* [PATCH] proc: fix map_files test on F29
@ 2019-04-14 12:36 Alexey Dobriyan
2019-04-14 15:24 ` Randy Dunlap
0 siblings, 1 reply; 4+ messages in thread
From: Alexey Dobriyan @ 2019-04-14 12:36 UTC (permalink / raw)
To: akpm; +Cc: linux-kernel
F29 bans mapping first 64KB even for root making test fail.
Iterate from address 0 until mmap() works.
Gentoo (root):
openat(AT_FDCWD, "/dev/zero", O_RDONLY) = 3
mmap(NULL, 4096, PROT_NONE, MAP_PRIVATE|MAP_FIXED, 3, 0) = 0
Gentoo (non-root):
openat(AT_FDCWD, "/dev/zero", O_RDONLY) = 3
mmap(NULL, 4096, PROT_NONE, MAP_PRIVATE|MAP_FIXED, 3, 0) = -1 EPERM (Operation not permitted)
mmap(0x1000, 4096, PROT_NONE, MAP_PRIVATE|MAP_FIXED, 3, 0) = 0x1000
F29 (root):
openat(AT_FDCWD, "/dev/zero", O_RDONLY) = 3
mmap(NULL, 4096, PROT_NONE, MAP_PRIVATE|MAP_FIXED, 3, 0) = -1 EACCES (Permission denied)
mmap(0x1000, 4096, PROT_NONE, MAP_PRIVATE|MAP_FIXED, 3, 0) = -1 EACCES (Permission denied)
mmap(0x2000, 4096, PROT_NONE, MAP_PRIVATE|MAP_FIXED, 3, 0) = -1 EACCES (Permission denied)
mmap(0x3000, 4096, PROT_NONE, MAP_PRIVATE|MAP_FIXED, 3, 0) = -1 EACCES (Permission denied)
mmap(0x4000, 4096, PROT_NONE, MAP_PRIVATE|MAP_FIXED, 3, 0) = -1 EACCES (Permission denied)
mmap(0x5000, 4096, PROT_NONE, MAP_PRIVATE|MAP_FIXED, 3, 0) = -1 EACCES (Permission denied)
mmap(0x6000, 4096, PROT_NONE, MAP_PRIVATE|MAP_FIXED, 3, 0) = -1 EACCES (Permission denied)
mmap(0x7000, 4096, PROT_NONE, MAP_PRIVATE|MAP_FIXED, 3, 0) = -1 EACCES (Permission denied)
mmap(0x8000, 4096, PROT_NONE, MAP_PRIVATE|MAP_FIXED, 3, 0) = -1 EACCES (Permission denied)
mmap(0x9000, 4096, PROT_NONE, MAP_PRIVATE|MAP_FIXED, 3, 0) = -1 EACCES (Permission denied)
mmap(0xa000, 4096, PROT_NONE, MAP_PRIVATE|MAP_FIXED, 3, 0) = -1 EACCES (Permission denied)
mmap(0xb000, 4096, PROT_NONE, MAP_PRIVATE|MAP_FIXED, 3, 0) = -1 EACCES (Permission denied)
mmap(0xc000, 4096, PROT_NONE, MAP_PRIVATE|MAP_FIXED, 3, 0) = -1 EACCES (Permission denied)
mmap(0xd000, 4096, PROT_NONE, MAP_PRIVATE|MAP_FIXED, 3, 0) = -1 EACCES (Permission denied)
mmap(0xe000, 4096, PROT_NONE, MAP_PRIVATE|MAP_FIXED, 3, 0) = -1 EACCES (Permission denied)
mmap(0xf000, 4096, PROT_NONE, MAP_PRIVATE|MAP_FIXED, 3, 0) = -1 EACCES (Permission denied)
mmap(0x10000, 4096, PROT_NONE, MAP_PRIVATE|MAP_FIXED, 3, 0) = 0x10000
Now all proc tests succeed on F29 if run as root, at last!
