x86/kprobes: Fix crash when probing CS CALL instructionsgrafted

When using eBPF to probe CS CALL instructions within a function,
a crash can be triggered.

The eBPF tool probes offset 257 of the __hrtimer_run_queues()
function:

<__hrtimer_run_queues+249>:  nopl   0x0(%rax,%rax,1)
<__hrtimer_run_queues+254>:  mov    %r14,%rdi
<__hrtimer_run_queues+257>:  cs call <__x86_indirect_thunk_r12>
<__hrtimer_run_queues+263>:  mov    %eax,%r12d
<__hrtimer_run_queues+266>:  xchg   %ax,%ax
<__hrtimer_run_queues+268>:  mov    %r13,%rdi

Which triggers this crash:

  BUG: unable to handle page fault for address: 00000000000f41c9
  #PF: supervisor write access in kernel mode
  #PF: error_code(0x0002) - not-present page
  PGD 0 P4D 0
  Oops: 0002 [#1] SMP NOPTI
  CPU: 1 PID: 0 Comm: swapper/1 Kdump: loaded Tainted: P
  RIP: 0010:__hrtimer_run_queues+0x106/0x230

Note that __hrtimer_run_queues+0x106 is __hrtimer_run_queues+262, which is
at the 6th byte of the above CS CALL instruction. Since the CS CALL
instruction occupies 6 bytes, the exception occurred in the middle of that
call instruction.

The root cause is that when using eBPF tools to probe in the middle of a
function, a kprobe with INT3 is used as the underlying implementation.

During single-step emulation of the original CALL instruction,
int3_emulate_call() assumes that the probed CALL instruction is 5 bytes
long. However, the actual CS-prefixed CALL instruction occupies 6 bytes,
so it constructs an incorrect exception return address. When the CPU
returns from the kprobe handler, the next instruction to be executed is at
the address of the last byte of that CS CALL instruction. Coincidentally,
starting from that address, the CPU fetches and decodes a completely
different instruction, which ultimately triggers a kernel crash.

Fix the issue by using the actual instruction length obtained from
the instruction decoder when constructing the exception return
address, rather than relying on the hardcoded CALL_INSN_SIZE macro.

[ mingo: Refined the changelog ]

Fixes: 6256e668b7af ("x86/kprobes: Use int3 instead of debug trap for single-step")
Suggested-by: Masami Hiramatsu (Google) <mhiramat@kernel.org>
Signed-off-by: Jinke Han <jinkehan@didiglobal.com>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Reviewed-by: Masami Hiramatsu (Google) <mhiramat@kernel.org>
Acked-by: Yafang Shao <laoar.shao@gmail.com>
Acked-by: Borislav Petkov <bp@alien8.de>
Cc: Peter Zijlstra <peterz@infradead.org>
Link: https://patch.msgid.link/20260908073742.GA10517@didi-ThinkCentre-M920t-N000

