Sometimes zero is a valid value for getauxval (e.g. AT_EXECFD). Make
sure that we can distinguish between a valid zero value and a not found
entry by setting errno.
Assumes that getauxval from sys/auxv.h sets errno correctly.
Signed-off-by: Vivian Wang
---
util/getauxval.c | 6 +-
1 file
qemu-devel/20240721090817.120888-1-...@dram.page/
Vivian Wang (2):
util/getauxval: Ensure setting errno if not found
linux-user/main: Check errno when getting AT_EXECFD
linux-user/main.c | 3 ++-
util/getauxval.c | 6 +-
2 files changed, 7 insertions(+), 2 deletions(-)
--
2.45.1
It's possible for AT_EXECFD to end up with a valid value of 0. Check
errno when using qemu_getauxval instead of return value to handle this
case.
Not handling this case leads to a confusing condition where the
executable ends up as fd 0, i.e. stdin.
Signed-off-by: Vivian Wang
Review
On 7/22/24 08:18, Xingtao Yao (Fujitsu) wrote:
>
>> -Original Message-
>> From: qemu-devel-bounces+yaoxt.fnst=fujitsu@nongnu.org
>> On Behalf Of Vivian
>> Wang
>> Sent: Sunday, July 21, 2024 5:08 PM
>> To: qemu-devel@nongnu.org
>> Cc:
It's possible for AT_EXECFD to end up with a valid value of 0. Check
errno when using qemu_getauxval instead of return value to handle this
case.
Not handling this case leads to a confusing condition where the
executable ends up as fd 0, i.e. stdin.
Signed-off-by: Vivian Wang
Sometimes zero is a valid value for getauxval (e.g. AT_EXECFD). Make
sure that we can distinguish between a valid zero value and a not found
entry by setting errno.
Ignore getauxval from sys/auxv.h on glibc < 2.19 because it does not set
errno.
Signed-off-by: Vivian Wang
---
util/getauxva
Found when trying to build coreutils under linux-user as binfmt_misc
interpreter with "open-binary" flag.
I'm unsure whether the version detection in patch 1 makes sense. Please
advise on how to handle situations like this.
Resolves: https://gitlab.com/qemu-project/qemu/-/issu
ch that .align+.org
> wasn't actually honored. This new form doesn't require the
> test bytes to be aligned in the binary.
>
>
> r~
I've confirmed that this fixes #1155
Tested-by: Vivian Wang
> Ilya Leoshkevich (4):
> linux-user: Clear translations a
On 8/17/22 23:05, Ilya Leoshkevich wrote:
> Hi,
>
> I noticed that when we get a SEGV due to jumping to non-readable
> memory, sometimes si_addr and program counter in siginfo_t are slightly
> off. I tracked this down to the assumption that translators stop before
> the end of a page, while in real
New Replica Watches Online
http://checkout.valentlne.com/replica/lib/
Europe was created by history. America was created by philosophy.
There is no such thing in anyone's life as an unimportant day.
Each individual woman's body demands to be accepted on its own terms.
Bad law is the worst
10 matches
Mail list logo