linux-user: Fix guest signal remapping after adjusting SIGABRT

linux-user: Implement VDSOs
 -----BEGIN PGP SIGNATURE-----
 
 iQFRBAABCgA7FiEEekgeeIaLTbaoWgXAZN846K9+IV8FAmVAHMsdHHJpY2hhcmQu
 aGVuZGVyc29uQGxpbmFyby5vcmcACgkQZN846K9+IV/GSgf/SiaCzl7FV2NsxA2h
 zHrgSYEf/4dyqjbgNhE9XSrIJ/cPEY47JrpMqJ0cK4BGc/d2IppUU0Zz3qZltXck
 CkTIPPXEWDvex+PSe5NXarxQtOazi21C+EySGFtFcCQ32C/LsdJBtNzrB+G/Tl/t
 QvPJBztXvS6FAdVci2TGBNk62nFq3NS/Uz477SD6Q/uSlczQQ5b1fu3YgZcCqM9D
 ncncHbuExUu+NMK02h8vyWwpxaTvUBSdRxx/6jnyctwVpWyMaIOfsrMooz0gBfoD
 Z7MqXhvtBYOnm4OjcQs4Nj1JBOdYoQS/y6dJ7ZP0kg10VSEwr48pduXZSvIypxbw
 hsaa8w==
 =wcWF
 -----END PGP SIGNATURE-----

Merge tag 'pull-lu-20231030' of https://gitlab.com/rth7680/qemu into staging

linux-user: Fix guest signal remapping after adjusting SIGABRT
linux-user: Implement VDSOs

* tag 'pull-lu-20231030' of https://gitlab.com/rth7680/qemu: (21 commits)
  build: Add update-linux-vdso makefile rule
  linux-user: Show vdso address in /proc/pid/maps
  linux-user/s390x: Add vdso
  linux-user/s390x: Rename __SIGNAL_FRAMESIZE to STACK_FRAME_OVERHEAD
  linux-user/ppc: Add vdso
  linux-user/loongarch64: Add vdso
  linux-user/riscv: Add vdso
  linux-user/hppa: Add vdso
  linux-user/arm: Add vdso
  linux-user/aarch64: Add vdso
  linux-user/x86_64: Add vdso
  linux-user/i386: Add vdso
  linux-user: Add gen-vdso tool
  linux-user: Load vdso image if available
  linux-user: Replace bprm->fd with bprm->src.fd
  linux-user: Use ImageSource in load_symbols
  linux-user: Use ImageSource in load_elf_image
  linux-user: Do not clobber bprm_buf swapping ehdr
  linux-user: Tidy loader_exec
  linux-user: Introduce imgsrc_read, imgsrc_read_alloc
  ...

Conflicts:
  linux-user/arm/signal.c
  Fix an #include context conflict.

Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>

Makefile

@@ -283,6 +283,13 @@ include $(SRC_PATH)/tests/vm/Makefile.include
 print-help-run = printf "  %-30s - %s\\n" "$1" "$2"
 print-help = @$(call print-help-run,$1,$2)
 
+.PHONY: update-linux-vdso
+update-linux-vdso:
+	@for m in $(SRC_PATH)/linux-user/*/Makefile.vdso; do \
+	  $(MAKE) $(SUBDIR_MAKEFLAGS) -C $$(dirname $$m) -f Makefile.vdso \
+		SRC_PATH=$(SRC_PATH) BUILD_DIR=$(BUILD_DIR); \
+	done
+
 .PHONY: help
 help:
 	@echo  'Generic targets:'
@@ -303,6 +310,9 @@ endif
 	$(call print-help,distclean,Remove all generated files)
 	$(call print-help,dist,Build a distributable tarball)
 	@echo  ''
+	@echo  'Linux-user targets:'
+	$(call print-help,update-linux-vdso,Build linux-user vdso images)
+	@echo  ''
 	@echo  'Test targets:'
 	$(call print-help,check,Run all tests (check-help for details))
 	$(call print-help,bench,Run all benchmarks)

linux-user/aarch64/Makefile.vdso

@@ -0,0 +1,15 @@
+include $(BUILD_DIR)/tests/tcg/aarch64-linux-user/config-target.mak
+
+SUBDIR = $(SRC_PATH)/linux-user/aarch64
+VPATH += $(SUBDIR)
+
+all: $(SUBDIR)/vdso-be.so $(SUBDIR)/vdso-le.so
+
+LDFLAGS = -nostdlib -shared -Wl,-h,linux-vdso.so.1 -Wl,--build-id=sha1 \
+	  -Wl,--hash-style=both -Wl,-T,$(SUBDIR)/vdso.ld
+
+$(SUBDIR)/vdso-be.so: vdso.S vdso.ld
+	$(CC) -o $@ $(LDFLAGS) -mbig-endian $<
+
+$(SUBDIR)/vdso-le.so: vdso.S vdso.ld
+	$(CC) -o $@ $(LDFLAGS) -mlittle-endian $<

linux-user/aarch64/meson.build

@@ -0,0 +1,11 @@
+# TARGET_BIG_ENDIAN is defined to 'n' for little-endian; which means it
+# is always true as far as source_set.apply() is concerned.  Always build
+# both header files and include the right one via #if.
+
+vdso_be_inc = gen_vdso.process('vdso-be.so',
+                               extra_args: ['-r', '__kernel_rt_sigreturn'])
+
+vdso_le_inc = gen_vdso.process('vdso-le.so',
+                               extra_args: ['-r', '__kernel_rt_sigreturn'])
+
+linux_user_ss.add(when: 'TARGET_AARCH64', if_true: [vdso_be_inc, vdso_le_inc])

linux-user/aarch64/vdso.S

@@ -0,0 +1,71 @@
+/*
+ * aarch64 linux replacement vdso.
+ *
+ * Copyright 2023 Linaro, Ltd.
+ *
+ * SPDX-License-Identifier: GPL-2.0-or-later
+ */
+
+#include <asm/unistd.h>
+
+/* ??? These are in include/elf.h, which is not ready for inclusion in asm. */
+#define NT_GNU_PROPERTY_TYPE_0  5
+#define GNU_PROPERTY_AARCH64_FEATURE_1_AND      0xc0000000
+#define GNU_PROPERTY_AARCH64_FEATURE_1_BTI      (1U << 0)
+#define GNU_PROPERTY_AARCH64_FEATURE_1_PAC      (1U << 1)
+
+#define GNU_PROPERTY_AARCH64_FEATURE_1_DEFAULT \
+    (GNU_PROPERTY_AARCH64_FEATURE_1_BTI | GNU_PROPERTY_AARCH64_FEATURE_1_PAC)
+
+	.section .note.gnu.property
+	.align	3
+	.long	2f - 1f
+	.long	6f - 3f
+	.long	NT_GNU_PROPERTY_TYPE_0
+1:	.string	"GNU"
+2:	.align	3
+3:	.long	GNU_PROPERTY_AARCH64_FEATURE_1_AND
+	.long	5f - 4f
+4:	.long	GNU_PROPERTY_AARCH64_FEATURE_1_DEFAULT
+5:	.align	3
+6:
+
+	.text
+
+.macro endf name
+	.globl	\name
+	.type	\name, @function
+	.size	\name, . - \name
+.endm
+
+.macro vdso_syscall name, nr
+\name:
+	bti	c
+	mov	x8, #\nr
+	svc	#0
+	ret
+endf	\name
+.endm
+
+	.cfi_startproc
+
+vdso_syscall __kernel_gettimeofday, __NR_gettimeofday
+vdso_syscall __kernel_clock_gettime, __NR_clock_gettime
+vdso_syscall __kernel_clock_getres, __NR_clock_getres
+
+	.cfi_endproc
+
+
+/*
+ * TODO: The kernel makes a big deal of turning off the .cfi directives,
+ * because they cause libgcc to crash, but that's because they're wrong.
+ *
+ * For now, elide the unwind info for __kernel_rt_sigreturn and rely on
+ * the libgcc fallback routine as we have always done.  This requires
+ * that the code sequence used be exact.
+ */
+__kernel_rt_sigreturn:
+	/* No BTI C insn here -- we arrive via RET. */
+	mov	x8, #__NR_rt_sigreturn
+	svc	#0
+endf	__kernel_rt_sigreturn

linux-user/aarch64/vdso.ld

@@ -0,0 +1,72 @@
+/*
+ * Linker script for linux aarch64 replacement vdso.
+ *
+ * Copyright 2021 Linaro, Ltd.
+ *
+ * SPDX-License-Identifier: GPL-2.0-or-later
+ */
+
+VERSION {
+        LINUX_2.6.39 {
+        global:
+                __kernel_rt_sigreturn;
+                __kernel_gettimeofday;
+                __kernel_clock_gettime;
+                __kernel_clock_getres;
+
+        local: *;
+        };
+}
+
+
+PHDRS {
+        phdr            PT_PHDR         FLAGS(4) PHDRS;
+        load            PT_LOAD         FLAGS(7) FILEHDR PHDRS;
+        dynamic         PT_DYNAMIC      FLAGS(4);
+        eh_frame_hdr    PT_GNU_EH_FRAME;
+        note            PT_NOTE         FLAGS(4);
+}
+
+SECTIONS {
+        /*
+         * We can't prelink to any address without knowing something about
+         * the virtual memory space of the host, since that leaks over into
+         * the available memory space of the guest.
+         */
+        . = SIZEOF_HEADERS;
+
+        /*
+         * The following, including the FILEHDRS and PHDRS, are modified
+         * when we relocate the binary.  We want them to be initially
+         * writable for the relocation; we'll force them read-only after.
+         */
+        .note           : { *(.note*) }         :load :note
+        .dynamic        : { *(.dynamic) }       :load :dynamic
+        .dynsym         : { *(.dynsym) }        :load
+        /*
+         * There ought not be any real read-write data.
+         * But since we manipulated the segment layout,
+         * we have to put these sections somewhere.
+         */
+        .data           : {
+                *(.data*)
+                *(.sdata*)
+                *(.got.plt) *(.got)
+                *(.gnu.linkonce.d.*)
+                *(.bss*)
+                *(.dynbss*)
+                *(.gnu.linkonce.b.*)
+        }
+
+        .rodata         : { *(.rodata*) }
+        .hash           : { *(.hash) }
+        .gnu.hash       : { *(.gnu.hash) }
+        .dynstr         : { *(.dynstr) }
+        .gnu.version    : { *(.gnu.version) }
+        .gnu.version_d  : { *(.gnu.version_d) }
+        .gnu.version_r  : { *(.gnu.version_r) }
+        .eh_frame_hdr   : { *(.eh_frame_hdr) }  :load :eh_frame_hdr
+        .eh_frame       : { *(.eh_frame) }      :load
+
+        .text           : { *(.text*) }         :load   =0xd503201f
+}

