[PATCH AUTOSEL 5.4 10/58] mips: vdso: fix 'jalr t9' crash in vdso code

[PATCH AUTOSEL 5.4 10/58] mips: vdso: fix 'jalr t9' crash in vdso code

[Sasha Levin]

From: Victor Kamensky <kamensky@cisco.com>

[ Upstream commit d3f703c4359ff06619b2322b91f69710453e6b6d ]

Observed that when kernel is built with Yocto mips64-poky-linux-gcc,
and mips64-poky-linux-gnun32-gcc toolchain, resulting vdso contains
'jalr t9' instructions in its code and since in vdso case nobody
sets GOT table code crashes when instruction reached. On other hand
observed that when kernel is built mips-poky-linux-gcc toolchain, the
same 'jalr t9' instruction are replaced with PC relative function
calls using 'bal' instructions.

The difference boils down to -mrelax-pic-calls and -mexplicit-relocs
gcc options that gets different default values depending on gcc
target triplets and corresponding binutils. -mrelax-pic-calls got
enabled by default only in mips-poky-linux-gcc case. MIPS binutils
ld relies on R_MIPS_JALR relocation to convert 'jalr t9' into 'bal'
and such relocation is generated only if -mrelax-pic-calls option
is on.

Please note 'jalr t9' conversion to 'bal' can happen only to static
functions. These static PIC calls use mips local GOT entries that
are supposed to be filled with start of DSO value by run-time linker
(missing in VDSO case) and they do not have dynamic relocations.
Global mips GOT entries must have dynamic relocations and they should
be prevented by cmd_vdso_check Makefile rule.

Solution call out -mrelax-pic-calls and -mexplicit-relocs options
explicitly while compiling MIPS vdso code. That would get correct
and consistent between different toolchains behaviour.

Reported-by: Bruce Ashfield <bruce.ashfield@gmail.com>
Signed-off-by: Victor Kamensky <kamensky@cisco.com>
Signed-off-by: Paul Burton <paulburton@kernel.org>
Cc: linux-mips@vger.kernel.org
Cc: Ralf Baechle <ralf@linux-mips.org>
Cc: James Hogan <jhogan@kernel.org>
Cc: Vincenzo Frascino <vincenzo.frascino@arm.com>
Cc: richard.purdie@linuxfoundation.org
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 arch/mips/vdso/Makefile | 1 +
 1 file changed, 1 insertion(+)

diff --git a/arch/mips/vdso/Makefile b/arch/mips/vdso/Makefile
index 996a934ece7d6..3fa4bbe1bae53 100644
--- a/arch/mips/vdso/Makefile
+++ b/arch/mips/vdso/Makefile
@@ -29,6 +29,7 @@ endif
 cflags-vdso := $(ccflags-vdso) \
 	$(filter -W%,$(filter-out -Wa$(comma)%,$(KBUILD_CFLAGS))) \
 	-O3 -g -fPIC -fno-strict-aliasing -fno-common -fno-builtin -G 0 \
+	-mrelax-pic-calls -mexplicit-relocs \
 	-fno-stack-protector -fno-jump-tables -DDISABLE_BRANCH_PROFILING \
 	$(call cc-option, -fno-asynchronous-unwind-tables) \
 	$(call cc-option, -fno-stack-protector)
-- 
2.20.1

[PATCH AUTOSEL 5.4 11/58] MIPS: Disable VDSO time functionality on microMIPS

[Sasha Levin]

From: Paul Burton <paulburton@kernel.org>

[ Upstream commit 07015d7a103c4420b69a287b8ef4d2535c0f4106 ]

A check we're about to add to pick up on function calls that depend on
bogus use of the GOT in the VDSO picked up on instances of such function
calls in microMIPS builds. Since the code appears genuinely problematic,
and given the relatively small amount of use & testing that microMIPS
sees, go ahead & disable the VDSO for microMIPS builds.