Signed-off-by: Alexey Dobriyan <adobriyan@gmail.com>
---
tools/testing/selftests/proc/proc-self-map-files-002.c | 20 ++++++++---------
1 file changed, 10 insertions(+), 10 deletions(-)
--- a/tools/testing/selftests/proc/proc-self-map-files-002.c
+++ b/tools/testing/selftests/proc/proc-self-map-files-002.c
@@ -46,12 +46,9 @@ static void fail(const char *fmt, unsigned long a, unsigned long b)
int main(void)
{
- const unsigned int PAGE_SIZE = sysconf(_SC_PAGESIZE);
-#ifdef __arm__
- unsigned long va = 2 * PAGE_SIZE;
-#else
- unsigned long va = 0;
-#endif
+ const int PAGE_SIZE = sysconf(_SC_PAGESIZE);
+ const unsigned long va_max = 1UL << 32;
+ unsigned long va;
void *p;
int fd;
unsigned long a, b;
@@ -60,10 +57,13 @@ int main(void)
if (fd == -1)
return 1;
- p = mmap((void *)va, PAGE_SIZE, PROT_NONE, MAP_PRIVATE|MAP_FILE|MAP_FIXED, fd, 0);
- if (p == MAP_FAILED) {
- if (errno == EPERM)
- return 4;
+ for (va = 0; va < va_max; va += PAGE_SIZE) {
+ p = mmap((void *)va, PAGE_SIZE, PROT_NONE, MAP_PRIVATE|MAP_FILE|MAP_FIXED, fd, 0);
+ if (p == (void *)va)
+ break;
+ }
+ if (va == va_max) {
+ fprintf(stderr, "error: mmap doesn't like you\n");
return 1;
}
^ permalink raw reply [flat|nested] 4+ messages in thread
* Re: [PATCH] proc: fix map_files test on F29
2019-04-14 12:36 [PATCH] proc: fix map_files test on F29 Alexey Dobriyan
@ 2019-04-14 15:24 ` Randy Dunlap
2019-04-14 15:58 ` Alexey Dobriyan
0 siblings, 1 reply; 4+ messages in thread
From: Randy Dunlap @ 2019-04-14 15:24 UTC (permalink / raw)
To: Alexey Dobriyan, akpm; +Cc: linux-kernel
On 4/14/19 5:36 AM, Alexey Dobriyan wrote:
> F29 bans mapping first 64KB even for root making test fail.
so there is an F29-only patch that prevents that?
something that is not in mainline?
thanks.
> Iterate from address 0 until mmap() works.
>
> Gentoo (root):
>
> openat(AT_FDCWD, "/dev/zero", O_RDONLY) = 3
> mmap(NULL, 4096, PROT_NONE, MAP_PRIVATE|MAP_FIXED, 3, 0) = 0
>
> Gentoo (non-root):
>
> openat(AT_FDCWD, "/dev/zero", O_RDONLY) = 3
> mmap(NULL, 4096, PROT_NONE, MAP_PRIVATE|MAP_FIXED, 3, 0) = -1 EPERM (Operation not permitted)
> mmap(0x1000, 4096, PROT_NONE, MAP_PRIVATE|MAP_FIXED, 3, 0) = 0x1000
>
> F29 (root):
>
> openat(AT_FDCWD, "/dev/zero", O_RDONLY) = 3
> mmap(NULL, 4096, PROT_NONE, MAP_PRIVATE|MAP_FIXED, 3, 0) = -1 EACCES (Permission denied)
> mmap(0x1000, 4096, PROT_NONE, MAP_PRIVATE|MAP_FIXED, 3, 0) = -1 EACCES (Permission denied)
> mmap(0x2000, 4096, PROT_NONE, MAP_PRIVATE|MAP_FIXED, 3, 0) = -1 EACCES (Permission denied)
> mmap(0x3000, 4096, PROT_NONE, MAP_PRIVATE|MAP_FIXED, 3, 0) = -1 EACCES (Permission denied)
> mmap(0x4000, 4096, PROT_NONE, MAP_PRIVATE|MAP_FIXED, 3, 0) = -1 EACCES (Permission denied)
> mmap(0x5000, 4096, PROT_NONE, MAP_PRIVATE|MAP_FIXED, 3, 0) = -1 EACCES (Permission denied)
> mmap(0x6000, 4096, PROT_NONE, MAP_PRIVATE|MAP_FIXED, 3, 0) = -1 EACCES (Permission denied)
> mmap(0x7000, 4096, PROT_NONE, MAP_PRIVATE|MAP_FIXED, 3, 0) = -1 EACCES (Permission denied)
> mmap(0x8000, 4096, PROT_NONE, MAP_PRIVATE|MAP_FIXED, 3, 0) = -1 EACCES (Permission denied)
> mmap(0x9000, 4096, PROT_NONE, MAP_PRIVATE|MAP_FIXED, 3, 0) = -1 EACCES (Permission denied)
> mmap(0xa000, 4096, PROT_NONE, MAP_PRIVATE|MAP_FIXED, 3, 0) = -1 EACCES (Permission denied)
> mmap(0xb000, 4096, PROT_NONE, MAP_PRIVATE|MAP_FIXED, 3, 0) = -1 EACCES (Permission denied)
> mmap(0xc000, 4096, PROT_NONE, MAP_PRIVATE|MAP_FIXED, 3, 0) = -1 EACCES (Permission denied)
> mmap(0xd000, 4096, PROT_NONE, MAP_PRIVATE|MAP_FIXED, 3, 0) = -1 EACCES (Permission denied)
> mmap(0xe000, 4096, PROT_NONE, MAP_PRIVATE|MAP_FIXED, 3, 0) = -1 EACCES (Permission denied)
> mmap(0xf000, 4096, PROT_NONE, MAP_PRIVATE|MAP_FIXED, 3, 0) = -1 EACCES (Permission denied)
> mmap(0x10000, 4096, PROT_NONE, MAP_PRIVATE|MAP_FIXED, 3, 0) = 0x10000
>
> Now all proc tests succeed on F29 if run as root, at last!
