mirror of https://github.com/torvalds/linux.git
94a17f2dc9
PTI uses separate ASIDs (aka. PCIDs) for kernel and user address
spaces. When the kernel needs to flush the user address space, it
just sets a bit in a bitmap and then flushes the entire PCID on
the next switch to userspace.
This bitmap is a single 'unsigned long' which is plenty for all 6
dynamic ASIDs. But, unfortunately, the INVLPGB support brings along a
bunch more user ASIDs, as many as ~2k more. The bitmap can't address
that many.
Fortunately, the bitmap is only needed for PTI and all the CPUs
with INVLPGB are AMD CPUs that aren't vulnerable to Meltdown and
don't need PTI. The only way someone can run into an issue in
practice is by booting with pti=on on a newer AMD CPU.
Disable INVLPGB if PTI is enabled. Avoid overrunning the small
bitmap.
Note: this will be fixed up properly by making the bitmap bigger.
For now, just avoid the mostly theoretical bug.
Fixes:
|
||
|---|---|---|
| .. | ||
| pat | ||
| Makefile | ||
| amdtopology.c | ||
| cpu_entry_area.c | ||
| debug_pagetables.c | ||
| dump_pagetables.c | ||
| extable.c | ||
| fault.c | ||
| hugetlbpage.c | ||
| ident_map.c | ||
| init.c | ||
| init_32.c | ||
| init_64.c | ||
| iomap_32.c | ||
| ioremap.c | ||
| kasan_init_64.c | ||
| kaslr.c | ||
| kmmio.c | ||
| kmsan_shadow.c | ||
| maccess.c | ||
| mem_encrypt.c | ||
| mem_encrypt_amd.c | ||
| mem_encrypt_boot.S | ||
| mm_internal.h | ||
| mmap.c | ||
| mmio-mod.c | ||
| numa.c | ||
| pf_in.c | ||
| pf_in.h | ||
| pgprot.c | ||
| pgtable.c | ||
| pgtable_32.c | ||
| physaddr.c | ||
| physaddr.h | ||
| pkeys.c | ||
| pti.c | ||
| srat.c | ||
| testmmiotrace.c | ||
| tlb.c | ||