Signed-off-by: Paul Burton <paulburton@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 arch/mips/vdso/Makefile | 19 +++++++++++++++++--
 1 file changed, 17 insertions(+), 2 deletions(-)

diff --git a/arch/mips/vdso/Makefile b/arch/mips/vdso/Makefile
index 3fa4bbe1bae53..b6b1eb638fb14 100644
--- a/arch/mips/vdso/Makefile
+++ b/arch/mips/vdso/Makefile
@@ -48,6 +48,8 @@ endif
 
 CFLAGS_REMOVE_vgettimeofday.o = -pg
 
+DISABLE_VDSO := n
+
 #
 # For the pre-R6 code in arch/mips/vdso/vdso.h for locating
 # the base address of VDSO, the linker will emit a R_MIPS_PC32
@@ -61,11 +63,24 @@ CFLAGS_REMOVE_vgettimeofday.o = -pg
 ifndef CONFIG_CPU_MIPSR6
   ifeq ($(call ld-ifversion, -lt, 225000000, y),y)
     $(warning MIPS VDSO requires binutils >= 2.25)
-    obj-vdso-y := $(filter-out vgettimeofday.o, $(obj-vdso-y))
-    ccflags-vdso += -DDISABLE_MIPS_VDSO
+    DISABLE_VDSO := y
   endif
 endif
 
+#
+# GCC (at least up to version 9.2) appears to emit function calls that make use
+# of the GOT when targeting microMIPS, which we can't use in the VDSO due to
+# the lack of relocations. As such, we disable the VDSO for microMIPS builds.
+#
+ifdef CONFIG_CPU_MICROMIPS
+  DISABLE_VDSO := y
+endif
+
+ifeq ($(DISABLE_VDSO),y)
+  obj-vdso-y := $(filter-out vgettimeofday.o, $(obj-vdso-y))
+  ccflags-vdso += -DDISABLE_MIPS_VDSO
+endif
+
 # VDSO linker flags.
 VDSO_LDFLAGS := \
 	-Wl,-Bsymbolic -Wl,--no-undefined -Wl,-soname=linux-vdso.so.1 \
-- 
2.20.1

[PATCH AUTOSEL 5.4 12/58] mips: vdso: add build time check that no 'jalr t9' calls left

[Sasha Levin]

From: Victor Kamensky <kamensky@cisco.com>

[ Upstream commit 976c23af3ee5bd3447a7bfb6c356ceb4acf264a6 ]

vdso shared object cannot have GOT based PIC 'jalr t9' calls
because nobody set GOT table in vdso. Contributing into vdso
.o files are compiled in PIC mode and as result for internal
static functions calls compiler will generate 'jalr t9'
instructions. Those are supposed to be converted into PC
relative 'bal' calls by linker when relocation are processed.

Mips global GOT entries do have dynamic relocations and they
will be caught by cmd_vdso_check Makefile rule. Static PIC
calls go through mips local GOT entries that do not have
dynamic relocations. For those 'jalr t9' calls could be present
but without dynamic relocations and they need to be converted
to 'bal' calls by linker.

Add additional build time check to make sure that no 'jalr t9'
slip through because of some toolchain misconfiguration that
prevents 'jalr t9' to 'bal' conversion.

Signed-off-by: Victor Kamensky <kamensky@cisco.com>
Signed-off-by: Paul Burton <paulburton@kernel.org>
Cc: linux-mips@vger.kernel.org
Cc: Ralf Baechle <ralf@linux-mips.org>
Cc: James Hogan <jhogan@kernel.org>
Cc: Vincenzo Frascino <vincenzo.frascino@arm.com>
Cc: bruce.ashfield@gmail.com
Cc: richard.purdie@linuxfoundation.org
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 arch/mips/vdso/Makefile | 8 +++++++-
 1 file changed, 7 insertions(+), 1 deletion(-)

diff --git a/arch/mips/vdso/Makefile b/arch/mips/vdso/Makefile
index b6b1eb638fb14..08c835d48520b 100644
--- a/arch/mips/vdso/Makefile
+++ b/arch/mips/vdso/Makefile
@@ -92,12 +92,18 @@ CFLAGS_REMOVE_vdso.o = -pg
 GCOV_PROFILE := n
 UBSAN_SANITIZE := n
 
