch_cpu_idle() as noinstr

2024-04-04T19:33:37+00:00

The following warning appears when using ftrace:

[89855.443413] RCU not on for: arch_cpu_idle+0x0/0x1c
[89855.445640] WARNING: CPU: 5 PID: 0 at include/linux/trace_recursion.h:162 arch_ftrace_ops_list_func+0x208/0x228
[89855.445824] Modules linked in: xt_conntrack(E) nft_chain_nat(E) xt_MASQUERADE(E) nf_conntrack_netlink(E) xt_addrtype(E) nft_compat(E) nf_tables(E) nfnetlink(E) br_netfilter(E) cfg80211(E) nls_iso8859_1(E) ofpart(E) redboot(E) cmdlinepart(E) cfi_cmdset_0001(E) virtio_net(E) cfi_probe(E) cfi_util(E) 9pnet_virtio(E) gen_probe(E) net_failover(E) virtio_rng(E) failover(E) 9pnet(E) physmap(E) map_funcs(E) chipreg(E) mtd(E) uio_pdrv_genirq(E) uio(E) dm_multipath(E) scsi_dh_rdac(E) scsi_dh_emc(E) scsi_dh_alua(E) drm(E) efi_pstore(E) backlight(E) ip_tables(E) x_tables(E) raid10(E) raid456(E) async_raid6_recov(E) async_memcpy(E) async_pq(E) async_xor(E) xor(E) async_tx(E) raid6_pq(E) raid1(E) raid0(E) virtio_blk(E)
[89855.451563] CPU: 5 PID: 0 Comm: swapper/5 Tainted: G            E      6.8.0-rc6ubuntu-defconfig #2
[89855.451726] Hardware name: riscv-virtio,qemu (DT)
[89855.451899] epc : arch_ftrace_ops_list_func+0x208/0x228
[89855.452016]  ra : arch_ftrace_ops_list_func+0x208/0x228
[89855.452119] epc : ffffffff8016b216 ra : ffffffff8016b216 sp : ffffaf808090fdb0
[89855.452171]  gp : ffffffff827c7680 tp : ffffaf808089ad40 t0 : ffffffff800c0dd8
[89855.452216]  t1 : 0000000000000001 t2 : 0000000000000000 s0 : ffffaf808090fe30
[89855.452306]  s1 : 0000000000000000 a0 : 0000000000000026 a1 : ffffffff82cd6ac8
[89855.452423]  a2 : ffffffff800458c8 a3 : ffffaf80b1870640 a4 : 0000000000000000
[89855.452646]  a5 : 0000000000000000 a6 : 00000000ffffffff a7 : ffffffffffffffff
[89855.452698]  s2 : ffffffff82766872 s3 : ffffffff80004caa s4 : ffffffff80ebea90
[89855.452743]  s5 : ffffaf808089bd40 s6 : 8000000a00006e00 s7 : 0000000000000008
[89855.452787]  s8 : 0000000000002000 s9 : 0000000080043700 s10: 0000000000000000
[89855.452831]  s11: 0000000000000000 t3 : 0000000000100000 t4 : 0000000000000064
[89855.452874]  t5 : 000000000000000c t6 : ffffaf80b182dbfc
[89855.452929] status: 0000000200000100 badaddr: 0000000000000000 cause: 0000000000000003
[89855.453053] [] arch_ftrace_ops_list_func+0x208/0x228
[89855.453191] [] ftrace_call+0x8/0x22
[89855.453265] [] do_idle+0x24c/0x2ca
[89855.453357] [] return_to_handler+0x0/0x26
[89855.453429] [] smp_callin+0x92/0xb6
[89855.453785] ---[ end trace 0000000000000000 ]---

To fix this, mark arch_cpu_idle() as noinstr, like it is done in commit
a9cbc1b471d2 ("s390/idle: mark arch_cpu_idle() noinstr").

