| Message ID | 20200924144246.12750-2-mdoucha@suse.cz |
|---|---|
| State | Superseded |
| Headers | show |
| Series | [1/2] Fix formatting signal name in tst_strstatus() | expand |
On Thu, Sep 24, 2020 at 10:43 PM Martin Doucha <mdoucha@suse.cz> wrote: > The first testcase keeps segfaulting on older kernels with only 4KB initial > stack mapping > > Signed-off-by: Martin Doucha <mdoucha@suse.cz> > Reviewed-by: Li Wang <liwang@redhat.com> > --- > > Note: Increasing initial stack size to 8KB fixes the segfaults but I'm not > entirely sure whether we shouldn't use PTHREAD_STACK_MIN as the minimal > size > instead. Pthread man page isn't entirely clear whether PTHREAD_STACK_MIN > is > the minimal value for stack size after it grows to the limit, or the > minimal > size of the initial stack mapping (before it starts growing). > I slightly think the latter, which means PTHREAD_STACK_MIN is the minimal size of the initial stack mapping. So maybe better to use the PTHREAD_STACK_MIN as mmaped_size, we even don't need to have a MAX() there. > > On x86_64, PTHREAD_STACK_MIN is 4 pages (16KB). On PPC64, it's 2 pages > (128KB). > testcases/kernel/syscalls/mmap/mmap18.c | 5 +++-- > 1 file changed, 3 insertions(+), 2 deletions(-) > > diff --git a/testcases/kernel/syscalls/mmap/mmap18.c > b/testcases/kernel/syscalls/mmap/mmap18.c > index 966bf673e..8b149fb93 100644 > --- a/testcases/kernel/syscalls/mmap/mmap18.c > +++ b/testcases/kernel/syscalls/mmap/mmap18.c > @@ -201,10 +201,11 @@ static void grow_stack_fail(size_t stack_size, > size_t mapped_size) > static void run_test(void) > { > size_t stack_size = 8 * PTHREAD_STACK_MIN; > + size_t minsize = MAX(page_size, 8192); > > - grow_stack_success(stack_size, page_size); > + grow_stack_success(stack_size, minsize); > grow_stack_success(stack_size, stack_size/2); > - grow_stack_fail(stack_size, page_size); > + grow_stack_fail(stack_size, minsize); > grow_stack_fail(stack_size, stack_size/2); > } > > -- > 2.28.0 > > > -- > Mailing list info: https://lists.linux.it/listinfo/ltp > >
Hi! > The first testcase keeps segfaulting on older kernels with only 4KB initial > stack mapping > > Signed-off-by: Martin Doucha <mdoucha@suse.cz> > --- > > Note: Increasing initial stack size to 8KB fixes the segfaults but I'm not > entirely sure whether we shouldn't use PTHREAD_STACK_MIN as the minimal size > instead. Pthread man page isn't entirely clear whether PTHREAD_STACK_MIN is > the minimal value for stack size after it grows to the limit, or the minimal > size of the initial stack mapping (before it starts growing). I vote for passing down PTHREAD_STACK_MIN for the initial mapping. > On x86_64, PTHREAD_STACK_MIN is 4 pages (16KB). On PPC64, it's 2 pages > (128KB). > > testcases/kernel/syscalls/mmap/mmap18.c | 5 +++-- > 1 file changed, 3 insertions(+), 2 deletions(-) > > diff --git a/testcases/kernel/syscalls/mmap/mmap18.c b/testcases/kernel/syscalls/mmap/mmap18.c > index 966bf673e..8b149fb93 100644 > --- a/testcases/kernel/syscalls/mmap/mmap18.c > +++ b/testcases/kernel/syscalls/mmap/mmap18.c > @@ -201,10 +201,11 @@ static void grow_stack_fail(size_t stack_size, size_t mapped_size) > static void run_test(void) > { > size_t stack_size = 8 * PTHREAD_STACK_MIN; > + size_t minsize = MAX(page_size, 8192); > > - grow_stack_success(stack_size, page_size); > + grow_stack_success(stack_size, minsize); > grow_stack_success(stack_size, stack_size/2); > - grow_stack_fail(stack_size, page_size); > + grow_stack_fail(stack_size, minsize); > grow_stack_fail(stack_size, stack_size/2); > } > > -- > 2.28.0 > > > -- > Mailing list info: https://lists.linux.it/listinfo/ltp
diff --git a/testcases/kernel/syscalls/mmap/mmap18.c b/testcases/kernel/syscalls/mmap/mmap18.c index 966bf673e..8b149fb93 100644 --- a/testcases/kernel/syscalls/mmap/mmap18.c +++ b/testcases/kernel/syscalls/mmap/mmap18.c @@ -201,10 +201,11 @@ static void grow_stack_fail(size_t stack_size, size_t mapped_size) static void run_test(void) { size_t stack_size = 8 * PTHREAD_STACK_MIN; + size_t minsize = MAX(page_size, 8192); - grow_stack_success(stack_size, page_size); + grow_stack_success(stack_size, minsize); grow_stack_success(stack_size, stack_size/2); - grow_stack_fail(stack_size, page_size); + grow_stack_fail(stack_size, minsize); grow_stack_fail(stack_size, stack_size/2); }
The first testcase keeps segfaulting on older kernels with only 4KB initial stack mapping Signed-off-by: Martin Doucha <mdoucha@suse.cz> --- Note: Increasing initial stack size to 8KB fixes the segfaults but I'm not entirely sure whether we shouldn't use PTHREAD_STACK_MIN as the minimal size instead. Pthread man page isn't entirely clear whether PTHREAD_STACK_MIN is the minimal value for stack size after it grows to the limit, or the minimal size of the initial stack mapping (before it starts growing). On x86_64, PTHREAD_STACK_MIN is 4 pages (16KB). On PPC64, it's 2 pages (128KB). testcases/kernel/syscalls/mmap/mmap18.c | 5 +++-- 1 file changed, 3 insertions(+), 2 deletions(-)