+# Check that we don't have PIC 'jalr t9' calls left
+quiet_cmd_vdso_mips_check = VDSOCHK $@
+      cmd_vdso_mips_check = if $(OBJDUMP) --disassemble $@ | egrep -h "jalr.*t9" > /dev/null; \
+		       then (echo >&2 "$@: PIC 'jalr t9' calls are not supported"; \
+			     rm -f $@; /bin/false); fi
+
 #
 # Shared build commands.
 #
 
 quiet_cmd_vdsold_and_vdso_check = LD      $@
-      cmd_vdsold_and_vdso_check = $(cmd_vdsold); $(cmd_vdso_check)
+      cmd_vdsold_and_vdso_check = $(cmd_vdsold); $(cmd_vdso_check); $(cmd_vdso_mips_check)
 
 quiet_cmd_vdsold = VDSO    $@
       cmd_vdsold = $(CC) $(c_flags) $(VDSO_LDFLAGS) \
-- 
2.20.1

[PATCH AUTOSEL 5.4 13/58] MIPS: VPE: Fix a double free and a memory leak in 'release_vpe()'

[Sasha Levin]

From: Christophe JAILLET <christophe.jaillet@wanadoo.fr>

[ Upstream commit bef8e2dfceed6daeb6ca3e8d33f9c9d43b926580 ]

Pointer on the memory allocated by 'alloc_progmem()' is stored in
'v->load_addr'. So this is this memory that should be freed by
'release_progmem()'.

'release_progmem()' is only a call to 'kfree()'.

With the current code, there is both a double free and a memory leak.
Fix it by passing the correct pointer to 'release_progmem()'.

Fixes: e01402b115ccc ("More AP / SP bits for the 34K, the Malta bits and things. Still wants")
Signed-off-by: Christophe JAILLET <christophe.jaillet@wanadoo.fr>
Signed-off-by: Paul Burton <paulburton@kernel.org>
Cc: ralf@linux-mips.org
Cc: linux-mips@vger.kernel.org
Cc: linux-kernel@vger.kernel.org
Cc: kernel-janitors@vger.kernel.org
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 arch/mips/kernel/vpe.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/arch/mips/kernel/vpe.c b/arch/mips/kernel/vpe.c
index 6176b9acba950..d0d832ab3d3b8 100644
--- a/arch/mips/kernel/vpe.c
+++ b/arch/mips/kernel/vpe.c
@@ -134,7 +134,7 @@ void release_vpe(struct vpe *v)
 {
 	list_del(&v->list);
 	if (v->load_addr)
-		release_progmem(v);
+		release_progmem(v->load_addr);
 	kfree(v);
 }
 
-- 
2.20.1

[PATCH AUTOSEL 5.4 17/58] MIPS: vdso: Wrap -mexplicit-relocs in cc-option

[Sasha Levin]

From: Nathan Chancellor <natechancellor@gmail.com>

[ Upstream commit 72cf3b3df423c1bbd8fa1056fed009d3a260f8a9 ]

Clang does not support this option and errors out:

clang-11: error: unknown argument: '-mexplicit-relocs'

Clang does not appear to need this flag like GCC does because the jalr
check that was added in commit 976c23af3ee5 ("mips: vdso: add build
time check that no 'jalr t9' calls left") passes just fine with

$ make ARCH=mips CC=clang CROSS_COMPILE=mipsel-linux-gnu- malta_defconfig arch/mips/vdso/

even before commit d3f703c4359f ("mips: vdso: fix 'jalr t9' crash in
vdso code").

-mrelax-pic-calls has been supported since clang 9, which is the
earliest version that could build a working MIPS kernel, and it is the
default for clang so just leave it be.