Reported-by: Evgenii Shatokhin 
Closes: https://lore.kernel.org/linux-riscv/51f21b87-ebed-4411-afbc-c00d3dea2bab@yadro.com/
Fixes: cfbc4f81c9d0 ("riscv: Select ARCH_WANTS_NO_INSTR")
Signed-off-by: Alexandre Ghiti 
Reviewed-by: Andy Chiu 
Tested-by: Andy Chiu 
Acked-by: Puranjay Mohan 
Link: https://lore.kernel.org/r/20240326203017.310422-2-alexghiti@rivosinc.com
Signed-off-by: Palmer Dabbelt

riscv: use KERN_INFO in do_trap

2024-04-04T19:12:14+00:00

Print the instruction dump with info instead of emergency level.  The
unhandled signal message is only for informational purpose.

Fixes: b8a03a634129 ("riscv: add userland instruction dump to RISC-V splats")
Signed-off-by: Andreas Schwab 
Reviewed-by: Conor Dooley 
Reviewed-by: Atish Patra 
Reviewed-by: Yunhui Cui 
Link: https://lore.kernel.org/r/mvmy1aegrhm.fsf@suse.de
Signed-off-by: Palmer Dabbelt

riscv: Fix vector state restore in rt_sigreturn()

2024-04-03T23:10:25+00:00

The RISC-V Vector specification states in "Appendix D: Calling
Convention for Vector State" [1] that "Executing a system call causes
all caller-saved vector registers (v0-v31, vl, vtype) and vstart to
become unspecified.". In the RISC-V kernel this is called "discarding
the vstate".

Returning from a signal handler via the rt_sigreturn() syscall, vector
discard is also performed. However, this is not an issue since the
vector state should be restored from the sigcontext, and therefore not
care about the vector discard.

The "live state" is the actual vector register in the running context,
and the "vstate" is the vector state of the task. A dirty live state,
means that the vstate and live state are not in synch.

When vectorized user_from_copy() was introduced, an bug sneaked in at
the restoration code, related to the discard of the live state.

An example when this go wrong:

  1. A userland application is executing vector code
  2. The application receives a signal, and the signal handler is
     entered.
  3. The application returns from the signal handler, using the
     rt_sigreturn() syscall.
  4. The live vector state is discarded upon entering the
     rt_sigreturn(), and the live state is marked as "dirty", indicating
     that the live state need to be synchronized with the current
     vstate.
  5. rt_sigreturn() restores the vstate, except the Vector registers,
     from the sigcontext
  6. rt_sigreturn() restores the Vector registers, from the sigcontext,
     and now the vectorized user_from_copy() is used. The dirty live
     state from the discard is saved to the vstate, making the vstate
     corrupt.
  7. rt_sigreturn() returns to the application, which crashes due to
     corrupted vstate.

Note that the vectorized user_from_copy() is invoked depending on the
value of CONFIG_RISCV_ISA_V_UCOPY_THRESHOLD. Default is 768, which
means that vlen has to be larger than 128b for this bug to trigger.

The fix is simply to mark the live state as non-dirty/clean prior
performing the vstate restore.

Link: https://github.com/riscv/riscv-isa-manual/releases/download/riscv-isa-release-8abdb41-2024-03-26/unpriv-isa-asciidoc.pdf # [1]
Reported-by: Charlie Jenkins 
Reported-by: Vineet Gupta 
Fixes: c2a658d41924 ("riscv: lib: vectorize copy_to_user/copy_from_user")
Signed-off-by: Björn Töpel 
Reviewed-by: Andy Chiu 
Tested-by: Vineet Gupta 
Link: https://lore.kernel.org/r/20240403072638.567446-1-bjorn@kernel.org
Cc: stable@vger.kernel.org
Signed-off-by: Palmer Dabbelt

riscv: compat_vdso: align VDSOAS build log

2024-03-27T14:08:38+00:00

Add one more space after "VDSOAS" for better alignment in the build log.