linux-user/arm/Makefile.vdso

@@ -0,0 +1,17 @@
+include $(BUILD_DIR)/tests/tcg/arm-linux-user/config-target.mak
+
+SUBDIR = $(SRC_PATH)/linux-user/arm
+VPATH += $(SUBDIR)
+
+all: $(SUBDIR)/vdso-be.so $(SUBDIR)/vdso-le.so
+
+# Adding -use-blx disables unneeded interworking without actually using blx.
+LDFLAGS = -nostdlib -shared -Wl,-use-blx \
+	  -Wl,-h,linux-vdso.so.1 -Wl,--build-id=sha1 \
+	  -Wl,--hash-style=both -Wl,-T,$(SUBDIR)/vdso.ld
+
+$(SUBDIR)/vdso-be.so: vdso.S vdso.ld vdso-asmoffset.h
+	$(CC) -o $@ $(LDFLAGS) -mbig-endian $<
+
+$(SUBDIR)/vdso-le.so: vdso.S vdso.ld vdso-asmoffset.h
+	$(CC) -o $@ $(LDFLAGS) -mlittle-endian $<

linux-user/arm/meson.build

@@ -5,3 +5,15 @@ syscall_nr_generators += {
                    arguments: [ meson.current_source_dir() / 'syscallhdr.sh', '@INPUT@', '@OUTPUT@', '@EXTRA_ARGS@' ],
                    output: '@BASENAME@_nr.h')
 }
+
+# TARGET_BIG_ENDIAN is defined to 'n' for little-endian; which means it
+# is always true as far as source_set.apply() is concerned.  Always build
+# both header files and include the right one via #if.
+
+vdso_be_inc = gen_vdso.process('vdso-be.so',
+                               extra_args: ['-s', 'sigreturn_codes'])
+
+vdso_le_inc = gen_vdso.process('vdso-le.so',
+                               extra_args: ['-s', 'sigreturn_codes'])
+
+linux_user_ss.add(when: 'TARGET_ARM', if_true: [vdso_be_inc, vdso_le_inc])

linux-user/arm/signal.c

@@ -22,6 +22,7 @@
 #include "signal-common.h"
 #include "linux-user/trace.h"
 #include "target/arm/cpu-features.h"
+#include "vdso-asmoffset.h"
 
 struct target_sigcontext {
     abi_ulong trap_no;
@@ -103,6 +104,11 @@ struct rt_sigframe
     struct sigframe sig;
 };
 
+QEMU_BUILD_BUG_ON(offsetof(struct sigframe, retcode[3])
+                  != SIGFRAME_RC3_OFFSET);
+QEMU_BUILD_BUG_ON(offsetof(struct rt_sigframe, sig.retcode[3])
+                  != RT_SIGFRAME_RC3_OFFSET);
+
 static abi_ptr sigreturn_fdpic_tramp;
 
 /*
@@ -161,6 +167,9 @@ get_sigframe(struct target_sigaction *ka, CPUARMState *regs, int framesize)
     return (sp - framesize) & ~7;
 }
 
+static void write_arm_sigreturn(uint32_t *rc, int syscall);
+static void write_arm_fdpic_sigreturn(uint32_t *rc, int ofs);
+
 static int
 setup_return(CPUARMState *env, struct target_sigaction *ka, int usig,
              struct sigframe *frame, abi_ulong sp_addr)
@@ -168,9 +177,9 @@ setup_return(CPUARMState *env, struct target_sigaction *ka, int usig,
     abi_ulong handler = 0;
     abi_ulong handler_fdpic_GOT = 0;
     abi_ulong retcode;
-    int thumb, retcode_idx;
+    bool is_fdpic = info_is_fdpic(((TaskState *)thread_cpu->opaque)->info);
-    int is_fdpic = info_is_fdpic(((TaskState *)thread_cpu->opaque)->info);
+    bool is_rt = ka->sa_flags & TARGET_SA_SIGINFO;
-    bool copy_retcode;
+    bool thumb;
 
     if (is_fdpic) {
         /* In FDPIC mode, ka->_sa_handler points to a function
@@ -185,9 +194,7 @@ setup_return(CPUARMState *env, struct target_sigaction *ka, int usig,
     } else {
         handler = ka->_sa_handler;
     }
-
     thumb = handler & 1;
-    retcode_idx = thumb + (ka->sa_flags & TARGET_SA_SIGINFO ? 2 : 0);
 
     uint32_t cpsr = cpsr_read(env);
 
@@ -203,24 +210,32 @@ setup_return(CPUARMState *env, struct target_sigaction *ka, int usig,
         cpsr &= ~CPSR_E;
     }
 
-    if (ka->sa_flags & TARGET_SA_RESTORER) {
+    /* Our vdso default_sigreturn label is a table of entry points. */
-        if (is_fdpic) {
+    retcode = default_sigreturn + (is_fdpic * 2 + is_rt) * 8;
-            __put_user((abi_ulong)ka->sa_restorer, &frame->retcode[3]);
+
-            retcode = (sigreturn_fdpic_tramp +
+    /*
-                       retcode_idx * RETCODE_BYTES + thumb);
+     * Put the sigreturn code on the stack no matter which return
-            copy_retcode = true;
+     * mechanism we use in order to remain ABI compliant.
-        } else {
+     * Because this is about ABI, always use the A32 instructions,
-            retcode = ka->sa_restorer;
+     * despite the fact that our actual vdso trampoline is T16.
-            copy_retcode = false;
+     */
-        }
+    if (is_fdpic) {
+        write_arm_fdpic_sigreturn(frame->retcode,
+                                  is_rt ? RT_SIGFRAME_RC3_OFFSET
+                                        : SIGFRAME_RC3_OFFSET);
     } else {
-        retcode = default_sigreturn + retcode_idx * RETCODE_BYTES + thumb;
+        write_arm_sigreturn(frame->retcode,
-        copy_retcode = true;
+                            is_rt ? TARGET_NR_rt_sigreturn
+                                  : TARGET_NR_sigreturn);
     }
 
-    /* Copy the code to the stack slot for ABI compatibility. */
+    if (ka->sa_flags & TARGET_SA_RESTORER) {
-    if (copy_retcode) {
+        if (is_fdpic) {
-        memcpy(frame->retcode, g2h_untagged(retcode & ~1), RETCODE_BYTES);
+            /* Place the function descriptor in slot 3. */
+            __put_user((abi_ulong)ka->sa_restorer, &frame->retcode[3]);
+        } else {
+            retcode = ka->sa_restorer;
+        }
     }
 
     env->regs[0] = usig;

linux-user/arm/vdso-asmoffset.h

@@ -0,0 +1,3 @@
+/* offsetof(struct sigframe, retcode[3]) */
+#define SIGFRAME_RC3_OFFSET     756
+#define RT_SIGFRAME_RC3_OFFSET  884

linux-user/arm/vdso.S

@@ -0,0 +1,174 @@
+/*
+ * arm linux replacement vdso.
+ *
+ * Copyright 2023 Linaro, Ltd.
+ *
+ * SPDX-License-Identifier: GPL-2.0-or-later
+ */
+
+#include <asm/unistd.h>
+#include "vdso-asmoffset.h"
+
+/*
+ * All supported cpus have T16 instructions: at least arm4t.
+ *
+ * We support user-user with m-profile cpus as an extension, because it
+ * is useful for testing gcc, which requires we avoid A32 instructions.
+ */
+	.thumb
+	.arch	armv4t
+	.eabi_attribute Tag_FP_arch, 0
+	.eabi_attribute Tag_ARM_ISA_use, 0
+
+	.text
+
+.macro	raw_syscall n
+	.ifne \n < 0x100
+	mov	r7, #\n
+	.elseif \n < 0x1ff
+	mov	r7, #0xff
+	add	r7, #(\n - 0xff)
+	.else
+	.err
+	.endif
+	swi	#0
+.endm
+
+.macro	fdpic_thunk ofs
+	ldr	r3, [sp, #\ofs]
+	ldmia	r2, {r2, r3}
+	mov	r9, r3
+	bx	r2
+.endm
+
+.macro	endf name
+	.globl	\name
+	.type	\name, %function
+	.size	\name, . - \name
+.endm
+
+/*
+ * We must save/restore r7 for the EABI syscall number.
+ * While we're doing that, we might as well save LR to get a free return,
+ * and a branch that is interworking back to ARMv5.
+ */
+
+.macro SYSCALL name, nr
+\name:
+	.cfi_startproc
+	push	{r7, lr}
+	.cfi_adjust_cfa_offset 8
+	.cfi_offset r7, -8
+	.cfi_offset lr, -4
+	raw_syscall \nr
+	pop	{r7, pc}
+	.cfi_endproc
+endf \name
+.endm
+
+SYSCALL	__vdso_clock_gettime, __NR_clock_gettime
+SYSCALL __vdso_clock_gettime64, __NR_clock_gettime64
+SYSCALL __vdso_clock_getres, __NR_clock_getres
+SYSCALL __vdso_gettimeofday, __NR_gettimeofday
+
+
+/*
+ * We, like the real kernel, use a table of sigreturn trampolines.
+ * Unlike the real kernel, we do not attempt to pack this into as
+ * few bytes as possible -- simply use 8 bytes per slot.
+ *
+ * Within each slot, use the exact same code sequence as the kernel,
+ * lest we trip up someone doing code inspection.
+ */
+
+.macro	slot n
+	.balign	8
+	.org	sigreturn_codes + 8 * \n
+.endm
+
+.macro cfi_fdpic_r9 ofs
+	/*
+	 * fd = *(r13 + ofs)
+         * r9 = *(fd + 4)
+	 *
+	 * DW_CFA_expression r9, length (7),
+	 *   DW_OP_breg13, ofs, DW_OP_deref,
+	 *   DW_OP_plus_uconst, 4, DW_OP_deref
+         */
+	.cfi_escape 0x10, 9, 7, 0x7d, (\ofs & 0x7f) + 0x80, (\ofs >> 7), 0x06, 0x23, 4, 0x06
+.endm
+
+.macro cfi_fdpic_pc ofs
+	/*
+	 * fd = *(r13 + ofs)
+         * pc = *fd
+	 *
+	 * DW_CFA_expression lr (14), length (5),
+	 *   DW_OP_breg13, ofs, DW_OP_deref, DW_OP_deref
+         */
+	.cfi_escape 0x10, 14, 5, 0x7d, (\ofs & 0x7f) + 0x80, (\ofs >> 7), 0x06, 0x06
+.endm
+
+/*
+ * Start the unwind info at least one instruction before the signal
+ * trampoline, because the unwinder will assume we are returning
+ * after a call site.
+ */
+	.cfi_startproc simple
+	.cfi_signal_frame
+	.cfi_return_column 15
+
+	.cfi_def_cfa	sp, 32 + 64
+	.cfi_offset	r0, -16 * 4
+	.cfi_offset	r1, -15 * 4
+	.cfi_offset	r2, -14 * 4
+	.cfi_offset	r3, -13 * 4
+	.cfi_offset	r4, -12 * 4
+	.cfi_offset	r5, -11 * 4
+	.cfi_offset	r6, -10 * 4
+	.cfi_offset	r7, -9 * 4
+	.cfi_offset	r8, -8 * 4
+	.cfi_offset	r9, -7 * 4
+	.cfi_offset	r10, -6 * 4
+	.cfi_offset	r11, -5 * 4
+	.cfi_offset	r12, -4 * 4
+	.cfi_offset	r13, -3 * 4
+	.cfi_offset	r14, -2 * 4
+	.cfi_offset	r15, -1 * 4
+
+	nop
+
+	.balign	16
+sigreturn_codes:
+	/* [EO]ABI sigreturn */
+	slot	0
+	raw_syscall __NR_sigreturn
+
+	.cfi_def_cfa_offset 160 + 64
+
+	/* [EO]ABI rt_sigreturn */
+	slot	1
+	raw_syscall __NR_rt_sigreturn
+
+	.cfi_endproc
+
+	/* FDPIC sigreturn */
+	.cfi_startproc
+	cfi_fdpic_pc SIGFRAME_RC3_OFFSET
+	cfi_fdpic_r9 SIGFRAME_RC3_OFFSET
+
+	slot	2
+	fdpic_thunk SIGFRAME_RC3_OFFSET
+	.cfi_endproc
+
+	/* FDPIC rt_sigreturn */
+	.cfi_startproc
+	cfi_fdpic_pc RT_SIGFRAME_RC3_OFFSET
+	cfi_fdpic_r9 RT_SIGFRAME_RC3_OFFSET
+
+	slot	3
+	fdpic_thunk RT_SIGFRAME_RC3_OFFSET
+	.cfi_endproc
+
+	.balign	16
+endf sigreturn_codes