Fixes: d3f703c4359f ("mips: vdso: fix 'jalr t9' crash in vdso code")
Link: https://github.com/ClangBuiltLinux/linux/issues/890
Signed-off-by: Nathan Chancellor <natechancellor@gmail.com>
Reviewed-by: Nick Desaulniers <ndesaulniers@google.com>
Tested-by: Nick Desaulniers <ndesaulniers@google.com>
Signed-off-by: Paul Burton <paulburton@kernel.org>
Cc: Ralf Baechle <ralf@linux-mips.org>
Cc: linux-mips@vger.kernel.org
Cc: linux-kernel@vger.kernel.org
Cc: clang-built-linux@googlegroups.com
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 arch/mips/vdso/Makefile | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/arch/mips/vdso/Makefile b/arch/mips/vdso/Makefile
index 08c835d48520b..3dcfdee678fb9 100644
--- a/arch/mips/vdso/Makefile
+++ b/arch/mips/vdso/Makefile
@@ -29,7 +29,7 @@ endif
 cflags-vdso := $(ccflags-vdso) \
 	$(filter -W%,$(filter-out -Wa$(comma)%,$(KBUILD_CFLAGS))) \
 	-O3 -g -fPIC -fno-strict-aliasing -fno-common -fno-builtin -G 0 \
-	-mrelax-pic-calls -mexplicit-relocs \
+	-mrelax-pic-calls $(call cc-option, -mexplicit-relocs) \
 	-fno-stack-protector -fno-jump-tables -DDISABLE_BRANCH_PROFILING \
 	$(call cc-option, -fno-asynchronous-unwind-tables) \
 	$(call cc-option, -fno-stack-protector)
-- 
2.20.1

Re: [PATCH AUTOSEL 5.4 10/58] mips: vdso: fix 'jalr t9' crash in vdso code

[Hauke Mehrtens]

[-- Attachment #1.1: Type: text/plain, Size: 2682 bytes --]

On 3/5/20 6:13 PM, Sasha Levin wrote:
> From: Victor Kamensky <kamensky@cisco.com>
> 
> [ Upstream commit d3f703c4359ff06619b2322b91f69710453e6b6d ]
> 
> Observed that when kernel is built with Yocto mips64-poky-linux-gcc,
> and mips64-poky-linux-gnun32-gcc toolchain, resulting vdso contains
> 'jalr t9' instructions in its code and since in vdso case nobody
> sets GOT table code crashes when instruction reached. On other hand
> observed that when kernel is built mips-poky-linux-gcc toolchain, the
> same 'jalr t9' instruction are replaced with PC relative function
> calls using 'bal' instructions.
> 
> The difference boils down to -mrelax-pic-calls and -mexplicit-relocs
> gcc options that gets different default values depending on gcc
> target triplets and corresponding binutils. -mrelax-pic-calls got
> enabled by default only in mips-poky-linux-gcc case. MIPS binutils
> ld relies on R_MIPS_JALR relocation to convert 'jalr t9' into 'bal'
> and such relocation is generated only if -mrelax-pic-calls option
> is on.
> 
> Please note 'jalr t9' conversion to 'bal' can happen only to static
> functions. These static PIC calls use mips local GOT entries that
> are supposed to be filled with start of DSO value by run-time linker
> (missing in VDSO case) and they do not have dynamic relocations.
> Global mips GOT entries must have dynamic relocations and they should
> be prevented by cmd_vdso_check Makefile rule.
> 
> Solution call out -mrelax-pic-calls and -mexplicit-relocs options
> explicitly while compiling MIPS vdso code. That would get correct
> and consistent between different toolchains behaviour.
> 
> Reported-by: Bruce Ashfield <bruce.ashfield@gmail.com>
> Signed-off-by: Victor Kamensky <kamensky@cisco.com>
> Signed-off-by: Paul Burton <paulburton@kernel.org>
> Cc: linux-mips@vger.kernel.org
> Cc: Ralf Baechle <ralf@linux-mips.org>
> Cc: James Hogan <jhogan@kernel.org>
> Cc: Vincenzo Frascino <vincenzo.frascino@arm.com>
> Cc: richard.purdie@linuxfoundation.org
> Signed-off-by: Sasha Levin <sashal@kernel.org>
> ---
>  arch/mips/vdso/Makefile | 1 +
>  1 file changed, 1 insertion(+)
> 

Hi Sasha,

Why was this not added to the 5.4 stable branch?

