2016-03-25 17:59 GMT+08:00 Jeff King <[email protected]>:
> On Wed, Mar 23, 2016 at 06:13:22PM +0800, Hui Yiqun wrote:
>
>> +/**
>> + * this function does the following:
>> + *
>> + * 1. if $XDG_RUNTIME_DIR is non-empty, `$XDG_RUNTIME_DIR/git` is used in
>> next
>> + * step, otherwise `/tmp/git-$uid` is taken.
>> + * 2. ensure that above directory does exist. what's more, it must has
>> correct
>> + * permission and ownership.
>> + * 3. a newly allocated string consisting of the path of above directory and
>> + * $filename is returned.
>> + *
>> + * Under following situation, NULL will be returned:
>> + *
>> + * + the directory mentioned in step 1 exists but have wrong permission or
>> + * ownership.
>> + * + the directory or its parent cannot be created.
>> + *
>> + * Notice:
>> + *
>> + * + the caller is responsible for deallocating the returned string.
>> + *
>> + */
>> +extern char *xdg_runtime_dir(const char *filename);
>
> There's a lot of "what" here that the caller doesn't really care about,
> and which may go stale with respect to the implementation over time. Can
> we make something more succinct like:
>
> /*
> * Return a path suitable for writing run-time files related to git,
> * or NULL if no such path can be established. The resulting string
> * should be freed by the caller.
> */
>
> ?
That's clearer, but if I were the caller, I would worry about the
security of the path.
How about adding:
The security of the path is ensured by file permission.
>
>> --- a/path.c
>> +++ b/path.c
>> @@ -5,6 +5,7 @@
>> #include "strbuf.h"
>> #include "string-list.h"
>> #include "dir.h"
>> +#include "git-compat-util.h"
>
> Why do we need this? It should generally be the first file included, as
> it sets up defines used by other header files. It looks like we include
> "cache.h" in this file, which is enough (it explicitly includes
> git-compat-util.h first to cover this case).
I include this header for `getuid` and `stat`. Now that there is an indirect
including, I will delete this one.
>
>> +char *xdg_runtime_dir(const char *filename)
>> +{
>> + struct strbuf sb = STRBUF_INIT;
>> + char *runtime_dir;
>> + struct stat st;
>> + uid_t uid = getuid();
>> +
>> + assert(filename);
>> + runtime_dir = getenv("XDG_RUNTIME_DIR");
>> + if (runtime_dir && *runtime_dir)
>> + strbuf_mkpath(&sb, "%s/git/", runtime_dir);
>> + else
>> + strbuf_mkpath(&sb, "/tmp/git-%d", uid);
>> +
>> + if (!lstat(sb.buf, &st)) {
>> + /*
>> + * As described in XDG base dir spec[1], the subdirectory
>> + * under $XDG_RUNTIME_DIR or its fallback MUST be owned by
>> + * the user, and its unix access mode MUST be 0700.
>> + *
>> + * Calling chmod or chown silently may cause security
>> + * problem if somebody chdir to it, sleep, and then, try
>> + * to open our protected runtime cache or socket.
>> + * So we just put warning and left it to user to solve.
>> + *
>
> There are some minor English problems here (and elsewhere). E.g., you
> probably want "So we just issue a warning and leave it to the user to
> solve.".
Sorry for my English.
>> + if ((st.st_mode & 0777) != S_IRWXU) {
>> + warning("permission of runtime directory '%s' "
>> + "MUST be 0700 instead of 0%o\n",
>> + sb.buf, (st.st_mode & 0777));
>> + return NULL;
>> + } else if (st.st_uid != uid) {
>> + warning("owner of runtime directory '%s' "
>> + "MUST be %d instead of %d\n",
>> + sb.buf, uid, st.st_uid);
>> + return NULL;
>> + }
>
> These cases still leak "sb", I think.
>
>> + /* TODO: check whether st.buf is an directory */
>
> Should we complete this todo? It's should just be S_ISDIR(st.st_mode).
>
>> + } else {
>> + if (safe_create_leading_directories_const(sb.buf) < 0) {
>> + warning("unable to create directories for '%s'\n",
>> + sb.buf);
>> + return NULL;
>> + }
>> + if (mkdir(sb.buf, 0700) < 0) {
>> + warning("unable to mkdir '%s'\n", sb.buf);
>> + return NULL;
>> + }
>
> These ones leak, too.
I will deal with it.
I find there are some similar leakage in this file. I'll fix them in
another patch.
Do you think we need some additional comments for the release of strbuf?
>
> -Peff
--
To unsubscribe from this list: send the line "unsubscribe git" in
the body of a message to [email protected]
More majordomo info at http://vger.kernel.org/majordomo-info.html
