"ps -e" without procfs(5)

Mikolaj Golub trociny at freebsd.org
Mon Oct 31 10:54:50 UTC 2011

On Mon, 31 Oct 2011 11:49:48 +0200 Kostik Belousov wrote:

 KB> On Sat, Oct 29, 2011 at 01:32:39PM +0300, Mikolaj Golub wrote:
 >> What do you think about the attached patch? This is a kernel
 >> part. COMPAT_FREEBSD32 has not been tested after the last update (just checked
 >> that it compiles): it looks I will not have access to amd64 box for testing
 >> during the weekend. I will test it after the weekend.
 >> Both kernel and userland parts are available here:
 >> http://people.freebsd.org/~trociny/env.sys.patch
 >> http://people.freebsd.org/~trociny/env.user.patch
 >> Currently there is an issue with procstat -x: if one tried to run it on 64 bit
 >> for a 32 bit process it would not detect this so would output a garbage. Could
 >> somebody recommend a way how to get this info about a process from userlend?

 KB> I think it is better to use sys/elf.h over the machine/elf.h.

 KB> Please change the comment for PROC_AUXV_MAX to "Safety limit on auxv size".
 KB> Also, it worth adding a comment saying that we are reading aux vectors twice,
 KB> first to get a size, second time to fetch a content, for simplicity.

 KB> When reading aux vector, if the PROC_AUXV_MAX entries are iterated over,
 KB> and we still not reached AT_NULL, the return error is 0. Was it intended ?

According to kern_exec.c it is possible that a process doesn't have auxv at
all. I don't know a way how to detect this. So because PROC_AUXV_MAX is much
larger than expected amount of aux entries and we have not reached AT_NULL it
is most likely the process doesn't have auxv and 0 length array (without
error) is returned. 

If you think I should return a error in this situation, I can add this. Please
tell me the error code I should return :-).

Also, may be there is a sane way to check on auxv existence?

 KB> For PROC_ARG and PROC_ENV, you blindly trust the read values of the arg and
 KB> env vector sizes. This can easily cause kernel panics due to unability to
 KB> malloc the requested memory. I recommend to put some clump, and twice
 KB> of (PATH_MAX + ARG_MAX) is probably enough (see kern_exec.c, in particular,
 KB> exec_alloc_args). Also, you might use the swappable memory for the strings
 KB> as well, in the style of exec_alloc_args().

 KB> I suspect this is my bug: Reading the GET_PS_STRINGS_CHUNK_SZ may validly
 KB> return EFAULT if the string is shorter than the chunk and aligned at
 KB> the end of the page, assuming the next page is not mapped. There should
 KB> be a fallback to fubyte() read loop. I remember that copyinstr() was
 KB> unsuitable.

 KB> The checks for P_WEXIT in the linprocfs routines look strange. Since
 KB> you are unlocking the process right after the check, it does not make
 KB> sense. In fact, the checks are not needed, I believe, since pseudofs
 KB> already did the hold (see e.g. pfs_read and pfs_visible).

Ah, right. Unintentionally added when was adding the P_SYSTEM check.

Thank you for all your comments. I will do this.

Mikolaj Golub

More information about the freebsd-hackers mailing list