Re: [PATCH 6.1 0000/1887] 6.1.191-rc1 review

[Sasha Levin]

On 10/28/26 14:26, Florian Fainelli wrote:
> On 10/28/26 14:12, Florian Fainelli wrote:
>> this is due to 341d4e6294d9 ("highmem: add memcpy_to_folio() and
>> memcpy_from_folio()"). We would need to also back port 00cdf76012ab

That one was already addressed earlier today. Rather than adding
00cdf76012ab, I dropped the highmem/hfs set from the 6.1 queue:

  - highmem: add memcpy_to_folio() and memcpy_from_folio()
  - highmem: fix a memory copy problem in memcpy_from_folio
  - hfs: fix general protection fault in hfs_find_init()
  - pagemap: add filemap_grab_folio()

> And there is this other build failure caused by 997f9c35fcad
> ("Bluetooth: MGMT: fix crash in set_mesh_sync and set_mesh_complete")
>
> net/bluetooth/mgmt.c: In function 'set_mesh_sync':
> net/bluetooth/mgmt.c:2188:36: error: implicit declaration of function
> '__struct_size'; did you mean 'struct_size'?

The 6.1 backport of that patch uses __struct_size(), which on 6.1 is only
available with CONFIG_FORTIFY_SOURCE enabled, so it breaks the build
without it. Fixed by dropping the five-patch Bluetooth MGMT cluster from
the 6.1 queue:

  - Bluetooth: MGMT: Fix possible UAFs
  - Bluetooth: MGMT: Fix dangling pointer on
    mgmt_add_adv_patterns_monitor_complete
  - Bluetooth: MGMT: Fix memory leak in set_ssp_complete
  - Bluetooth: MGMT: fix crash in set_mesh_sync and set_mesh_complete
  - Bluetooth: MGMT: Fix list corruption and UAF in command complete
    handlers

-- 
Thanks,
Sasha