1 files changed, 263 insertions, 0 deletions

diff --git a/drivers/clk/sprd/pll.c b/drivers/clk/sprd/pll.c
new file mode 100644
index 000000000..4aec30777
--- /dev/null
+++ b/drivers/clk/sprd/pll.c
@@ -0,0 +1,263 @@
+// SPDX-License-Identifier: GPL-2.0
+//
+// Spreadtrum pll clock driver
+//
+// Copyright (C) 2015~2017 Spreadtrum, Inc.
+// Author: Chunyan Zhang <chunyan.zhang@spreadtrum.com>
+
+#include <linux/delay.h>
+#include <linux/err.h>
+#include <linux/regmap.h>
+#include <linux/slab.h>
+
+#include "pll.h"
+
+#define CLK_PLL_1M	1000000
+#define CLK_PLL_10M	(CLK_PLL_1M * 10)
+
+#define pindex(pll, member)		\
+	(pll->factors[member].shift / (8 * sizeof(pll->regs_num)))
+
+#define pshift(pll, member)		\
+	(pll->factors[member].shift % (8 * sizeof(pll->regs_num)))
+
+#define pwidth(pll, member)		\
+	pll->factors[member].width
+
+#define pmask(pll, member)					\
+	((pwidth(pll, member)) ?				\
+	GENMASK(pwidth(pll, member) + pshift(pll, member) - 1,	\
+	pshift(pll, member)) : 0)
+
+#define pinternal(pll, cfg, member)	\
+	(cfg[pindex(pll, member)] & pmask(pll, member))
+
+#define pinternal_val(pll, cfg, member)	\
+	(pinternal(pll, cfg, member) >> pshift(pll, member))
+
+static inline unsigned int
+sprd_pll_read(const struct sprd_pll *pll, u8 index)
+{
+	const struct sprd_clk_common *common = &pll->common;
+	unsigned int val = 0;
+
+	if (WARN_ON(index >= pll->regs_num))
+		return 0;
+
+	regmap_read(common->regmap, common->reg + index * 4, &val);
+
+	return val;
+}
+
+static inline void
+sprd_pll_write(const struct sprd_pll *pll, u8 index,
+				  u32 msk, u32 val)
+{
+	const struct sprd_clk_common *common = &pll->common;
+	unsigned int offset, reg;
+	int ret = 0;
+
+	if (WARN_ON(index >= pll->regs_num))
+		return;
+
+	offset = common->reg + index * 4;
+	ret = regmap_read(common->regmap, offset, &reg);
+	if (!ret)
+		regmap_write(common->regmap, offset, (reg & ~msk) | val);
+}
+
+static unsigned long pll_get_refin(const struct sprd_pll *pll)
+{
+	u32 shift, mask, index, refin_id = 3;
+	const unsigned long refin[4] = { 2, 4, 13, 26 };
+
+	if (pwidth(pll, PLL_REFIN)) {
+		index = pindex(pll, PLL_REFIN);
+		shift = pshift(pll, PLL_REFIN);
+		mask = pmask(pll, PLL_REFIN);
+		refin_id = (sprd_pll_read(pll, index) & mask) >> shift;
+		if (refin_id > 3)
+			refin_id = 3;
+	}
+
+	return refin[refin_id];
+}
+
+static u32 pll_get_ibias(u64 rate, const u64 *table)
+{
+	u32 i, num = table[0];
+
+	/* table[0] indicates the number of items in this table */
+	for (i = 0; i < num; i++)
+		if (rate <= table[i + 1])
+			break;
+
+	return i == num ? num - 1 : i;
+}
+
+static unsigned long _sprd_pll_recalc_rate(const struct sprd_pll *pll,
+					   unsigned long parent_rate)
+{
+	u32 *cfg;
+	u32 i, mask, regs_num = pll->regs_num;
+	unsigned long rate, nint, kint = 0;
+	u64 refin;
+	u16 k1, k2;
+
+	cfg = kcalloc(regs_num, sizeof(*cfg), GFP_KERNEL);
+	if (!cfg)
+		return parent_rate;
+
+	for (i = 0; i < regs_num; i++)
+		cfg[i] = sprd_pll_read(pll, i);
+
+	refin = pll_get_refin(pll);
+
+	if (pinternal(pll, cfg, PLL_PREDIV))
+		refin = refin * 2;
+
+	if (pwidth(pll, PLL_POSTDIV) &&
+	    ((pll->fflag == 1 && pinternal(pll, cfg, PLL_POSTDIV)) ||
+	     (!pll->fflag && !pinternal(pll, cfg, PLL_POSTDIV))))
+		refin = refin / 2;
+
+	if (!pinternal(pll, cfg, PLL_DIV_S)) {
+		rate = refin * pinternal_val(pll, cfg, PLL_N) * CLK_PLL_10M;
+	} else {
+		nint = pinternal_val(pll, cfg, PLL_NINT);