linux-user/arm/vdso.ld

@@ -0,0 +1,67 @@
+/*
+ * Linker script for linux arm replacement vdso.
+ *
+ * Copyright 2023 Linaro, Ltd.
+ *
+ * SPDX-License-Identifier: GPL-2.0-or-later
+ */
+
+VERSION {
+        LINUX_2.6 {
+        global:
+                __vdso_clock_gettime;
+                __vdso_gettimeofday;
+                __vdso_clock_getres;
+                __vdso_clock_gettime64;
+
+        local: *;
+        };
+}
+
+
+PHDRS {
+        phdr            PT_PHDR         FLAGS(4) PHDRS;
+        load            PT_LOAD         FLAGS(7) FILEHDR PHDRS; /* FLAGS=RWX */
+        dynamic         PT_DYNAMIC      FLAGS(4);
+        eh_frame_hdr    PT_GNU_EH_FRAME;
+        note            PT_NOTE         FLAGS(4);
+}
+
+SECTIONS {
+        . = SIZEOF_HEADERS;
+
+        /*
+         * The following, including the FILEHDRS and PHDRS, are modified
+         * when we relocate the binary.  We want them to be initially
+         * writable for the relocation; we'll force them read-only after.
+         */
+        .note           : { *(.note*) }         :load :note
+        .dynamic        : { *(.dynamic) }       :load :dynamic
+        .dynsym         : { *(.dynsym) }        :load
+        /*
+         * There ought not be any real read-write data.
+         * But since we manipulated the segment layout,
+         * we have to put these sections somewhere.
+         */
+        .data           : {
+                *(.data*)
+                *(.sdata*)
+                *(.got.plt) *(.got)
+                *(.gnu.linkonce.d.*)
+                *(.bss*)
+                *(.dynbss*)
+                *(.gnu.linkonce.b.*)
+        }
+
+        .rodata         : { *(.rodata*) }
+        .hash           : { *(.hash) }
+        .gnu.hash       : { *(.gnu.hash) }
+        .dynstr         : { *(.dynstr) }
+        .gnu.version    : { *(.gnu.version) }
+        .gnu.version_d  : { *(.gnu.version_d) }
+        .gnu.version_r  : { *(.gnu.version_r) }
+        .eh_frame_hdr   : { *(.eh_frame_hdr) }  :load :eh_frame_hdr
+        .eh_frame       : { *(.eh_frame) }      :load
+
+        .text           : { *(.text*) }         :load
+}

linux-user/elfload.c

@@ -37,6 +37,19 @@
 #undef ELF_ARCH
 #endif
 
+#ifndef TARGET_ARCH_HAS_SIGTRAMP_PAGE
+#define TARGET_ARCH_HAS_SIGTRAMP_PAGE 0
+#endif
+
+typedef struct {
+    const uint8_t *image;
+    const uint32_t *relocs;
+    unsigned image_size;
+    unsigned reloc_count;
+    unsigned sigreturn_ofs;
+    unsigned rt_sigreturn_ofs;
+} VdsoImageInfo;
+
 #define ELF_OSABI   ELFOSABI_SYSV
 
 /* from personality.h */
@@ -296,12 +309,27 @@ static void elf_core_copy_regs(target_elf_gregset_t *regs, const CPUX86State *en
     (*regs)[15] = tswapreg(env->regs[R_ESP]);
     (*regs)[16] = tswapreg(env->segs[R_SS].selector & 0xffff);
 }
-#endif
+
+/*
+ * i386 is the only target which supplies AT_SYSINFO for the vdso.
+ * All others only supply AT_SYSINFO_EHDR.
+ */
+#define DLINFO_ARCH_ITEMS (vdso_info != NULL)
+#define ARCH_DLINFO                                     \
+    do {                                                \
+        if (vdso_info) {                                \
+            NEW_AUX_ENT(AT_SYSINFO, vdso_info->entry);  \
+        }                                               \
+    } while (0)
+
+#endif /* TARGET_X86_64 */
+
+#define VDSO_HEADER "vdso.c.inc"
 
 #define USE_ELF_CORE_DUMP
 #define ELF_EXEC_PAGESIZE       4096
 
-#endif
+#endif /* TARGET_I386 */
 
 #ifdef TARGET_ARM
 
@@ -921,6 +949,13 @@ const char *elf_hwcap2_str(uint32_t bit)
 #undef GET_FEATURE_ID
 
 #endif /* not TARGET_AARCH64 */
+
+#if TARGET_BIG_ENDIAN
+# define VDSO_HEADER  "vdso-be.c.inc"
+#else
+# define VDSO_HEADER  "vdso-le.c.inc"
+#endif
+
 #endif /* TARGET_ARM */
 
 #ifdef TARGET_SPARC
@@ -1156,6 +1191,14 @@ static void elf_core_copy_regs(target_elf_gregset_t *regs, const CPUPPCState *en
 #define USE_ELF_CORE_DUMP
 #define ELF_EXEC_PAGESIZE       4096
 
+#ifndef TARGET_PPC64
+# define VDSO_HEADER  "vdso-32.c.inc"
+#elif TARGET_BIG_ENDIAN
+# define VDSO_HEADER  "vdso-64.c.inc"
+#else
+# define VDSO_HEADER  "vdso-64le.c.inc"
+#endif
+
 #endif
 
 #ifdef TARGET_LOONGARCH64
@@ -1166,6 +1209,8 @@ static void elf_core_copy_regs(target_elf_gregset_t *regs, const CPUPPCState *en
 
 #define elf_check_arch(x) ((x) == EM_LOONGARCH)
 
+#define VDSO_HEADER "vdso.c.inc"
+
 static inline void init_thread(struct target_pt_regs *regs,
                                struct image_info *infop)
 {
@@ -1853,6 +1898,8 @@ static void elf_core_copy_regs(target_elf_gregset_t *regs,
 #define USE_ELF_CORE_DUMP
 #define ELF_EXEC_PAGESIZE 4096
 
+#define VDSO_HEADER "vdso.c.inc"
+
 #endif /* TARGET_S390X */
 
 #ifdef TARGET_RISCV
@@ -1861,8 +1908,10 @@ static void elf_core_copy_regs(target_elf_gregset_t *regs,
 
 #ifdef TARGET_RISCV32
 #define ELF_CLASS ELFCLASS32
+#define VDSO_HEADER "vdso-32.c.inc"
 #else
 #define ELF_CLASS ELFCLASS64
+#define VDSO_HEADER "vdso-64.c.inc"
 #endif
 
 #define ELF_HWCAP get_elf_hwcap()
@@ -1898,6 +1947,8 @@ static inline void init_thread(struct target_pt_regs *regs,
 #define STACK_GROWS_DOWN 0
 #define STACK_ALIGNMENT  64
 
+#define VDSO_HEADER "vdso.c.inc"
+
 static inline void init_thread(struct target_pt_regs *regs,
                                struct image_info *infop)
 {
@@ -2205,7 +2256,8 @@ static inline void bswap_mips_abiflags(Mips_elf_abiflags_v0 *abiflags) { }
 #ifdef USE_ELF_CORE_DUMP
 static int elf_core_dump(int, const CPUArchState *);
 #endif /* USE_ELF_CORE_DUMP */
-static void load_symbols(struct elfhdr *hdr, int fd, abi_ulong load_bias);
+static void load_symbols(struct elfhdr *hdr, const ImageSource *src,
+                         abi_ulong load_bias);
 
 /* Verify the portions of EHDR within E_IDENT for the target.
    This can be performed before bswapping the entire header.  */
@@ -2474,7 +2526,8 @@ static abi_ulong loader_build_fdpic_loadmap(struct image_info *info, abi_ulong s
 static abi_ulong create_elf_tables(abi_ulong p, int argc, int envc,
                                    struct elfhdr *exec,
                                    struct image_info *info,
-                                   struct image_info *interp_info)
+                                   struct image_info *interp_info,
+                                   struct image_info *vdso_info)
 {
     abi_ulong sp;
     abi_ulong u_argc, u_argv, u_envp, u_auxv;
@@ -2562,10 +2615,15 @@ static abi_ulong create_elf_tables(abi_ulong p, int argc, int envc,
     }
 
     size = (DLINFO_ITEMS + 1) * 2;
-    if (k_base_platform)
+    if (k_base_platform) {
         size += 2;
-    if (k_platform)
+    }
+    if (k_platform) {
         size += 2;
+    }
+    if (vdso_info) {
+        size += 2;
+    }
 #ifdef DLINFO_ARCH_ITEMS
     size += DLINFO_ARCH_ITEMS * 2;
 #endif
@@ -2647,6 +2705,9 @@ static abi_ulong create_elf_tables(abi_ulong p, int argc, int envc,
     if (u_platform) {
         NEW_AUX_ENT(AT_PLATFORM, u_platform);
     }
+    if (vdso_info) {
+        NEW_AUX_ENT(AT_SYSINFO_EHDR, vdso_info->load_addr);
+    }
     NEW_AUX_ENT (AT_NULL, 0);
 #undef NEW_AUX_ENT
 
@@ -3106,10 +3167,9 @@ static bool parse_elf_property(const uint32_t *data, int *off, int datasz,
 }
 
