EFI fix for v6.9 (#1)

- Avoid triggering the softlockup watchdog when accepting large chunks
   of memory from the hypervisor
 -----BEGIN PGP SIGNATURE-----
 
 iHUEABYIAB0WIQQQm/3uucuRGn1Dmh0wbglWLn0tXAUCZjS4yAAKCRAwbglWLn0t
 XIOQAPkBZlRdhlbegKjguyvVk9S6EKAgA1j94L/gK8ZL9olbUwD/fEuBIm+U2wtv
 Ntulu0l5kiHSJzC8lwsUdtUSd+ovUQg=
 =o5dU
 -----END PGP SIGNATURE-----

Merge tag 'efi-urgent-for-v6.9-1' of git://git.kernel.org/pub/scm/linux/kernel/git/efi/efi

Pull EFI fix from Ard Biesheuvel:
 "This works around a shortcoming in the memory acceptation API, which
  may apparently hog the CPU for long enough to trigger the softlockup
  watchdog.

  Note that this only affects confidential VMs running under the Intel
  TDX hypervisor, which is why I accepted this for now, but this should
  obviously be fixed properly in the future"

* tag 'efi-urgent-for-v6.9-1' of git://git.kernel.org/pub/scm/linux/kernel/git/efi/efi:
  efi/unaccepted: touch soft lockup during memory accept
This commit is contained in:
Linus Torvalds 2024-05-03 12:05:19 -07:00
commit f094ee78e0

View File

@ -4,6 +4,7 @@
#include <linux/memblock.h>
#include <linux/spinlock.h>
#include <linux/crash_dump.h>
#include <linux/nmi.h>
#include <asm/unaccepted_memory.h>
/* Protects unaccepted memory bitmap and accepting_list */
@ -149,6 +150,9 @@ void accept_memory(phys_addr_t start, phys_addr_t end)
}
list_del(&range.list);
touch_softlockup_watchdog();
spin_unlock_irqrestore(&unaccepted_memory_lock, flags);
}