CVE-2024-35938

In the Linux kernel, the following vulnerability has been resolved:<br /> <br /> wifi: ath11k: decrease MHI channel buffer length to 8KB<br /> <br /> Currently buf_len field of ath11k_mhi_config_qca6390 is assigned<br /> with 0, making MHI use a default size, 64KB, to allocate channel<br /> buffers. This is likely to fail in some scenarios where system<br /> memory is highly fragmented and memory compaction or reclaim is<br /> not allowed.<br /> <br /> There is a fail report which is caused by it:<br /> kworker/u32:45: page allocation failure: order:4, mode:0x40c00(GFP_NOIO|__GFP_COMP), nodemask=(null),cpuset=/,mems_allowed=0<br /> CPU: 0 PID: 19318 Comm: kworker/u32:45 Not tainted 6.8.0-rc3-1.gae4495f-default #1 openSUSE Tumbleweed (unreleased) 493b6d5b382c603654d7a81fc3c144d59a1dfceb<br /> Workqueue: events_unbound async_run_entry_fn<br /> Call Trace:<br /> <br /> dump_stack_lvl+0x47/0x60<br /> warn_alloc+0x13a/0x1b0<br /> ? srso_alias_return_thunk+0x5/0xfbef5<br /> ? __alloc_pages_direct_compact+0xab/0x210<br /> __alloc_pages_slowpath.constprop.0+0xd3e/0xda0<br /> __alloc_pages+0x32d/0x350<br /> ? mhi_prepare_channel+0x127/0x2d0 [mhi 40df44e07c05479f7a6e7b90fba9f0e0031a7814]<br /> __kmalloc_large_node+0x72/0x110<br /> __kmalloc+0x37c/0x480<br /> ? mhi_map_single_no_bb+0x77/0xf0 [mhi 40df44e07c05479f7a6e7b90fba9f0e0031a7814]<br /> ? mhi_prepare_channel+0x127/0x2d0 [mhi 40df44e07c05479f7a6e7b90fba9f0e0031a7814]<br /> mhi_prepare_channel+0x127/0x2d0 [mhi 40df44e07c05479f7a6e7b90fba9f0e0031a7814]<br /> __mhi_prepare_for_transfer+0x44/0x80 [mhi 40df44e07c05479f7a6e7b90fba9f0e0031a7814]<br /> ? __pfx_____mhi_prepare_for_transfer+0x10/0x10 [mhi 40df44e07c05479f7a6e7b90fba9f0e0031a7814]<br /> device_for_each_child+0x5c/0xa0<br /> ? __pfx_pci_pm_resume+0x10/0x10<br /> ath11k_core_resume+0x65/0x100 [ath11k a5094e22d7223135c40d93c8f5321cf09fd85e4e]<br /> ? srso_alias_return_thunk+0x5/0xfbef5<br /> ath11k_pci_pm_resume+0x32/0x60 [ath11k_pci 830b7bfc3ea80ebef32e563cafe2cb55e9cc73ec]<br /> ? srso_alias_return_thunk+0x5/0xfbef5<br /> dpm_run_callback+0x8c/0x1e0<br /> device_resume+0x104/0x340<br /> ? __pfx_dpm_watchdog_handler+0x10/0x10<br /> async_resume+0x1d/0x30<br /> async_run_entry_fn+0x32/0x120<br /> process_one_work+0x168/0x330<br /> worker_thread+0x2f5/0x410<br /> ? __pfx_worker_thread+0x10/0x10<br /> kthread+0xe8/0x120<br /> ? __pfx_kthread+0x10/0x10<br /> ret_from_fork+0x34/0x50<br /> ? __pfx_kthread+0x10/0x10<br /> ret_from_fork_asm+0x1b/0x30<br /> <br /> <br /> Actually those buffers are used only by QMI target -&gt; host communication.<br /> And for WCN6855 and QCA6390, the largest packet size for that is less<br /> than 6KB. So change buf_len field to 8KB, which results in order 1<br /> allocation if page size is 4KB. In this way, we can at least save some<br /> memory, and as well as decrease the possibility of allocation failure<br /> in those scenarios.<br /> <br /> Tested-on: WCN6855 hw2.0 PCI WLAN.HSP.1.1-03125-QCAHSPSWPL_V1_V2_SILICONZ_LITE-3.6510.30