 /* Process NT_GNU_PROPERTY_TYPE_0. */
-static bool parse_elf_properties(int image_fd,
+static bool parse_elf_properties(const ImageSource *src,
                                  struct image_info *info,
                                  const struct elf_phdr *phdr,
-                                 char bprm_buf[BPRM_BUF_SIZE],
                                  Error **errp)
 {
     union {
@@ -3137,14 +3197,8 @@ static bool parse_elf_properties(int image_fd,
         return false;
     }
 
-    if (phdr->p_offset + n <= BPRM_BUF_SIZE) {
+    if (!imgsrc_read(&note, phdr->p_offset, n, src, errp)) {
-        memcpy(&note, bprm_buf + phdr->p_offset, n);
+        return false;
-    } else {
-        ssize_t len = pread(image_fd, &note, n, phdr->p_offset);
-        if (len != n) {
-            error_setg_errno(errp, errno, "Error reading file header");
-            return false;
-        }
     }
 
     /*
@@ -3190,29 +3244,34 @@ static bool parse_elf_properties(int image_fd,
     }
 }
 
-/* Load an ELF image into the address space.
+/**
+ * load_elf_image: Load an ELF image into the address space.
+ * @image_name: the filename of the image, to use in error messages.
+ * @src: the ImageSource from which to read.
+ * @info: info collected from the loaded image.
+ * @ehdr: the ELF header, not yet bswapped.
+ * @pinterp_name: record any PT_INTERP string found.
+ *
+ * On return: @info values will be filled in, as necessary or available.
+ */
 
-   IMAGE_NAME is the filename of the image, to use in error messages.
+static void load_elf_image(const char *image_name, const ImageSource *src,
-   IMAGE_FD is the open file descriptor for the image.
+                           struct image_info *info, struct elfhdr *ehdr,
-
+                           char **pinterp_name)
-   BPRM_BUF is a copy of the beginning of the file; this of course
-   contains the elf file header at offset 0.  It is assumed that this
-   buffer is sufficiently aligned to present no problems to the host
-   in accessing data at aligned offsets within the buffer.
-
-   On return: INFO values will be filled in, as necessary or available.  */
-
-static void load_elf_image(const char *image_name, int image_fd,
-                           struct image_info *info, char **pinterp_name,
-                           char bprm_buf[BPRM_BUF_SIZE])
 {
-    struct elfhdr *ehdr = (struct elfhdr *)bprm_buf;
+    g_autofree struct elf_phdr *phdr = NULL;
-    struct elf_phdr *phdr;
     abi_ulong load_addr, load_bias, loaddr, hiaddr, error;
-    int i, retval, prot_exec;
+    int i, prot_exec;
     Error *err = NULL;
 
-    /* First of all, some simple consistency checks */
+    /*
+     * First of all, some simple consistency checks.
+     * Note that we rely on the bswapped ehdr staying in bprm_buf,
+     * for later use by load_elf_binary and create_elf_tables.
+     */
+    if (!imgsrc_read(ehdr, 0, sizeof(*ehdr), src, &err)) {
+        goto exit_errmsg;
+    }
     if (!elf_check_ident(ehdr)) {
         error_setg(&err, "Invalid ELF image for this architecture");
         goto exit_errmsg;
@@ -3223,15 +3282,11 @@ static void load_elf_image(const char *image_name, int image_fd,
         goto exit_errmsg;
     }
 
-    i = ehdr->e_phnum * sizeof(struct elf_phdr);
+    phdr = imgsrc_read_alloc(ehdr->e_phoff,
-    if (ehdr->e_phoff + i <= BPRM_BUF_SIZE) {
+                             ehdr->e_phnum * sizeof(struct elf_phdr),
-        phdr = (struct elf_phdr *)(bprm_buf + ehdr->e_phoff);
+                             src, &err);
-    } else {
+    if (phdr == NULL) {
-        phdr = (struct elf_phdr *) alloca(i);
+        goto exit_errmsg;
-        retval = pread(image_fd, phdr, i, ehdr->e_phoff);
-        if (retval != i) {
-            goto exit_read;
-        }
     }
     bswap_phdr(phdr, ehdr->e_phnum);
 
@@ -3268,17 +3323,10 @@ static void load_elf_image(const char *image_name, int image_fd,
                 goto exit_errmsg;
             }
 
-            interp_name = g_malloc(eppnt->p_filesz);
+            interp_name = imgsrc_read_alloc(eppnt->p_offset, eppnt->p_filesz,
-
+                                            src, &err);
-            if (eppnt->p_offset + eppnt->p_filesz <= BPRM_BUF_SIZE) {
+            if (interp_name == NULL) {
-                memcpy(interp_name, bprm_buf + eppnt->p_offset,
+                goto exit_errmsg;
-                       eppnt->p_filesz);
-            } else {
-                retval = pread(image_fd, interp_name, eppnt->p_filesz,
-                               eppnt->p_offset);
-                if (retval != eppnt->p_filesz) {
-                    goto exit_read;
-                }
             }
             if (interp_name[eppnt->p_filesz - 1] != 0) {
                 error_setg(&err, "Invalid PT_INTERP entry");
@@ -3286,7 +3334,7 @@ static void load_elf_image(const char *image_name, int image_fd,
             }
             *pinterp_name = g_steal_pointer(&interp_name);
         } else if (eppnt->p_type == PT_GNU_PROPERTY) {
-            if (!parse_elf_properties(image_fd, info, eppnt, bprm_buf, &err)) {
+            if (!parse_elf_properties(src, info, eppnt, &err)) {
                 goto exit_errmsg;
             }
         } else if (eppnt->p_type == PT_GNU_STACK) {
@@ -3439,9 +3487,9 @@ static void load_elf_image(const char *image_name, int image_fd,
              * but no backing file segment.
              */
             if (eppnt->p_filesz != 0) {
-                error = target_mmap(vaddr_ps, eppnt->p_filesz + vaddr_po,
+                error = imgsrc_mmap(vaddr_ps, eppnt->p_filesz + vaddr_po,
                                     elf_prot, MAP_PRIVATE | MAP_FIXED,
-                                    image_fd, eppnt->p_offset - vaddr_po);
+                                    src, eppnt->p_offset - vaddr_po);
                 if (error == -1) {
                     goto exit_mmap;
                 }
@@ -3473,20 +3521,11 @@ static void load_elf_image(const char *image_name, int image_fd,
 #ifdef TARGET_MIPS
         } else if (eppnt->p_type == PT_MIPS_ABIFLAGS) {
             Mips_elf_abiflags_v0 abiflags;
-            if (eppnt->p_filesz < sizeof(Mips_elf_abiflags_v0)) {
+
-                error_setg(&err, "Invalid PT_MIPS_ABIFLAGS entry");
+            if (!imgsrc_read(&abiflags, eppnt->p_offset, sizeof(abiflags),
+                             src, &err)) {
                 goto exit_errmsg;
             }
-            if (eppnt->p_offset + eppnt->p_filesz <= BPRM_BUF_SIZE) {
-                memcpy(&abiflags, bprm_buf + eppnt->p_offset,
-                       sizeof(Mips_elf_abiflags_v0));
-            } else {
-                retval = pread(image_fd, &abiflags, sizeof(Mips_elf_abiflags_v0),
-                               eppnt->p_offset);
-                if (retval != sizeof(Mips_elf_abiflags_v0)) {
-                    goto exit_read;
-                }
-            }
             bswap_mips_abiflags(&abiflags);
             info->fp_abi = abiflags.fp_abi;
 #endif
@@ -3499,23 +3538,16 @@ static void load_elf_image(const char *image_name, int image_fd,
     }
 
     if (qemu_log_enabled()) {
-        load_symbols(ehdr, image_fd, load_bias);
+        load_symbols(ehdr, src, load_bias);
     }
 
-    debuginfo_report_elf(image_name, image_fd, load_bias);
+    debuginfo_report_elf(image_name, src->fd, load_bias);
 
     mmap_unlock();
 
-    close(image_fd);
+    close(src->fd);
     return;
 
- exit_read:
-    if (retval >= 0) {
-        error_setg(&err, "Incomplete read of file header");
-    } else {
-        error_setg_errno(&err, errno, "Error reading file header");
-    }
-    goto exit_errmsg;
  exit_mmap:
     error_setg_errno(&err, errno, "Error mapping file");
     goto exit_errmsg;
@@ -3527,6 +3559,8 @@ static void load_elf_image(const char *image_name, int image_fd,
 static void load_elf_interp(const char *filename, struct image_info *info,
                             char bprm_buf[BPRM_BUF_SIZE])
 {
+    struct elfhdr ehdr;
+    ImageSource src;
     int fd, retval;
     Error *err = NULL;
 
@@ -3544,11 +3578,57 @@ static void load_elf_interp(const char *filename, struct image_info *info,
         exit(-1);
     }
 
-    if (retval < BPRM_BUF_SIZE) {
+    src.fd = fd;
-        memset(bprm_buf + retval, 0, BPRM_BUF_SIZE - retval);
+    src.cache = bprm_buf;
+    src.cache_size = retval;
+
+    load_elf_image(filename, &src, info, &ehdr, NULL);
+}
+
+#ifdef VDSO_HEADER
+#include VDSO_HEADER
+#define  vdso_image_info()  &vdso_image_info
+#else
+#define  vdso_image_info()  NULL
+#endif
+
+static void load_elf_vdso(struct image_info *info, const VdsoImageInfo *vdso)
+{
+    ImageSource src;
+    struct elfhdr ehdr;
+    abi_ulong load_bias, load_addr;
+
+    src.fd = -1;
+    src.cache = vdso->image;
+    src.cache_size = vdso->image_size;
+
+    load_elf_image("<internal-vdso>", &src, info, &ehdr, NULL);
+    load_addr = info->load_addr;
+    load_bias = info->load_bias;
+
+    /*
+     * We need to relocate the VDSO image.  The one built into the kernel
+     * is built for a fixed address.  The one built for QEMU is not, since
+     * that requires close control of the guest address space.
+     * We pre-processed the image to locate all of the addresses that need
+     * to be updated.
+     */
+    for (unsigned i = 0, n = vdso->reloc_count; i < n; i++) {
+        abi_ulong *addr = g2h_untagged(load_addr + vdso->relocs[i]);
+        *addr = tswapal(tswapal(*addr) + load_bias);
     }
 
-    load_elf_image(filename, fd, info, NULL, bprm_buf);
+    /* Install signal trampolines, if present. */
+    if (vdso->sigreturn_ofs) {
+        default_sigreturn = load_addr + vdso->sigreturn_ofs;
+    }
+    if (vdso->rt_sigreturn_ofs) {
+        default_rt_sigreturn = load_addr + vdso->rt_sigreturn_ofs;
+    }
+
+    /* Remove write from VDSO segment. */
+    target_mprotect(info->start_data, info->end_data - info->start_data,
+                    PROT_READ | PROT_EXEC);
 }
 
 static int symfind(const void *s0, const void *s1)
@@ -3595,19 +3675,20 @@ static int symcmp(const void *s0, const void *s1)
 }
 