[Before]

  LDS     arch/riscv/kernel/compat_vdso/compat_vdso.lds
  VDSOAS arch/riscv/kernel/compat_vdso/rt_sigreturn.o
  VDSOAS arch/riscv/kernel/compat_vdso/getcpu.o
  VDSOAS arch/riscv/kernel/compat_vdso/flush_icache.o
  VDSOAS arch/riscv/kernel/compat_vdso/note.o
  VDSOLD  arch/riscv/kernel/compat_vdso/compat_vdso.so.dbg
  VDSOSYM include/generated/compat_vdso-offsets.h

[After]

  LDS     arch/riscv/kernel/compat_vdso/compat_vdso.lds
  VDSOAS  arch/riscv/kernel/compat_vdso/rt_sigreturn.o
  VDSOAS  arch/riscv/kernel/compat_vdso/getcpu.o
  VDSOAS  arch/riscv/kernel/compat_vdso/flush_icache.o
  VDSOAS  arch/riscv/kernel/compat_vdso/note.o
  VDSOLD  arch/riscv/kernel/compat_vdso/compat_vdso.so.dbg
  VDSOSYM include/generated/compat_vdso-offsets.h

Signed-off-by: Masahiro Yamada 
Reviewed-by: Alexandre Ghiti 
Link: https://lore.kernel.org/r/20231117125843.1058553-1-masahiroy@kernel.org
Signed-off-by: Palmer Dabbelt

riscv: hwprobe: do not produce frtace relocation

2024-03-26T21:06:59+00:00

Such relocation causes crash of android linker similar to one
described in commit e05d57dcb8c7
("riscv: Fixup __vdso_gettimeofday broke dynamic ftrace").

Looks like this relocation is added by CONFIG_DYNAMIC_FTRACE which is
disabled in the default android kernel.

Before:

readelf -rW arch/riscv/kernel/vdso/vdso.so:

Relocation section '.rela.dyn' at offset 0xd00 contains 1 entry:
    Offset             Info             Type
0000000000000d20  0000000000000003 R_RISCV_RELATIVE

objdump:
0000000000000c86 <__vdso_riscv_hwprobe@@LINUX_4.15>:
 c86:   0001                    nop
 c88:   0001                    nop
 c8a:   0001                    nop
 c8c:   0001                    nop
 c8e:   e211                    bnez    a2,c92 <__vdso_riscv_hwprobe...

After:
readelf -rW arch/riscv/kernel/vdso/vdso.so:

There are no relocations in this file.

objdump:
0000000000000c86 <__vdso_riscv_hwprobe@@LINUX_4.15>:
 c86:   e211                    bnez    a2,c8a <__vdso_riscv_hwprobe...
 c88:   c6b9                    beqz    a3,cd6 <__vdso_riscv_hwprobe...
 c8a:   e739                    bnez    a4,cd8 <__vdso_riscv_hwprobe...
 c8c:   ffffd797                auipc   a5,0xffffd

Also disable SCS since it also should not be available in vdso.

Fixes: aa5af0aa90ba ("RISC-V: Add hwprobe vDSO function and data")
Signed-off-by: Roman Artemev 
Signed-off-by: Vladimir Isaev 
Reviewed-by: Alexandre Ghiti 
Reviewed-by: Guo Ren 
Link: https://lore.kernel.org/r/20240313085843.17661-1-vladimir.isaev@syntacore.com
Signed-off-by: Palmer Dabbelt

Merge tag 'riscv-for-linus-6.9-mw2' of git://git.kernel.org/pub/scm/linux/kernel/git/riscv/linux

2024-03-22T17:41:13+00:00

Pull RISC-V updates from Palmer Dabbelt:

 - Support for various vector-accelerated crypto routines

 - Hibernation is now enabled for portable kernel builds

 - mmap_rnd_bits_max is larger on systems with larger VAs

 - Support for fast GUP

 - Support for membarrier-based instruction cache synchronization

 - Support for the Andes hart-level interrupt controller and PMU

 - Some cleanups around unaligned access speed probing and Kconfig
   settings

 - Support for ACPI LPI and CPPC

 - Various cleanus related to barriers

 - A handful of fixes

* tag 'riscv-for-linus-6.9-mw2' of git://git.kernel.org/pub/scm/linux/kernel/git/riscv/linux: (66 commits)
  riscv: Fix syscall wrapper for >word-size arguments
  crypto: riscv - add vector crypto accelerated AES-CBC-CTS
  crypto: riscv - parallelize AES-CBC decryption
  riscv: Only flush the mm icache when setting an exec pte
  riscv: Use kcalloc() instead of kzalloc()
  riscv/barrier: Add missing space after ','
  riscv/barrier: Consolidate fence definitions
  riscv/barrier: Define RISCV_FULL_BARRIER
  riscv/barrier: Define __{mb,rmb,wmb}
  RISC-V: defconfig: Enable CONFIG_ACPI_CPPC_CPUFREQ
  cpufreq: Move CPPC configs to common Kconfig and add RISC-V
  ACPI: RISC-V: Add CPPC driver
  ACPI: Enable ACPI_PROCESSOR for RISC-V
  ACPI: RISC-V: Add LPI driver
  cpuidle: RISC-V: Move few functions to arch/riscv
  riscv: Introduce set_compat_task() in asm/compat.h
  riscv: Introduce is_compat_thread() into compat.h
  riscv: add compile-time test into is_compat_task()
  riscv: Replace direct thread flag check with is_compat_task()
  riscv: Improve arch_get_mmap_end() macro
  ...

Merge tag 'kbuild-v6.9' of git://git.kernel.org/pub/scm/linux/kernel/git/masahiroy/linux-kbuild

2024-03-21T21:41:00+00:00

Pull Kbuild updates from Masahiro Yamada:

 - Generate a list of built DTB files (arch/*/boot/dts/dtbs-list)

 - Use more threads when building Debian packages in parallel

 - Fix warnings shown during the RPM kernel package uninstallation

 - Change OBJECT_FILES_NON_STANDARD_*.o etc. to take a relative path to
   Makefile

 - Support GCC's -fmin-function-alignment flag

 - Fix a null pointer dereference bug in modpost

 - Add the DTB support to the RPM package

 - Various fixes and cleanups in Kconfig

* tag 'kbuild-v6.9' of git://git.kernel.org/pub/scm/linux/kernel/git/masahiroy/linux-kbuild: (67 commits)
  kconfig: tests: test dependency after shuffling choices
  kconfig: tests: add a test for randconfig with dependent choices
  kconfig: tests: support KCONFIG_SEED for the randconfig runner
  kbuild: rpm-pkg: add dtb files in kernel rpm
  kconfig: remove unneeded menu_is_visible() call in conf_write_defconfig()
  kconfig: check prompt for choice while parsing
  kconfig: lxdialog: remove unused dialog colors
  kconfig: lxdialog: fix button color for blackbg theme
  modpost: fix null pointer dereference
  kbuild: remove GCC's default -Wpacked-bitfield-compat flag
  kbuild: unexport abs_srctree and abs_objtree
  kbuild: Move -Wenum-{compare-conditional,enum-conversion} into W=1
  kconfig: remove named choice support
  kconfig: use linked list in get_symbol_str() to iterate over menus
  kconfig: link menus to a symbol
  kbuild: fix inconsistent indentation in top Makefile
  kbuild: Use -fmin-function-alignment when available
  alpha: merge two entries for CONFIG_ALPHA_GAMMA
  alpha: merge two entries for CONFIG_ALPHA_EV4
  kbuild: change DTC_FLAGS_.o to take the path relative to $(obj)
  ...

riscv: Use kcalloc() instead of kzalloc()

2024-03-20T15:56:07+00:00

As noted in the "Deprecated Interfaces, Language Features, Attributes,
and Conventions" documentation [1], size calculations (especially
multiplication) should not be performed in memory allocator (or similar)
function arguments due to the risk of them overflowing. This could lead
to values wrapping around and a smaller allocation being made than the
caller was expecting. Using those allocations could lead to linear
overflows of heap memory and other misbehaviors.

So, use the purpose specific kcalloc() function instead of the argument
count * size in the kzalloc() function.

Also, it is preferred to use sizeof(*pointer) instead of sizeof(type)
due to the type of the variable can change and one needs not change the
former (unlike the latter).

Link: https://www.kernel.org/doc/html/next/process/deprecated.html#open-coded-arithmetic-in-allocator-arguments [1]
Link: https://github.com/KSPP/linux/issues/162
Signed-off-by: Erick Archer 
Reviewed-by: Alexandre Ghiti 
Reviewed-by: Andrew Jones 
Reviewed-by: Gustavo A. R. Silva 
Link: https://lore.kernel.org/r/20240120135400.4710-1-erick.archer@gmx.com
Signed-off-by: Palmer Dabbelt