#asahi · 2021-02-27 16:37
marcanbut somehow linux gets away with dmb for DMA ordering, so... who's wrong?#asahi · 2021-02-27 16:37
marcanmodwizcode: I think I found that PDF, if it's the one I mentioned on the list on the v1 thread#asahi · 2021-02-26 04:17
marcanthey use that too#asahi · 2021-02-26 04:17
marcanI doubt they'd break vdmtool...#asahi · 2021-02-25 18:38
marcanYou can use a C to A cable too, it doesn't have to be C to C#asahi · 2021-02-25 13:14
marcanjust plug it in#asahi · 2021-02-25 12:26
marcanyeah#asahi · 2021-02-25 12:25
marcanthis replaces the old target disk mode which worked as a block device#asahi · 2021-02-25 12:25
marcanjeff_miw: macOS already supports being a device and behaving as a network device for data transfer. it is also used for DFU mode.#asahi · 2021-02-25 12:24
marcanPXN is useful for e.g. W^X in kernel mode#asahi · 2021-02-25 10:51
marcan(they're gone)#asahi · 2021-02-24 16:37
marcanI'll probably end up writing a litmus test-ish thing for this to prove it to myself that dmb is enough#asahi · 2021-02-24 16:37
marcanand that is also where many guarantees can easily break down#asahi · 2021-02-24 16:37
marcanit all gets very confusing when you step outside CPUs and plain memory#asahi · 2021-02-24 16:37
marcani.e. what I *really* want to know is whether A writing memory and writing to peripheral B is ordered wrt C reading from peripheral B and reading the memory that A wrote#asahi · 2021-02-24 16:36
marcanthere is an additional confounding factor though, that this is really a 3-party behavior#asahi · 2021-02-24 16:36
marcanor e.g. only between normal-normal and device-device within a peripheral window (there is a reference to an imp-def window size in the spec)#asahi · 2021-02-24 16:35
marcanmodwizcode: my question is whether it guarantees ordering between device memory and normal memory ops#asahi · 2021-02-24 16:26
marcaneither that or Tantatrul got to me#asahi · 2021-02-24 16:22
marcanoh right, for some reason I thought Advanced#asahi · 2021-02-24 16:21
marcanApple just blatantly called their bits by the SSE names, ARM integrated it into the existing FPCR in a slightly less obvious way#asahi · 2021-02-24 16:21
marcanthe feature is *very obviously* an x86 emulation feature either way#asahi · 2021-02-24 16:20
marcanmaybe someone at ARM went "eh... "Advanced FP" works too" :p#asahi · 2021-02-24 16:20
marcanamusingly, they both call it "AFP", but I'm pretty sure the "A" stands for "Apple" in Apple's case#asahi · 2021-02-24 16:20
marcanwouldn't be surprised if it came from them#asahi · 2021-02-24 16:20
marcanthat came out last year, so presumably Apple's is basically a prerelease version of the same#asahi · 2021-02-24 16:19
marcannope#asahi · 2021-02-24 16:17
marcanemail sent#asahi · 2021-02-24 16:00
marcanit does trap in VMs if I'm not mistaken, so at least there's that, but I need to double check#asahi · 2021-02-24 15:47
marcanaaand that's a macOS CVE. hoo boy.#asahi · 2021-02-24 15:36
marcanmaz: I don't know what s3_5_c15_c10_1 is, but it's also 2 bits, and writable by EL0 by default#asahi · 2021-02-24 15:35
marcanuh-oh#asahi · 2021-02-24 12:49
marcanpushed m1n1 changes for EL0 support#asahi · 2021-02-24 12:42
marcanI guess this makes sense to some extent; it means the kernel save/restore code can just unconditionally save/restore these flags, regardless of whether a given process has the feature enabled#asahi · 2021-02-24 12:42
marcanmaz: okay, yeah, this stuff traps from EL0 if not enabled, even with TGE=1#asahi · 2021-02-24 12:37
marcanwell okay, it's alias time then#asahi · 2021-02-24 12:37
marcan"architecturally secure"#asahi · 2021-02-24 12:37
marcanlovely#asahi · 2021-02-24 12:37
marcanhahahaha#asahi · 2021-02-24 12:37
marcanargh, so this is architectural?#asahi · 2021-02-24 12:34
marcanis there any mechanism that would forbid EL2/1 execution from EL0-accessible pages, other than the PXN bit in the page tables?#asahi · 2021-02-24 12:34
marcannow I wonder if this is an apple-ism somewhere...#asahi · 2021-02-24 12:31
marcanyeah, it's execution that breaks#asahi · 2021-02-24 12:31
marcanbut I am setting it; however, things break before I take an exception normally anyway#asahi · 2021-02-24 12:30
marcanactually, that bit *dis*ables enabling PAN on exceptions#asahi · 2021-02-24 12:29
marcansec, doing an actual controlled test with only some mappings, not the whole thing#asahi · 2021-02-24 12:29
marcanj`ey: yes, but I set PSTATE.PAN to 0 and it still breaks#asahi · 2021-02-24 12:29
marcanno wait, but I don't set that either#asahi · 2021-02-24 12:28
marcanoh wait, PXN#asahi · 2021-02-24 12:26
marcanI remember this from back when I was messing with page permissions, but setting bit 6 in the PTEs breaks everything and it doesn't look like it's PAN?