 /* Best attempt to load symbols from this ELF object. */
-static void load_symbols(struct elfhdr *hdr, int fd, abi_ulong load_bias)
+static void load_symbols(struct elfhdr *hdr, const ImageSource *src,
+                         abi_ulong load_bias)
 {
     int i, shnum, nsyms, sym_idx = 0, str_idx = 0;
-    uint64_t segsz;
+    g_autofree struct elf_shdr *shdr = NULL;
-    struct elf_shdr *shdr;
     char *strings = NULL;
-    struct syminfo *s = NULL;
+    struct elf_sym *syms = NULL;
-    struct elf_sym *new_syms, *syms = NULL;
+    struct elf_sym *new_syms;
+    uint64_t segsz;
 
     shnum = hdr->e_shnum;
-    i = shnum * sizeof(struct elf_shdr);
+    shdr = imgsrc_read_alloc(hdr->e_shoff, shnum * sizeof(struct elf_shdr),
-    shdr = (struct elf_shdr *)alloca(i);
+                             src, NULL);
-    if (pread(fd, shdr, i, hdr->e_shoff) != i) {
+    if (shdr == NULL) {
         return;
     }
 
@@ -3625,31 +3706,33 @@ static void load_symbols(struct elfhdr *hdr, int fd, abi_ulong load_bias)
 
  found:
     /* Now know where the strtab and symtab are.  Snarf them.  */
-    s = g_try_new(struct syminfo, 1);
-    if (!s) {
-        goto give_up;
-    }
 
     segsz = shdr[str_idx].sh_size;
-    s->disas_strtab = strings = g_try_malloc(segsz);
+    strings = g_try_malloc(segsz);
-    if (!strings ||
+    if (!strings) {
-        pread(fd, strings, segsz, shdr[str_idx].sh_offset) != segsz) {
+        goto give_up;
+    }
+    if (!imgsrc_read(strings, shdr[str_idx].sh_offset, segsz, src, NULL)) {
         goto give_up;
     }
 
     segsz = shdr[sym_idx].sh_size;
-    syms = g_try_malloc(segsz);
-    if (!syms || pread(fd, syms, segsz, shdr[sym_idx].sh_offset) != segsz) {
-        goto give_up;
-    }
-
     if (segsz / sizeof(struct elf_sym) > INT_MAX) {
-        /* Implausibly large symbol table: give up rather than ploughing
+        /*
-         * on with the number of symbols calculation overflowing
+         * Implausibly large symbol table: give up rather than ploughing
+         * on with the number of symbols calculation overflowing.
          */
         goto give_up;
     }
     nsyms = segsz / sizeof(struct elf_sym);
+    syms = g_try_malloc(segsz);
+    if (!syms) {
+        goto give_up;
+    }
+    if (!imgsrc_read(syms, shdr[sym_idx].sh_offset, segsz, src, NULL)) {
+        goto give_up;
+    }
+
     for (i = 0; i < nsyms; ) {
         bswap_sym(syms + i);
         /* Throw away entries which we do not need.  */
@@ -3674,10 +3757,12 @@ static void load_symbols(struct elfhdr *hdr, int fd, abi_ulong load_bias)
         goto give_up;
     }
 
-    /* Attempt to free the storage associated with the local symbols
+    /*
-       that we threw away.  Whether or not this has any effect on the
+     * Attempt to free the storage associated with the local symbols
-       memory allocation depends on the malloc implementation and how
+     * that we threw away.  Whether or not this has any effect on the
-       many symbols we managed to discard.  */
+     * memory allocation depends on the malloc implementation and how
+     * many symbols we managed to discard.
+     */
     new_syms = g_try_renew(struct elf_sym, syms, nsyms);
     if (new_syms == NULL) {
         goto give_up;
@@ -3686,20 +3771,23 @@ static void load_symbols(struct elfhdr *hdr, int fd, abi_ulong load_bias)
 
     qsort(syms, nsyms, sizeof(*syms), symcmp);
 
-    s->disas_num_syms = nsyms;
+    {
-#if ELF_CLASS == ELFCLASS32
+        struct syminfo *s = g_new(struct syminfo, 1);
-    s->disas_symtab.elf32 = syms;
-#else
-    s->disas_symtab.elf64 = syms;
-#endif
-    s->lookup_symbol = lookup_symbolxx;
-    s->next = syminfos;
-    syminfos = s;
 
+        s->disas_strtab = strings;
+        s->disas_num_syms = nsyms;
+#if ELF_CLASS == ELFCLASS32
+        s->disas_symtab.elf32 = syms;
+#else
+        s->disas_symtab.elf64 = syms;
+#endif
+        s->lookup_symbol = lookup_symbolxx;
+        s->next = syminfos;
+        syminfos = s;
+    }
     return;
 
-give_up:
+ give_up:
-    g_free(s);
     g_free(strings);
     g_free(syms);
 }
@@ -3741,8 +3829,14 @@ uint32_t get_elf_eflags(int fd)
 
 int load_elf_binary(struct linux_binprm *bprm, struct image_info *info)
 {
-    struct image_info interp_info;
+    /*
-    struct elfhdr elf_ex;
+     * We need a copy of the elf header for passing to create_elf_tables.
+     * We will have overwritten the original when we re-use bprm->buf
+     * while loading the interpreter.  Allocate the storage for this now
+     * and let elf_load_image do any swapping that may be required.
+     */
+    struct elfhdr ehdr;
+    struct image_info interp_info, vdso_info;
     char *elf_interpreter = NULL;
     char *scratch;
 
@@ -3751,13 +3845,7 @@ int load_elf_binary(struct linux_binprm *bprm, struct image_info *info)
     interp_info.fp_abi = MIPS_ABI_FP_UNKNOWN;
 #endif
 
-    load_elf_image(bprm->filename, bprm->fd, info,
+    load_elf_image(bprm->filename, &bprm->src, info, &ehdr, &elf_interpreter);
-                   &elf_interpreter, bprm->buf);
-
-    /* ??? We need a copy of the elf header for passing to create_elf_tables.
-       If we do nothing, we'll have overwritten this when we re-use bprm->buf
-       when we load the interpreter.  */
-    elf_ex = *(struct elfhdr *)bprm->buf;
 
     /* Do this so that we can load the interpreter, if need be.  We will
        change some of these later */
@@ -3829,10 +3917,14 @@ int load_elf_binary(struct linux_binprm *bprm, struct image_info *info)
     }
 
     /*
-     * TODO: load a vdso, which would also contain the signal trampolines.
+     * Load a vdso if available, which will amongst other things contain the
-     * Otherwise, allocate a private page to hold them.
+     * signal trampolines.  Otherwise, allocate a separate page for them.
      */
-    if (TARGET_ARCH_HAS_SIGTRAMP_PAGE) {
+    const VdsoImageInfo *vdso = vdso_image_info();
+    if (vdso) {
+        load_elf_vdso(&vdso_info, vdso);
+        info->vdso = vdso_info.load_bias;
+    } else if (TARGET_ARCH_HAS_SIGTRAMP_PAGE) {
         abi_long tramp_page = target_mmap(0, TARGET_PAGE_SIZE,
                                           PROT_READ | PROT_WRITE,
                                           MAP_PRIVATE | MAP_ANON, -1, 0);
@@ -3844,8 +3936,9 @@ int load_elf_binary(struct linux_binprm *bprm, struct image_info *info)
         target_mprotect(tramp_page, TARGET_PAGE_SIZE, PROT_READ | PROT_EXEC);
     }
 
-    bprm->p = create_elf_tables(bprm->p, bprm->argc, bprm->envc, &elf_ex,
+    bprm->p = create_elf_tables(bprm->p, bprm->argc, bprm->envc, &ehdr, info,
-                                info, (elf_interpreter ? &interp_info : NULL));
+                                elf_interpreter ? &interp_info : NULL,
+                                vdso ? &vdso_info : NULL);
     info->start_stack = bprm->p;
 
     /* If we have an interpreter, set that as the program's entry point.

linux-user/flatload.c

@@ -463,7 +463,7 @@ static int load_flat_file(struct linux_binprm * bprm,
         DBG_FLT("BINFMT_FLAT: ROM mapping of file (we hope)\n");
 
         textpos = target_mmap(0, text_len, PROT_READ|PROT_EXEC,
-                              MAP_PRIVATE, bprm->fd, 0);
+                              MAP_PRIVATE, bprm->src.fd, 0);
         if (textpos == -1) {
             fprintf(stderr, "Unable to mmap process text\n");
             return -1;
@@ -490,7 +490,7 @@ static int load_flat_file(struct linux_binprm * bprm,
         } else
 #endif
         {
-            result = target_pread(bprm->fd, datapos,
+            result = target_pread(bprm->src.fd, datapos,
                                   data_len + (relocs * sizeof(abi_ulong)),
                                   fpos);
         }
@@ -540,10 +540,10 @@ static int load_flat_file(struct linux_binprm * bprm,
         else
 #endif
         {
-            result = target_pread(bprm->fd, textpos,
+            result = target_pread(bprm->src.fd, textpos,
                                   text_len, 0);
             if (result >= 0) {
-                result = target_pread(bprm->fd, datapos,
+                result = target_pread(bprm->src.fd, datapos,
                     data_len + (relocs * sizeof(abi_ulong)),
                     ntohl(hdr->data_start));
             }