Some OpenWrt users ran into this problem with kernel 5.4 on MIPS64 [0].
We backported this patch on our own in OpenWrt [1], but it should be
added to the sable branch in my opinion as it fixes a real problem.

@Sasha: Can you add it to the 5.4 stable branch or should I send some
special email?

[0]: https://bugs.openwrt.org/index.php?do=details&task_id=3277
[1]: https://git.openwrt.org/2932b4d05e1d212259c2903fd9cebcee5616680b

Hauke


[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 488 bytes --]

Re: [PATCH AUTOSEL 5.4 10/58] mips: vdso: fix 'jalr t9' crash in vdso code

[Sasha Levin]

On Sat, Aug 29, 2020 at 03:08:01PM +0200, Hauke Mehrtens wrote:
>On 3/5/20 6:13 PM, Sasha Levin wrote:
>> From: Victor Kamensky <kamensky@cisco.com>
>>
>> [ Upstream commit d3f703c4359ff06619b2322b91f69710453e6b6d ]
>>
>> Observed that when kernel is built with Yocto mips64-poky-linux-gcc,
>> and mips64-poky-linux-gnun32-gcc toolchain, resulting vdso contains
>> 'jalr t9' instructions in its code and since in vdso case nobody
>> sets GOT table code crashes when instruction reached. On other hand
>> observed that when kernel is built mips-poky-linux-gcc toolchain, the
>> same 'jalr t9' instruction are replaced with PC relative function
>> calls using 'bal' instructions.
>>
>> The difference boils down to -mrelax-pic-calls and -mexplicit-relocs
>> gcc options that gets different default values depending on gcc
>> target triplets and corresponding binutils. -mrelax-pic-calls got
>> enabled by default only in mips-poky-linux-gcc case. MIPS binutils
>> ld relies on R_MIPS_JALR relocation to convert 'jalr t9' into 'bal'
>> and such relocation is generated only if -mrelax-pic-calls option
>> is on.
>>
>> Please note 'jalr t9' conversion to 'bal' can happen only to static
>> functions. These static PIC calls use mips local GOT entries that
>> are supposed to be filled with start of DSO value by run-time linker
>> (missing in VDSO case) and they do not have dynamic relocations.
>> Global mips GOT entries must have dynamic relocations and they should
>> be prevented by cmd_vdso_check Makefile rule.
>>
>> Solution call out -mrelax-pic-calls and -mexplicit-relocs options
>> explicitly while compiling MIPS vdso code. That would get correct
>> and consistent between different toolchains behaviour.
>>
>> Reported-by: Bruce Ashfield <bruce.ashfield@gmail.com>
>> Signed-off-by: Victor Kamensky <kamensky@cisco.com>
>> Signed-off-by: Paul Burton <paulburton@kernel.org>
>> Cc: linux-mips@vger.kernel.org
>> Cc: Ralf Baechle <ralf@linux-mips.org>
>> Cc: James Hogan <jhogan@kernel.org>
>> Cc: Vincenzo Frascino <vincenzo.frascino@arm.com>
>> Cc: richard.purdie@linuxfoundation.org
>> Signed-off-by: Sasha Levin <sashal@kernel.org>
>> ---
>>  arch/mips/vdso/Makefile | 1 +
>>  1 file changed, 1 insertion(+)
>>
>
>Hi Sasha,
>
>Why was this not added to the 5.4 stable branch?
>
>Some OpenWrt users ran into this problem with kernel 5.4 on MIPS64 [0].
>We backported this patch on our own in OpenWrt [1], but it should be
>added to the sable branch in my opinion as it fixes a real problem.
>
>@Sasha: Can you add it to the 5.4 stable branch or should I send some
>special email?

It failed building on 5.4. If you'd like it included, please send me a
tested backport for 5.4.

