{"uuid": "0f1125b4-e9f4-4841-9829-75cb8d5320a4", "vulnerability_lookup_origin": "1a89b78e-f703-45f3-bb86-59eb712668bd", "author": "2a075640-a300-48a4-bb44-bc6130783b9b", "vulnerability": "CVE-2024-57889", "type": "published-proof-of-concept", "source": "https://t.me/DarkWebInformer_CVEAlerts/1757", "content": "\ud83d\udd17 DarkWebInformer.com - Cyber Threat Intelligence\n\ud83d\udccc CVE ID: CVE-2024-57889\n\ud83d\udd39 Description: In the Linux kernel, the following vulnerability has been resolved:\n\npinctrl: mcp23s08: Fix sleeping in atomic context due to regmap locking\n\nIf a device uses MCP23xxx IO expander to receive IRQs, the following\nbug can happen:\n\n  BUG: sleeping function called from invalid context\n    at kernel/locking/mutex.c:283\n  in_atomic(): 1, irqs_disabled(): 1, non_block: 0, ...\n  preempt_count: 1, expected: 0\n  ...\n  Call Trace:\n  ...\n  __might_resched+0x104/0x10e\n  __might_sleep+0x3e/0x62\n  mutex_lock+0x20/0x4c\n  regmap_lock_mutex+0x10/0x18\n  regmap_update_bits_base+0x2c/0x66\n  mcp23s08_irq_set_type+0x1ae/0x1d6\n  __irq_set_trigger+0x56/0x172\n  __setup_irq+0x1e6/0x646\n  request_threaded_irq+0xb6/0x160\n  ...\n\nWe observed the problem while experimenting with a touchscreen driver which\nused MCP23017 IO expander (I2C).\n\nThe regmap in the pinctrl-mcp23s08 driver uses a mutex for protection from\nconcurrent accesses, which is the default for regmaps without .fast_io,\n.disable_locking, etc.\n\nmcp23s08_irq_set_type() calls regmap_update_bits_base(), and the latter\nlocks the mutex.\n\nHowever, __setup_irq() locks desc-&gt;lock spinlock before calling these\nfunctions. As a result, the system tries to lock the mutex whole holding\nthe spinlock.\n\nIt seems, the internal regmap locks are not needed in this driver at all.\nmcp-&gt;lock seems to protect the regmap from concurrent accesses already,\nexcept, probably, in mcp_pinconf_get/set.\n\nmcp23s08_irq_set_type() and mcp23s08_irq_mask/unmask() are called under\nchip_bus_lock(), which calls mcp23s08_irq_bus_lock(). The latter takes\nmcp-&gt;lock and enables regmap caching, so that the potentially slow I2C\naccesses are deferred until chip_bus_unlock().\n\nThe accesses to the regmap from mcp23s08_probe_one() do not need additional\nlocking.\n\nIn all remaining places where the regmap is accessed, except\nmcp_pinconf_get/set(), the driver already takes mcp-&gt;lock.\n\nThis patch adds locking in mcp_pinconf_get/set() and disables internal\nlocking in the regmap config. Among other things, it fixes the sleeping\nin atomic context described above.\n\ud83d\udccf Published: 2025-01-15T13:05:41.769Z\n\ud83d\udccf Modified: 2025-01-15T13:05:41.769Z\n\ud83d\udd17 References:\n1. https://git.kernel.org/stable/c/788d9e9a41b81893d6bb8faa05f045c975278318\n2. https://git.kernel.org/stable/c/c55d186376a87b468c9ee30f2195e0f3857f61a0\n3. https://git.kernel.org/stable/c/9372e160d8211a7e17f2abff8370794f182df785\n4. https://git.kernel.org/stable/c/0310cbad163a908d09d99c26827859365cd71fcb\n5. https://git.kernel.org/stable/c/8c6fd5803b988a5e78c9b9e42c70a936d7cfc6ec\n6. https://git.kernel.org/stable/c/830f838589522404cd7c2f0f540602f25034af61\n7. https://git.kernel.org/stable/c/a37eecb705f33726f1fb7cd2a67e514a15dfe693", "creation_timestamp": "2025-01-15T14:27:23.000000Z"}