linux-user/gen-vdso-elfn.c.inc

@@ -0,0 +1,314 @@
+/*
+ * Post-process a vdso elf image for inclusion into qemu.
+ * Elf size specialization.
+ *
+ * Copyright 2023 Linaro, Ltd.
+ *
+ * SPDX-License-Identifier: GPL-2.0-or-later
+ */
+
+static void elfN(bswap_ehdr)(ElfN(Ehdr) *ehdr)
+{
+    bswaps(&ehdr->e_type);            /* Object file type */
+    bswaps(&ehdr->e_machine);         /* Architecture */
+    bswaps(&ehdr->e_version);         /* Object file version */
+    bswaps(&ehdr->e_entry);           /* Entry point virtual address */
+    bswaps(&ehdr->e_phoff);           /* Program header table file offset */
+    bswaps(&ehdr->e_shoff);           /* Section header table file offset */
+    bswaps(&ehdr->e_flags);           /* Processor-specific flags */
+    bswaps(&ehdr->e_ehsize);          /* ELF header size in bytes */
+    bswaps(&ehdr->e_phentsize);       /* Program header table entry size */
+    bswaps(&ehdr->e_phnum);           /* Program header table entry count */
+    bswaps(&ehdr->e_shentsize);       /* Section header table entry size */
+    bswaps(&ehdr->e_shnum);           /* Section header table entry count */
+    bswaps(&ehdr->e_shstrndx);        /* Section header string table index */
+}
+
+static void elfN(bswap_phdr)(ElfN(Phdr) *phdr)
+{
+    bswaps(&phdr->p_type);            /* Segment type */
+    bswaps(&phdr->p_flags);           /* Segment flags */
+    bswaps(&phdr->p_offset);          /* Segment file offset */
+    bswaps(&phdr->p_vaddr);           /* Segment virtual address */
+    bswaps(&phdr->p_paddr);           /* Segment physical address */
+    bswaps(&phdr->p_filesz);          /* Segment size in file */
+    bswaps(&phdr->p_memsz);           /* Segment size in memory */
+    bswaps(&phdr->p_align);           /* Segment alignment */
+}
+
+static void elfN(bswap_shdr)(ElfN(Shdr) *shdr)
+{
+    bswaps(&shdr->sh_name);
+    bswaps(&shdr->sh_type);
+    bswaps(&shdr->sh_flags);
+    bswaps(&shdr->sh_addr);
+    bswaps(&shdr->sh_offset);
+    bswaps(&shdr->sh_size);
+    bswaps(&shdr->sh_link);
+    bswaps(&shdr->sh_info);
+    bswaps(&shdr->sh_addralign);
+    bswaps(&shdr->sh_entsize);
+}
+
+static void elfN(bswap_sym)(ElfN(Sym) *sym)
+{
+    bswaps(&sym->st_name);
+    bswaps(&sym->st_value);
+    bswaps(&sym->st_size);
+    bswaps(&sym->st_shndx);
+}
+
+static void elfN(bswap_dyn)(ElfN(Dyn) *dyn)
+{
+    bswaps(&dyn->d_tag);              /* Dynamic type tag */
+    bswaps(&dyn->d_un.d_ptr);         /* Dynamic ptr or val, in union */
+}
+
+static void elfN(search_symtab)(ElfN(Shdr) *shdr, unsigned sym_idx,
+                                void *buf, bool need_bswap)
+{
+    unsigned str_idx = shdr[sym_idx].sh_link;
+    ElfN(Sym) *sym = buf + shdr[sym_idx].sh_offset;
+    unsigned sym_n = shdr[sym_idx].sh_size / sizeof(*sym);
+    const char *str = buf + shdr[str_idx].sh_offset;
+
+    for (unsigned i = 0; i < sym_n; ++i) {
+        const char *name;
+
+        if (need_bswap) {
+            elfN(bswap_sym)(sym + i);
+        }
+        name = str + sym[i].st_name;
+
+        if (sigreturn_sym && strcmp(sigreturn_sym, name) == 0) {
+            sigreturn_addr = sym[i].st_value;
+        }
+        if (rt_sigreturn_sym && strcmp(rt_sigreturn_sym, name) == 0) {
+            rt_sigreturn_addr = sym[i].st_value;
+        }
+    }
+}
+
+static void elfN(process)(FILE *outf, void *buf, bool need_bswap)
+{
+    ElfN(Ehdr) *ehdr = buf;
+    ElfN(Phdr) *phdr;
+    ElfN(Shdr) *shdr;
+    unsigned phnum, shnum;
+    unsigned dynamic_ofs = 0;
+    unsigned dynamic_addr = 0;
+    unsigned symtab_idx = 0;
+    unsigned dynsym_idx = 0;
+    unsigned first_segsz = 0;
+    int errors = 0;
+
+    if (need_bswap) {
+        elfN(bswap_ehdr)(ehdr);
+    }
+
+    phnum = ehdr->e_phnum;
+    phdr = buf + ehdr->e_phoff;
+    if (need_bswap) {
+        for (unsigned i = 0; i < phnum; ++i) {
+            elfN(bswap_phdr)(phdr + i);
+        }
+    }
+
+    shnum = ehdr->e_shnum;
+    shdr = buf + ehdr->e_shoff;
+    if (need_bswap) {
+        for (unsigned i = 0; i < shnum; ++i) {
+            elfN(bswap_shdr)(shdr + i);
+        }
+    }
+    for (unsigned i = 0; i < shnum; ++i) {
+        switch (shdr[i].sh_type) {
+        case SHT_SYMTAB:
+            symtab_idx = i;
+            break;
+        case SHT_DYNSYM:
+            dynsym_idx = i;
+            break;
+        }
+    }
+
+    /*
+     * Validate the VDSO is created as we expect: that PT_PHDR,
+     * PT_DYNAMIC, and PT_NOTE located in a writable data segment.
+     * PHDR and DYNAMIC require relocation, and NOTE will get the
+     * linux version number.
+     */
+    for (unsigned i = 0; i < phnum; ++i) {
+        if (phdr[i].p_type != PT_LOAD) {
+            continue;
+        }
+        if (first_segsz != 0) {
+            fprintf(stderr, "Multiple LOAD segments\n");
+            errors++;
+        }
+        if (phdr[i].p_offset != 0) {
+            fprintf(stderr, "LOAD segment does not cover EHDR\n");
+            errors++;
+        }
+        if (phdr[i].p_vaddr != 0) {
+            fprintf(stderr, "LOAD segment not loaded at address 0\n");
+            errors++;
+        }
+        first_segsz = phdr[i].p_filesz;
+        if (first_segsz < ehdr->e_phoff + phnum * sizeof(*phdr)) {
+            fprintf(stderr, "LOAD segment does not cover PHDRs\n");
+            errors++;
+        }
+        if ((phdr[i].p_flags & (PF_R | PF_W)) != (PF_R | PF_W)) {
+            fprintf(stderr, "LOAD segment is not read-write\n");
+            errors++;
+        }
+    }
+    for (unsigned i = 0; i < phnum; ++i) {
+        const char *which;
+
+        switch (phdr[i].p_type) {
+        case PT_PHDR:
+            which = "PT_PHDR";
+            break;
+        case PT_NOTE:
+            which = "PT_NOTE";
+            break;
+        case PT_DYNAMIC:
+            dynamic_ofs = phdr[i].p_offset;
+            dynamic_addr = phdr[i].p_vaddr;
+            which = "PT_DYNAMIC";
+            break;
+        default:
+            continue;
+        }
+        if (first_segsz < phdr[i].p_vaddr + phdr[i].p_filesz) {
+            fprintf(stderr, "LOAD segment does not cover %s\n", which);
+            errors++;
+        }
+    }
+    if (errors) {
+        exit(EXIT_FAILURE);
+    }
+
+    /* Relocate the program headers. */
+    for (unsigned i = 0; i < phnum; ++i) {
+        output_reloc(outf, buf, &phdr[i].p_vaddr);
+        output_reloc(outf, buf, &phdr[i].p_paddr);
+    }
+
+    /* Relocate the DYNAMIC entries. */
+    if (dynamic_addr) {
+        ElfN(Dyn) *dyn = buf + dynamic_ofs;
+        __typeof(dyn->d_tag) tag;
+
+        do {
+
+            if (need_bswap) {
+                elfN(bswap_dyn)(dyn);
+            }
+            tag = dyn->d_tag;
+
+            switch (tag) {
+            case DT_HASH:
+            case DT_SYMTAB:
+            case DT_STRTAB:
+            case DT_VERDEF:
+            case DT_VERSYM:
+            case DT_PLTGOT:
+            case DT_ADDRRNGLO ... DT_ADDRRNGHI:
+                /* These entries store an address in the entry. */
+                output_reloc(outf, buf, &dyn->d_un.d_val);
+                break;
+
+            case DT_NULL:
+            case DT_STRSZ:
+            case DT_SONAME:
+            case DT_DEBUG:
+            case DT_FLAGS:
+            case DT_FLAGS_1:
+            case DT_SYMBOLIC:
+            case DT_BIND_NOW:
+            case DT_VERDEFNUM:
+            case DT_VALRNGLO ... DT_VALRNGHI:
+                /* These entries store an integer in the entry. */
+                break;
+
+            case DT_SYMENT:
+                if (dyn->d_un.d_val != sizeof(ElfN(Sym))) {
+                    fprintf(stderr, "VDSO has incorrect dynamic symbol size\n");
+                    errors++;
+                }
+                break;
+
+            case DT_REL:
+            case DT_RELSZ:
+            case DT_RELA:
+            case DT_RELASZ:
+                /*
+                 * These entries indicate that the VDSO was built incorrectly.
+                 * It should not have any real relocations.
+                 * ??? The RISC-V toolchain will emit these even when there
+                 * are no relocations.  Validate zeros.
+                 */
+                if (dyn->d_un.d_val != 0) {
+                    fprintf(stderr, "VDSO has dynamic relocations\n");
+                    errors++;
+                }
+                break;
+            case DT_RELENT:
+            case DT_RELAENT:
+            case DT_TEXTREL:
+                /* These entries store an integer in the entry. */
+                /* Should not be required; see above. */
+                break;
+
+            case DT_NEEDED:
+            case DT_VERNEED:
+            case DT_PLTREL:
+            case DT_JMPREL:
+            case DT_RPATH:
+            case DT_RUNPATH:
+                fprintf(stderr, "VDSO has external dependencies\n");
+                errors++;
+                break;
+
+            case PT_LOPROC + 3:
+                if (ehdr->e_machine == EM_PPC64) {
+                    break;  /* DT_PPC64_OPT: integer bitmask */
+                }
+                goto do_default;
+
+            default:
+            do_default:
+                /* This is probably something target specific. */
+                fprintf(stderr, "VDSO has unknown DYNAMIC entry (%lx)\n",
+                        (unsigned long)tag);
+                errors++;
+                break;
+            }
+            dyn++;
+        } while (tag != DT_NULL);
+        if (errors) {
+            exit(EXIT_FAILURE);
+        }
+    }
+
+    /* Relocate the dynamic symbol table. */
+    if (dynsym_idx) {
+        ElfN(Sym) *sym = buf + shdr[dynsym_idx].sh_offset;
+        unsigned sym_n = shdr[dynsym_idx].sh_size / sizeof(*sym);
+
+        for (unsigned i = 0; i < sym_n; ++i) {
+            output_reloc(outf, buf, &sym[i].st_value);
+        }
+    }
+
+    /* Search both dynsym and symtab for the signal return symbols. */
+    if (dynsym_idx) {
+        elfN(search_symtab)(shdr, dynsym_idx, buf, need_bswap);
+    }
+    if (symtab_idx) {
+        elfN(search_symtab)(shdr, symtab_idx, buf, need_bswap);
+    }
+}