+		if (pinternal(pll, cfg, PLL_SDM_EN))
+			kint = pinternal_val(pll, cfg, PLL_KINT);
+
+		mask = pmask(pll, PLL_KINT);
+
+		k1 = pll->k1;
+		k2 = pll->k2;
+		rate = DIV_ROUND_CLOSEST_ULL(refin * kint * k1,
+					 ((mask >> __ffs(mask)) + 1)) *
+					 k2 + refin * nint * CLK_PLL_1M;
+	}
+
+	kfree(cfg);
+	return rate;
+}
+
+#define SPRD_PLL_WRITE_CHECK(pll, i, mask, val)		\
+	(((sprd_pll_read(pll, i) & mask) == val) ? 0 : (-EFAULT))
+
+static int _sprd_pll_set_rate(const struct sprd_pll *pll,
+			      unsigned long rate,
+			      unsigned long parent_rate)
+{
+	struct reg_cfg *cfg;
+	int ret = 0;
+	u32 mask, shift, width, ibias_val, index;
+	u32 regs_num = pll->regs_num, i = 0;
+	unsigned long kint, nint;
+	u64 tmp, refin, fvco = rate;
+
+	cfg = kzalloc_objs(*cfg, regs_num);
+	if (!cfg)
+		return -ENOMEM;
+
+	refin = pll_get_refin(pll);
+
+	mask = pmask(pll, PLL_PREDIV);
+	index = pindex(pll, PLL_PREDIV);
+	width = pwidth(pll, PLL_PREDIV);
+	if (width && (sprd_pll_read(pll, index) & mask))
+		refin = refin * 2;
+
+	mask = pmask(pll, PLL_POSTDIV);
+	index = pindex(pll, PLL_POSTDIV);
+	width = pwidth(pll, PLL_POSTDIV);
+	cfg[index].msk = mask;
+	if (width && ((pll->fflag == 1 && fvco <= pll->fvco) ||
+		      (pll->fflag == 0 && fvco > pll->fvco)))
+		cfg[index].val |= mask;
+
+	if (width && fvco <= pll->fvco)
+		fvco = fvco * 2;
+
+	mask = pmask(pll, PLL_DIV_S);
+	index = pindex(pll, PLL_DIV_S);
+	cfg[index].val |= mask;
+	cfg[index].msk |= mask;
+
+	mask = pmask(pll, PLL_SDM_EN);
+	index = pindex(pll, PLL_SDM_EN);
+	cfg[index].val |= mask;
+	cfg[index].msk |= mask;
+
+	nint = do_div(fvco, refin * CLK_PLL_1M);
+	mask = pmask(pll, PLL_NINT);
+	index = pindex(pll, PLL_NINT);
+	shift = pshift(pll, PLL_NINT);
+	cfg[index].val |= (nint << shift) & mask;
+	cfg[index].msk |= mask;
+
+	mask = pmask(pll, PLL_KINT);
+	index = pindex(pll, PLL_KINT);
+	width = pwidth(pll, PLL_KINT);
+	shift = pshift(pll, PLL_KINT);
+	tmp = fvco - refin * nint * CLK_PLL_1M;
+	tmp = do_div(tmp, 10000) * ((mask >> shift) + 1);
+	kint = DIV_ROUND_CLOSEST_ULL(tmp, refin * 100);
+	cfg[index].val |= (kint << shift) & mask;
+	cfg[index].msk |= mask;
+
+	ibias_val = pll_get_ibias(fvco, pll->itable);
+
+	mask = pmask(pll, PLL_IBIAS);
+	index = pindex(pll, PLL_IBIAS);
+	shift = pshift(pll, PLL_IBIAS);
+	cfg[index].val |= ibias_val << shift & mask;
+	cfg[index].msk |= mask;
+
+	for (i = 0; i < regs_num; i++) {
+		if (cfg[i].msk) {
+			sprd_pll_write(pll, i, cfg[i].msk, cfg[i].val);
+			ret |= SPRD_PLL_WRITE_CHECK(pll, i, cfg[i].msk,
+						   cfg[i].val);
+		}
+	}
+
+	if (!ret)
+		udelay(pll->udelay);
+
+	kfree(cfg);
+	return ret;
+}
+
+static unsigned long sprd_pll_recalc_rate(struct clk_hw *hw,
+					  unsigned long parent_rate)
+{
+	struct sprd_pll *pll = hw_to_sprd_pll(hw);
+
+	return _sprd_pll_recalc_rate(pll, parent_rate);
+}
+
+static int sprd_pll_set_rate(struct clk_hw *hw,
+			     unsigned long rate,
+			     unsigned long parent_rate)
+{
+	struct sprd_pll *pll = hw_to_sprd_pll(hw);
+
+	return _sprd_pll_set_rate(pll, rate, parent_rate);
+}
+
+static int sprd_pll_clk_prepare(struct clk_hw *hw)
+{
+	struct sprd_pll *pll = hw_to_sprd_pll(hw);
+
+	udelay(pll->udelay);
+
+	return 0;
+}
+
+const struct clk_ops sprd_pll_ops = {
+	.prepare = sprd_pll_clk_prepare,
+	.recalc_rate = sprd_pll_recalc_rate,
+	.determine_rate = clk_determine_rate_noop,
+	.set_rate = sprd_pll_set_rate,
+};
+EXPORT_SYMBOL_GPL(sprd_pll_ops);