>
> Signed-off-by: Alexey Dobriyan <adobriyan@gmail.com>
> ---
>
> tools/testing/selftests/proc/proc-self-map-files-002.c | 20 ++++++++---------
> 1 file changed, 10 insertions(+), 10 deletions(-)
>
> --- a/tools/testing/selftests/proc/proc-self-map-files-002.c
> +++ b/tools/testing/selftests/proc/proc-self-map-files-002.c
> @@ -46,12 +46,9 @@ static void fail(const char *fmt, unsigned long a, unsigned long b)
>
> int main(void)
> {
> - const unsigned int PAGE_SIZE = sysconf(_SC_PAGESIZE);
> -#ifdef __arm__
> - unsigned long va = 2 * PAGE_SIZE;
> -#else
> - unsigned long va = 0;
> -#endif
> + const int PAGE_SIZE = sysconf(_SC_PAGESIZE);
> + const unsigned long va_max = 1UL << 32;
> + unsigned long va;
> void *p;
> int fd;
> unsigned long a, b;
> @@ -60,10 +57,13 @@ int main(void)
> if (fd == -1)
> return 1;
>
> - p = mmap((void *)va, PAGE_SIZE, PROT_NONE, MAP_PRIVATE|MAP_FILE|MAP_FIXED, fd, 0);
> - if (p == MAP_FAILED) {
> - if (errno == EPERM)
> - return 4;
> + for (va = 0; va < va_max; va += PAGE_SIZE) {
> + p = mmap((void *)va, PAGE_SIZE, PROT_NONE, MAP_PRIVATE|MAP_FILE|MAP_FIXED, fd, 0);
> + if (p == (void *)va)
> + break;
> + }
> + if (va == va_max) {
> + fprintf(stderr, "error: mmap doesn't like you\n");
> return 1;
> }
>
>
--
~Randy
^ permalink raw reply [flat|nested] 4+ messages in thread
* Re: [PATCH] proc: fix map_files test on F29
2019-04-14 15:24 ` Randy Dunlap
@ 2019-04-14 15:58 ` Alexey Dobriyan
2019-04-14 16:06 ` Randy Dunlap
0 siblings, 1 reply; 4+ messages in thread
From: Alexey Dobriyan @ 2019-04-14 15:58 UTC (permalink / raw)
To: Randy Dunlap; +Cc: akpm, linux-kernel
On Sun, Apr 14, 2019 at 08:24:53AM -0700, Randy Dunlap wrote:
> On 4/14/19 5:36 AM, Alexey Dobriyan wrote:
> > F29 bans mapping first 64KB even for root making test fail.
>
> so there is an F29-only patch that prevents that?
> something that is not in mainline?
Ehh, no. F29 ships stable kernels actually.
Test assumes mmap(0, MAP_FIXED) succeeds as root. F29 uses SELinux
to ban that.
^ permalink raw reply [flat|nested] 4+ messages in thread
* Re: [PATCH] proc: fix map_files test on F29
2019-04-14 15:58 ` Alexey Dobriyan
@ 2019-04-14 16:06 ` Randy Dunlap
0 siblings, 0 replies; 4+ messages in thread
From: Randy Dunlap @ 2019-04-14 16:06 UTC (permalink / raw)
To: Alexey Dobriyan; +Cc: akpm, linux-kernel
On 4/14/19 8:58 AM, Alexey Dobriyan wrote:
> On Sun, Apr 14, 2019 at 08:24:53AM -0700, Randy Dunlap wrote:
>> On 4/14/19 5:36 AM, Alexey Dobriyan wrote:
>>> F29 bans mapping first 64KB even for root making test fail.
>>
>> so there is an F29-only patch that prevents that?
>> something that is not in mainline?
>
> Ehh, no. F29 ships stable kernels actually.
>
> Test assumes mmap(0, MAP_FIXED) succeeds as root. F29 uses SELinux
> to ban that.
>
Thanks for the clarification.
--
~Randy
^ permalink raw reply [flat|nested] 4+ messages in thread
end of thread, other threads:[~2019-04-14 16:06 UTC | newest]
Thread overview: 4+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2019-04-14 12:36 [PATCH] proc: fix map_files test on F29 Alexey Dobriyan
2019-04-14 15:24 ` Randy Dunlap
2019-04-14 15:58 ` Alexey Dobriyan
2019-04-14 16:06 ` Randy Dunlap
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).