linux-user/gen-vdso.c

@@ -0,0 +1,223 @@
+/*
+ * Post-process a vdso elf image for inclusion into qemu.
+ *
+ * Copyright 2023 Linaro, Ltd.
+ *
+ * SPDX-License-Identifier: GPL-2.0-or-later
+ */
+
+#include <stdlib.h>
+#include <stdbool.h>
+#include <stdint.h>
+#include <stdio.h>
+#include <string.h>
+#include <errno.h>
+#include <endian.h>
+#include <unistd.h>
+#include "elf.h"
+
+
+#define bswap_(p)  _Generic(*(p), \
+                            uint16_t: __builtin_bswap16,       \
+                            uint32_t: __builtin_bswap32,       \
+                            uint64_t: __builtin_bswap64,       \
+                            int16_t: __builtin_bswap16,        \
+                            int32_t: __builtin_bswap32,        \
+                            int64_t: __builtin_bswap64)
+#define bswaps(p) (*(p) = bswap_(p)(*(p)))
+
+static void output_reloc(FILE *outf, void *buf, void *loc)
+{
+    fprintf(outf, "    0x%08tx,\n", loc - buf);
+}
+
+static const char *sigreturn_sym;
+static const char *rt_sigreturn_sym;
+
+static unsigned sigreturn_addr;
+static unsigned rt_sigreturn_addr;
+
+#define N 32
+#define elfN(x)  elf32_##x
+#define ElfN(x)  Elf32_##x
+#include "gen-vdso-elfn.c.inc"
+#undef N
+#undef elfN
+#undef ElfN
+
+#define N 64
+#define elfN(x)  elf64_##x
+#define ElfN(x)  Elf64_##x
+#include "gen-vdso-elfn.c.inc"
+#undef N
+#undef elfN
+#undef ElfN
+
+
+int main(int argc, char **argv)
+{
+    FILE *inf, *outf;
+    long total_len;
+    const char *prefix = "vdso";
+    const char *inf_name;
+    const char *outf_name = NULL;
+    unsigned char *buf;
+    bool need_bswap;
+
+    while (1) {
+        int opt = getopt(argc, argv, "o:p:r:s:");
+        if (opt < 0) {
+            break;
+        }
+        switch (opt) {
+        case 'o':
+            outf_name = optarg;
+            break;
+        case 'p':
+            prefix = optarg;
+            break;
+        case 'r':
+            rt_sigreturn_sym = optarg;
+            break;
+        case 's':
+            sigreturn_sym = optarg;
+            break;
+        default:
+        usage:
+            fprintf(stderr, "usage: [-p prefix] [-r rt-sigreturn-name] "
+                    "[-s sigreturn-name] -o output-file input-file\n");
+            return EXIT_FAILURE;
+        }
+    }
+
+    if (optind >= argc || outf_name == NULL) {
+        goto usage;
+    }
+    inf_name = argv[optind];
+
+    /*
+     * Open the input and output files.
+     */
+    inf = fopen(inf_name, "rb");
+    if (inf == NULL) {
+        goto perror_inf;
+    }
+    outf = fopen(outf_name, "w");
+    if (outf == NULL) {
+        goto perror_outf;
+    }
+
+    /*
+     * Read the input file into a buffer.
+     * We expect the vdso to be small, on the order of one page,
+     * therefore we do not expect a partial read.
+     */
+    fseek(inf, 0, SEEK_END);
+    total_len = ftell(inf);
+    fseek(inf, 0, SEEK_SET);
+
+    buf = malloc(total_len);
+    if (buf == NULL) {
+        goto perror_inf;
+    }
+
+    errno = 0;
+    if (fread(buf, 1, total_len, inf) != total_len) {
+        if (errno) {
+            goto perror_inf;
+        }
+        fprintf(stderr, "%s: incomplete read\n", inf_name);
+        return EXIT_FAILURE;
+    }
+    fclose(inf);
+
+    /*
+     * Write out the vdso image now, before we make local changes.
+     */
+
+    fprintf(outf,
+            "/* Automatically generated from linux-user/gen-vdso.c. */\n"
+            "\n"
+            "static const uint8_t %s_image[] = {",
+            prefix);
+    for (long i = 0; i < total_len; ++i) {
+        if (i % 12 == 0) {
+            fputs("\n   ", outf);
+        }
+        fprintf(outf, " 0x%02x,", buf[i]);
+    }
+    fprintf(outf, "\n};\n\n");
+
+    /*
+     * Identify which elf flavor we're processing.
+     * The first 16 bytes of the file are e_ident.
+     */
+
+    if (buf[EI_MAG0] != ELFMAG0 || buf[EI_MAG1] != ELFMAG1 ||
+        buf[EI_MAG2] != ELFMAG2 || buf[EI_MAG3] != ELFMAG3) {
+        fprintf(stderr, "%s: not an elf file\n", inf_name);
+        return EXIT_FAILURE;
+    }
+    switch (buf[EI_DATA]) {
+    case ELFDATA2LSB:
+        need_bswap = BYTE_ORDER != LITTLE_ENDIAN;
+        break;
+    case ELFDATA2MSB:
+        need_bswap = BYTE_ORDER != BIG_ENDIAN;
+        break;
+    default:
+        fprintf(stderr, "%s: invalid elf EI_DATA (%u)\n",
+                inf_name, buf[EI_DATA]);
+        return EXIT_FAILURE;
+    }
+
+    /*
+     * We need to relocate the VDSO image.  The one built into the kernel
+     * is built for a fixed address.  The one we built for QEMU is not,
+     * since that requires close control of the guest address space.
+     *
+     * Output relocation addresses as we go.
+     */
+
+    fprintf(outf, "static const unsigned %s_relocs[] = {\n", prefix);
+
+    switch (buf[EI_CLASS]) {
+    case ELFCLASS32:
+        elf32_process(outf, buf, need_bswap);
+        break;
+    case ELFCLASS64:
+        elf64_process(outf, buf, need_bswap);
+        break;
+    default:
+        fprintf(stderr, "%s: invalid elf EI_CLASS (%u)\n",
+                inf_name, buf[EI_CLASS]);
+        return EXIT_FAILURE;
+    }
+
+    fprintf(outf, "};\n\n");   /* end vdso_relocs. */
+
+    fprintf(outf, "static const VdsoImageInfo %s_image_info = {\n", prefix);
+    fprintf(outf, "    .image = %s_image,\n", prefix);
+    fprintf(outf, "    .relocs = %s_relocs,\n", prefix);
+    fprintf(outf, "    .image_size = sizeof(%s_image),\n", prefix);
+    fprintf(outf, "    .reloc_count = ARRAY_SIZE(%s_relocs),\n", prefix);
+    fprintf(outf, "    .sigreturn_ofs = 0x%x,\n", sigreturn_addr);
+    fprintf(outf, "    .rt_sigreturn_ofs = 0x%x,\n", rt_sigreturn_addr);
+    fprintf(outf, "};\n");
+
+    /*
+     * Everything should have gone well.
+     */
+    if (fclose(outf)) {
+        goto perror_outf;
+    }
+    return EXIT_SUCCESS;
+
+ perror_inf:
+    perror(inf_name);
+    return EXIT_FAILURE;
+
+ perror_outf:
+    perror(outf_name);
+    return EXIT_FAILURE;
+}

linux-user/hppa/Makefile.vdso

@@ -0,0 +1,11 @@
+include $(BUILD_DIR)/tests/tcg/hppa-linux-user/config-target.mak
+
+SUBDIR = $(SRC_PATH)/linux-user/hppa
+VPATH += $(SUBDIR)
+
+all: $(SUBDIR)/vdso.so
+
+$(SUBDIR)/vdso.so: vdso.S vdso.ld vdso-asmoffset.h
+	$(CC) -o $@ -nostdlib -shared -Wl,-h,linux-vdso32.so.1 \
+	  -Wl,--build-id=sha1 -Wl,--hash-style=both \
+	  -Wl,-T,$(SUBDIR)/vdso.ld $<

linux-user/hppa/meson.build

@@ -3,3 +3,8 @@ syscall_nr_generators += {
                     arguments: [ meson.current_source_dir() / 'syscallhdr.sh', '@INPUT@', '@OUTPUT@', '@EXTRA_ARGS@' ],
                     output: '@BASENAME@_nr.h')
 }
+
+vdso_inc = gen_vdso.process('vdso.so',
+                            extra_args: [ '-r', '__kernel_sigtramp_rt' ])
+
+linux_user_ss.add(when: 'TARGET_HPPA', if_true: vdso_inc)

linux-user/hppa/signal.c

@@ -21,6 +21,7 @@
 #include "user-internals.h"
 #include "signal-common.h"
 #include "linux-user/trace.h"
+#include "vdso-asmoffset.h"
 
 struct target_sigcontext {
     abi_ulong sc_flags;
@@ -47,6 +48,19 @@ struct target_rt_sigframe {
     /* hidden location of upper halves of pa2.0 64-bit gregs */
 };
 
+QEMU_BUILD_BUG_ON(sizeof(struct target_rt_sigframe) != sizeof_rt_sigframe);
+QEMU_BUILD_BUG_ON(offsetof(struct target_rt_sigframe, uc.tuc_mcontext)
+                  != offsetof_sigcontext);
+QEMU_BUILD_BUG_ON(offsetof(struct target_sigcontext, sc_gr)
+                  != offsetof_sigcontext_gr);
+QEMU_BUILD_BUG_ON(offsetof(struct target_sigcontext, sc_fr)
+                  != offsetof_sigcontext_fr);
+QEMU_BUILD_BUG_ON(offsetof(struct target_sigcontext, sc_iaoq)
+                  != offsetof_sigcontext_iaoq);
+QEMU_BUILD_BUG_ON(offsetof(struct target_sigcontext, sc_sar)
+                  != offsetof_sigcontext_sar);
+
+
 static void setup_sigcontext(struct target_sigcontext *sc, CPUArchState *env)
 {
     int i;
@@ -91,16 +105,6 @@ static void restore_sigcontext(CPUArchState *env, struct target_sigcontext *sc)
     __get_user(env->cr[CR_SAR], &sc->sc_sar);
 }
 