-- 
Thanks,
Sasha

Re: [PATCH AUTOSEL 5.4 10/58] mips: vdso: fix 'jalr t9' crash in vdso code

[Hauke Mehrtens]

[-- Attachment #1.1: Type: text/plain, Size: 3587 bytes --]

On 8/29/20 3:56 PM, Sasha Levin wrote:
> On Sat, Aug 29, 2020 at 03:08:01PM +0200, Hauke Mehrtens wrote:
>> On 3/5/20 6:13 PM, Sasha Levin wrote:
>>> From: Victor Kamensky <kamensky@cisco.com>
>>>
>>> [ Upstream commit d3f703c4359ff06619b2322b91f69710453e6b6d ]
>>>
>>> Observed that when kernel is built with Yocto mips64-poky-linux-gcc,
>>> and mips64-poky-linux-gnun32-gcc toolchain, resulting vdso contains
>>> 'jalr t9' instructions in its code and since in vdso case nobody
>>> sets GOT table code crashes when instruction reached. On other hand
>>> observed that when kernel is built mips-poky-linux-gcc toolchain, the
>>> same 'jalr t9' instruction are replaced with PC relative function
>>> calls using 'bal' instructions.
>>>
>>> The difference boils down to -mrelax-pic-calls and -mexplicit-relocs
>>> gcc options that gets different default values depending on gcc
>>> target triplets and corresponding binutils. -mrelax-pic-calls got
>>> enabled by default only in mips-poky-linux-gcc case. MIPS binutils
>>> ld relies on R_MIPS_JALR relocation to convert 'jalr t9' into 'bal'
>>> and such relocation is generated only if -mrelax-pic-calls option
>>> is on.
>>>
>>> Please note 'jalr t9' conversion to 'bal' can happen only to static
>>> functions. These static PIC calls use mips local GOT entries that
>>> are supposed to be filled with start of DSO value by run-time linker
>>> (missing in VDSO case) and they do not have dynamic relocations.
>>> Global mips GOT entries must have dynamic relocations and they should
>>> be prevented by cmd_vdso_check Makefile rule.
>>>
>>> Solution call out -mrelax-pic-calls and -mexplicit-relocs options
>>> explicitly while compiling MIPS vdso code. That would get correct
>>> and consistent between different toolchains behaviour.
>>>
>>> Reported-by: Bruce Ashfield <bruce.ashfield@gmail.com>
>>> Signed-off-by: Victor Kamensky <kamensky@cisco.com>
>>> Signed-off-by: Paul Burton <paulburton@kernel.org>
>>> Cc: linux-mips@vger.kernel.org
>>> Cc: Ralf Baechle <ralf@linux-mips.org>
>>> Cc: James Hogan <jhogan@kernel.org>
>>> Cc: Vincenzo Frascino <vincenzo.frascino@arm.com>
>>> Cc: richard.purdie@linuxfoundation.org
>>> Signed-off-by: Sasha Levin <sashal@kernel.org>
>>> ---
>>>  arch/mips/vdso/Makefile | 1 +
>>>  1 file changed, 1 insertion(+)
>>>
>>
>> Hi Sasha,
>>
>> Why was this not added to the 5.4 stable branch?
>>
>> Some OpenWrt users ran into this problem with kernel 5.4 on MIPS64 [0].
>> We backported this patch on our own in OpenWrt [1], but it should be
>> added to the sable branch in my opinion as it fixes a real problem.
>>
>> @Sasha: Can you add it to the 5.4 stable branch or should I send some
>> special email?
> 
> It failed building on 5.4. If you'd like it included, please send me a
> tested backport for 5.4.
> 

I successfully compiled a kernel 5.4.61 with this patch on top with GCC
8.4 for MIPS 64 big and little Endian.

What was broken in your compile test?

Someone complained that it does not compile with LLVM any more, there is
this additional fix to solve that problem:
https://git.kernel.org/linus/72cf3b3df423c1bbd8fa1056fed009d3a260f8a9
OpenWrt does not support LLVM, so we did not see this problem.

Could you please backport these two commits to kernel 5.4:
https://git.kernel.org/linus/d3f703c4359ff06619b2322b91f69710453e6b6d
https://git.kernel.org/linus/72cf3b3df423c1bbd8fa1056fed009d3a260f8a9

We haven't seen this problem with kernel 4.19 or older, but I am not
sure if this was luck.

