[apparmor] [patch 04/24] make the parse_sub_mode code more generic

John Johansen john.johansen at canonical.com
Wed Mar 12 08:15:46 UTC 2014


On 03/11/2014 11:19 PM, Steve Beattie wrote:
> On Fri, Mar 07, 2014 at 09:31:25AM -0800, john.johansen at canonical.com wrote:
>> Make it more generic so that it can be shared with signals.
>>
>> Signed-off-by: John Johansen <john.johansen at canonical.com>
> 
> So this looks good enough for Acked-by: Steve Beattie <steve at nxnw.org>,
> just one question:
> 
>> +static int parse_X_sub_mode(const char *X, const char *str_mode, int *result, int fail, const char *mode_desc __unused)
>> +{
>> +	int mode = 0;
>> +	const char *p;
>> +
>> +	PDEBUG("Parsing X mode: %s\n", X, str_mode);
>> +
>> +	if (!str_mode)
>> +		return 0;
>> +
>> +	p = str_mode;
>> +	while (*p) {
>> +		char current = *p;
>> +		char lower;
>> +
>> +reeval:
>> +		switch (current) {
>> +		case COD_READ_CHAR:
>> +			PDEBUG("Parsing %s mode: found %s READ\n", X, mode_desc);
>> +			mode |= AA_DBUS_RECEIVE;
>> +			break;
>> +
>> +		case COD_WRITE_CHAR:
>> +			PDEBUG("Parsing %s mode: found %s WRITE\n", X,
>> +			       mode_desc);
>> +			mode |= AA_DBUS_SEND;
> 
> If this function is to be shared between dbus and signals, should
> AA_DBUS_RECEIVE and AA_DBUS SEND be morphed into something more generic?
> Or does that occur in the patch that introduces signal support?
> 
Sigh, yeah we need to patch it into something more generic. Signal is
using it as AA_MAY_SEND, but they need to be explicitly tied.




More information about the AppArmor mailing list