Wind River Support Network

Description

In the Linux kernel, the following vulnerability has been resolved:[EOL][EOL]mm: hugetlb: avoid soft lockup when mprotect to large memory area[EOL][EOL]When calling mprotect() to a large hugetlb memory area in our customer's[EOL]workload (~300GB hugetlb memory), soft lockup was observed:[EOL][EOL]watchdog: BUG: soft lockup - CPU#98 stuck for 23s! [t2_new_sysv:126916][EOL][EOL]CPU: 98 PID: 126916 Comm: t2_new_sysv Kdump: loaded Not tainted 6.17-rc7[EOL]Hardware name: GIGACOMPUTING R2A3-T40-AAV1/Jefferson CIO, BIOS 5.4.4.1 07/15/2025[EOL]pstate: 20400009 (nzCv daif +PAN -UAO -TCO -DIT -SSBS BTYPE=--)[EOL]pcÂ\xa0: mte_clear_page_tags+0x14/0x24[EOL]lrÂ\xa0: mte_sync_tags+0x1c0/0x240[EOL]spÂ\xa0: ffff80003150bb80[EOL]x29: ffff80003150bb80 x28: ffff00739e9705a8 x27: 0000ffd2d6a00000[EOL]x26: 0000ff8e4bc00000 x25: 00e80046cde00f45 x24: 0000000000022458[EOL]x23: 0000000000000000 x22: 0000000000000004 x21: 000000011b380000[EOL]x20: ffff000000000000 x19: 000000011b379f40 x18: 0000000000000000[EOL]x17: 0000000000000000 x16: 0000000000000000 x15: 0000000000000000[EOL]x14: 0000000000000000 x13: 0000000000000000 x12: 0000000000000000[EOL]x11: 0000000000000000 x10: 0000000000000000 x9 : ffffc875e0aa5e2c[EOL]x8Â\xa0: 0000000000000000 x7 : 0000000000000000 x6 : 0000000000000000[EOL]x5Â\xa0: fffffc01ce7a5c00 x4 : 00000000046cde00 x3 : fffffc0000000000[EOL]x2Â\xa0: 0000000000000004 x1 : 0000000000000040 x0 : ffff0046cde7c000[EOL][EOL]Call trace:[EOL]Â\xa0Â\xa0mte_clear_page_tags+0x14/0x24[EOL]Â\xa0Â\xa0set_huge_pte_at+0x25c/0x280[EOL]Â\xa0Â\xa0hugetlb_change_protection+0x220/0x430[EOL]Â\xa0Â\xa0change_protection+0x5c/0x8c[EOL]Â\xa0Â\xa0mprotect_fixup+0x10c/0x294[EOL]Â\xa0Â\xa0do_mprotect_pkey.constprop.0+0x2e0/0x3d4[EOL]Â\xa0Â\xa0__arm64_sys_mprotect+0x24/0x44[EOL]Â\xa0Â\xa0invoke_syscall+0x50/0x160[EOL]Â\xa0Â\xa0el0_svc_common+0x48/0x144[EOL]Â\xa0Â\xa0do_el0_svc+0x30/0xe0[EOL]Â\xa0Â\xa0el0_svc+0x30/0xf0[EOL]Â\xa0Â\xa0el0t_64_sync_handler+0xc4/0x148[EOL]Â\xa0Â\xa0el0t_64_sync+0x1a4/0x1a8[EOL][EOL]Soft lockup is not triggered with THP or base page because there is[EOL]cond_resched() called for each PMD size.[EOL][EOL]Although the soft lockup was triggered by MTE, it should be not MTE[EOL]specific.  The other processing which takes long time in the loop may[EOL]trigger soft lockup too.[EOL][EOL]So add cond_resched() for hugetlb to avoid soft lockup.