>On Tue 2023-04-04 19:12:01, Jaewon Kim wrote: >> >From: Jaewon Kim >> >> Sent: 03 April 2023 15:40 >> >... >> >> I wanted to print phys_addr_t type value only when kptr_restrict sysctl is >> >> allowed. So I thought I could use %pK for that purpose. And the physical >> >> address is not that long. I wanted to make that length short like 9 hex. >> > >> >Isn't that is the wrong format for physical addresses anyway? >> >They can be larger than virtual ones (eg x86 with PAE). >> >> Yeah, correct. I just used %pK to hide physical address, I thought it could be >> leak in security perspective. Could you give me advice how I can hide the >> address by default and look the address if kptr_restrict allow it? > >Could you please send a patch that would show what value you want to print? > >The initial mail mentioned converting %09lx to %09pK in some driver. >Then it included a warning when compiling from mm/page_alloc.o > O Oh I just found something I'm interested. I was printing rmem->base and rmem->name and wanted to hide the rmem->base. The commit aeb9267eb6b1 ("of: reserved-mem: print out reserved-mem details during boot - v6.3-rc1") just used %pa. If %pa is not good enough in security perspecitve, I think I can use %pa. Yes it was just my curiosity. I don't actually need the length. I'm OK now. Thank you >Honestly, I think that you could just use %pK or %p. IMHO, it does not >make sense to optimize it for the length. > >Anyway, there is still the question if the address is really worth >printing. Will it really help to locate a potential problem? > >Best Regards, >Petr