Hauke


[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 488 bytes --]

Re: [PATCH AUTOSEL 5.4 10/58] mips: vdso: fix 'jalr t9' crash in vdso code

[Sasha Levin]

On Sat, Aug 29, 2020 at 04:37:32PM +0200, Hauke Mehrtens wrote:
>On 8/29/20 3:56 PM, Sasha Levin wrote:
>> On Sat, Aug 29, 2020 at 03:08:01PM +0200, Hauke Mehrtens wrote:
>>> On 3/5/20 6:13 PM, Sasha Levin wrote:
>>>> From: Victor Kamensky <kamensky@cisco.com>
>>>>
>>>> [ Upstream commit d3f703c4359ff06619b2322b91f69710453e6b6d ]
>>>>
>>>> Observed that when kernel is built with Yocto mips64-poky-linux-gcc,
>>>> and mips64-poky-linux-gnun32-gcc toolchain, resulting vdso contains
>>>> 'jalr t9' instructions in its code and since in vdso case nobody
>>>> sets GOT table code crashes when instruction reached. On other hand
>>>> observed that when kernel is built mips-poky-linux-gcc toolchain, the
>>>> same 'jalr t9' instruction are replaced with PC relative function
>>>> calls using 'bal' instructions.
>>>>
>>>> The difference boils down to -mrelax-pic-calls and -mexplicit-relocs
>>>> gcc options that gets different default values depending on gcc
>>>> target triplets and corresponding binutils. -mrelax-pic-calls got
>>>> enabled by default only in mips-poky-linux-gcc case. MIPS binutils
>>>> ld relies on R_MIPS_JALR relocation to convert 'jalr t9' into 'bal'
>>>> and such relocation is generated only if -mrelax-pic-calls option
>>>> is on.
>>>>
>>>> Please note 'jalr t9' conversion to 'bal' can happen only to static
>>>> functions. These static PIC calls use mips local GOT entries that
>>>> are supposed to be filled with start of DSO value by run-time linker
>>>> (missing in VDSO case) and they do not have dynamic relocations.
>>>> Global mips GOT entries must have dynamic relocations and they should
>>>> be prevented by cmd_vdso_check Makefile rule.
>>>>
>>>> Solution call out -mrelax-pic-calls and -mexplicit-relocs options
>>>> explicitly while compiling MIPS vdso code. That would get correct
>>>> and consistent between different toolchains behaviour.
>>>>
>>>> Reported-by: Bruce Ashfield <bruce.ashfield@gmail.com>
>>>> Signed-off-by: Victor Kamensky <kamensky@cisco.com>
>>>> Signed-off-by: Paul Burton <paulburton@kernel.org>
>>>> Cc: linux-mips@vger.kernel.org
>>>> Cc: Ralf Baechle <ralf@linux-mips.org>
>>>> Cc: James Hogan <jhogan@kernel.org>
>>>> Cc: Vincenzo Frascino <vincenzo.frascino@arm.com>
>>>> Cc: richard.purdie@linuxfoundation.org
>>>> Signed-off-by: Sasha Levin <sashal@kernel.org>
>>>> ---
>>>>  arch/mips/vdso/Makefile | 1 +
>>>>  1 file changed, 1 insertion(+)
>>>>
>>>
>>> Hi Sasha,
>>>
>>> Why was this not added to the 5.4 stable branch?
>>>
>>> Some OpenWrt users ran into this problem with kernel 5.4 on MIPS64 [0].
>>> We backported this patch on our own in OpenWrt [1], but it should be
>>> added to the sable branch in my opinion as it fixes a real problem.
>>>
>>> @Sasha: Can you add it to the 5.4 stable branch or should I send some
>>> special email?
>>
>> It failed building on 5.4. If you'd like it included, please send me a
>> tested backport for 5.4.
>>
>
>I successfully compiled a kernel 5.4.61 with this patch on top with GCC
>8.4 for MIPS 64 big and little Endian.
>
>What was broken in your compile test?

See https://lore.kernel.org/stable/bfdce3ef-5fe9-8dab-1695-be3d33727529@roeck-us.net/

-- 
Thanks,
Sasha