-#if TARGET_ABI_BITS == 32
-#define SIGFRAME                64
-#define FUNCTIONCALLFRAME       48
-#else
-#define SIGFRAME                128
-#define FUNCTIONCALLFRAME       96
-#endif
-#define PARISC_RT_SIGFRAME_SIZE32 \
-    ((sizeof(struct target_rt_sigframe) + FUNCTIONCALLFRAME + SIGFRAME) & -SIGFRAME)
-
 void setup_rt_frame(int sig, struct target_sigaction *ka,
                     target_siginfo_t *info,
                     target_sigset_t *set, CPUArchState *env)

linux-user/hppa/vdso-asmoffset.h

@@ -0,0 +1,12 @@
+#define sizeof_rt_sigframe              584
+#define offsetof_sigcontext             160
+#define offsetof_sigcontext_gr          0x4
+#define offsetof_sigcontext_fr          0x88
+#define offsetof_sigcontext_iaoq        0x190
+#define offsetof_sigcontext_sar         0x198
+
+/* arch/parisc/include/asm/rt_sigframe.h */
+#define SIGFRAME                        64
+#define FUNCTIONCALLFRAME               48
+#define PARISC_RT_SIGFRAME_SIZE32 \
+    (((sizeof_rt_sigframe) + FUNCTIONCALLFRAME + SIGFRAME) & -SIGFRAME)

linux-user/hppa/vdso.S

@@ -0,0 +1,165 @@
+/*
+ * hppa linux kernel vdso replacement.
+ *
+ * Copyright 2023 Linaro, Ltd.
+ *
+ * SPDX-License-Identifier: GPL-2.0-or-later
+ */
+
+#include <asm/unistd.h>
+#include "vdso-asmoffset.h"
+
+	.text
+
+
+/*
+ * arch/parisc/kernel/vdso32/sigtramp.S:
+ * Gdb expects the trampoline is on the stack and the pc is offset from
+ * a 64-byte boundary by 0, 4 or 5 instructions. Since the vdso trampoline
+ * is not on the stack, we need a new variant with different offsets and
+ * data to tell gdb where to find the signal context on the stack.
+ *
+ * Here we put the offset to the context data at the start of the trampoline
+ * region and offset the first trampoline by 2 instructions. Please do
+ * not change the trampoline as the code in gdb depends on the following
+ * instruction sequence exactly.
+ */
+
+/* arch/parisc/kernel/asm-offsets.c */
+#define SIGFRAME_CONTEXT_REGS32 \
+    (offsetof_sigcontext - PARISC_RT_SIGFRAME_SIZE32)
+
+        .align	64
+        .word	SIGFRAME_CONTEXT_REGS32
+
+/*
+ * All that said, we can provide a proper unwind record, which means that
+ * GDB should not actually need the offset magic.
+ *
+ * The return address that arrived here, from the inner frame, is
+ * not marked as a signal frame and so the unwinder still tries to
+ * subtract 1 to examine the presumed call insn.  Thus we must
+ * extend the unwind info to a nop before the start.
+ */
+
+	.cfi_startproc simple
+	.cfi_signal_frame
+
+	/* Compare pa32_fallback_frame_state from libgcc. */
+
+	/*
+	 * Place the CFA at the start of sigcontext for convenience.
+	 * The previous CFA will be restored from the saved stack pointer.
+	 */
+	.cfi_def_cfa	30, -PARISC_RT_SIGFRAME_SIZE32 + offsetof_sigcontext
+
+	/* Record save offset of general registers. */
+	.cfi_offset	1, offsetof_sigcontext_gr + 1 * 4
+	.cfi_offset	2, offsetof_sigcontext_gr + 2 * 4
+	.cfi_offset	3, offsetof_sigcontext_gr + 3 * 4
+	.cfi_offset	4, offsetof_sigcontext_gr + 4 * 4
+	.cfi_offset	5, offsetof_sigcontext_gr + 5 * 4
+	.cfi_offset	6, offsetof_sigcontext_gr + 6 * 4
+	.cfi_offset	7, offsetof_sigcontext_gr + 7 * 4
+	.cfi_offset	8, offsetof_sigcontext_gr + 8 * 4
+	.cfi_offset	9, offsetof_sigcontext_gr + 9 * 4
+	.cfi_offset	10, offsetof_sigcontext_gr + 10 * 4
+	.cfi_offset	11, offsetof_sigcontext_gr + 11 * 4
+	.cfi_offset	12, offsetof_sigcontext_gr + 12 * 4
+	.cfi_offset	13, offsetof_sigcontext_gr + 13 * 4
+	.cfi_offset	14, offsetof_sigcontext_gr + 14 * 4
+	.cfi_offset	15, offsetof_sigcontext_gr + 15 * 4
+	.cfi_offset	16, offsetof_sigcontext_gr + 16 * 4
+	.cfi_offset	17, offsetof_sigcontext_gr + 17 * 4
+	.cfi_offset	18, offsetof_sigcontext_gr + 18 * 4
+	.cfi_offset	19, offsetof_sigcontext_gr + 19 * 4
+	.cfi_offset	20, offsetof_sigcontext_gr + 20 * 4
+	.cfi_offset	21, offsetof_sigcontext_gr + 21 * 4
+	.cfi_offset	22, offsetof_sigcontext_gr + 22 * 4
+	.cfi_offset	23, offsetof_sigcontext_gr + 23 * 4
+	.cfi_offset	24, offsetof_sigcontext_gr + 24 * 4
+	.cfi_offset	25, offsetof_sigcontext_gr + 25 * 4
+	.cfi_offset	26, offsetof_sigcontext_gr + 26 * 4
+	.cfi_offset	27, offsetof_sigcontext_gr + 27 * 4
+	.cfi_offset	28, offsetof_sigcontext_gr + 28 * 4
+	.cfi_offset	29, offsetof_sigcontext_gr + 29 * 4
+	.cfi_offset	30, offsetof_sigcontext_gr + 30 * 4
+	.cfi_offset	31, offsetof_sigcontext_gr + 31 * 4
+
+	/* Record save offset of fp registers, left and right halves. */
+	.cfi_offset	32, offsetof_sigcontext_fr + 4 * 8
+	.cfi_offset	33, offsetof_sigcontext_fr + 4 * 8 + 4
+	.cfi_offset	34, offsetof_sigcontext_fr + 5 * 8
+	.cfi_offset	35, offsetof_sigcontext_fr + 5 * 8 + 4
+	.cfi_offset	36, offsetof_sigcontext_fr + 6 * 8
+	.cfi_offset	37, offsetof_sigcontext_fr + 6 * 8 + 4
+	.cfi_offset	38, offsetof_sigcontext_fr + 7 * 8
+	.cfi_offset	39, offsetof_sigcontext_fr + 7 * 8 + 4
+	.cfi_offset	40, offsetof_sigcontext_fr + 8 * 8
+	.cfi_offset	41, offsetof_sigcontext_fr + 8 * 8 + 4
+	.cfi_offset	42, offsetof_sigcontext_fr + 9 * 8
+	.cfi_offset	43, offsetof_sigcontext_fr + 9 * 8 + 4
+	.cfi_offset	44, offsetof_sigcontext_fr + 10 * 8
+	.cfi_offset	45, offsetof_sigcontext_fr + 10 * 8 + 4
+	.cfi_offset	46, offsetof_sigcontext_fr + 11 * 8
+	.cfi_offset	47, offsetof_sigcontext_fr + 11 * 8 + 4
+	.cfi_offset	48, offsetof_sigcontext_fr + 12 * 8
+	.cfi_offset	49, offsetof_sigcontext_fr + 12 * 8 + 4
+	.cfi_offset	50, offsetof_sigcontext_fr + 13 * 8
+	.cfi_offset	51, offsetof_sigcontext_fr + 13 * 8 + 4
+	.cfi_offset	52, offsetof_sigcontext_fr + 14 * 8
+	.cfi_offset	53, offsetof_sigcontext_fr + 14 * 8 + 4
+	.cfi_offset	54, offsetof_sigcontext_fr + 15 * 8
+	.cfi_offset	55, offsetof_sigcontext_fr + 15 * 8 + 4
+	.cfi_offset	56, offsetof_sigcontext_fr + 16 * 8
+	.cfi_offset	57, offsetof_sigcontext_fr + 16 * 8 + 4
+	.cfi_offset	58, offsetof_sigcontext_fr + 17 * 8
+	.cfi_offset	59, offsetof_sigcontext_fr + 17 * 8 + 4
+	.cfi_offset	60, offsetof_sigcontext_fr + 18 * 8
+	.cfi_offset	61, offsetof_sigcontext_fr + 18 * 8 + 4
+	.cfi_offset	62, offsetof_sigcontext_fr + 19 * 8
+	.cfi_offset	63, offsetof_sigcontext_fr + 19 * 8 + 4
+	.cfi_offset	64, offsetof_sigcontext_fr + 20 * 8
+	.cfi_offset	65, offsetof_sigcontext_fr + 20 * 8 + 4
+	.cfi_offset	66, offsetof_sigcontext_fr + 21 * 8
+	.cfi_offset	67, offsetof_sigcontext_fr + 21 * 8 + 4
+	.cfi_offset	68, offsetof_sigcontext_fr + 22 * 8
+	.cfi_offset	69, offsetof_sigcontext_fr + 22 * 8 + 4
+	.cfi_offset	70, offsetof_sigcontext_fr + 23 * 8
+	.cfi_offset	71, offsetof_sigcontext_fr + 23 * 8 + 4
+	.cfi_offset	72, offsetof_sigcontext_fr + 24 * 8
+	.cfi_offset	73, offsetof_sigcontext_fr + 24 * 8 + 4
+	.cfi_offset	74, offsetof_sigcontext_fr + 25 * 8
+	.cfi_offset	75, offsetof_sigcontext_fr + 25 * 8 + 4
+	.cfi_offset	76, offsetof_sigcontext_fr + 26 * 8
+	.cfi_offset	77, offsetof_sigcontext_fr + 26 * 8 + 4
+	.cfi_offset	78, offsetof_sigcontext_fr + 27 * 8
+	.cfi_offset	79, offsetof_sigcontext_fr + 27 * 8 + 4
+	.cfi_offset	80, offsetof_sigcontext_fr + 28 * 8
+	.cfi_offset	81, offsetof_sigcontext_fr + 28 * 8 + 4
+	.cfi_offset	82, offsetof_sigcontext_fr + 29 * 8
+	.cfi_offset	83, offsetof_sigcontext_fr + 29 * 8 + 4
+	.cfi_offset	84, offsetof_sigcontext_fr + 30 * 8
+	.cfi_offset	85, offsetof_sigcontext_fr + 30 * 8 + 4
+	.cfi_offset	86, offsetof_sigcontext_fr + 31 * 8
+	.cfi_offset	87, offsetof_sigcontext_fr + 31 * 8 + 4
+
+	/* Record save offset of %sar */
+	.cfi_offset	88, offsetof_sigcontext_sar
+
+	/* Record save offset of return address, iaoq[0]. */
+	.cfi_return_column 89
+	.cfi_offset	89, offsetof_sigcontext_iaoq
+
+	nop
+
+__kernel_sigtramp_rt:
+	ldi	0, %r25
+	ldi	__NR_rt_sigreturn, %r20
+	be,l	0x100(%sr2, %r0), %sr0, %r31
+	nop
+
+	.cfi_endproc
+	.size	__kernel_sigtramp_rt, . - __kernel_sigtramp_rt
+	.type	__kernel_sigtramp_rt, @function
+	.globl	__kernel_sigtramp_rt

linux-user/hppa/vdso.ld

@@ -0,0 +1,77 @@
+/*
+ * Linker script for linux hppa vdso.
+ *
+ * This library is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU Lesser General Public
+ * License as published by the Free Software Foundation; either
+ * version 2 of the License, or (at your option) any later version.
+ *
+ * This library is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+ * Lesser General Public License for more details.
+ *
+ * You should have received a copy of the GNU Lesser General Public
+ * License along with this library; if not, see <http://www.gnu.org/licenses/>.
+ */
+
+VERSION {
+        /*
+         * The kernel's vdso32.lds.S attempts to export
+         *     __kernel_sigtramp_rt32
+         *     __kernel_restart_syscall32
+         * except that those symbols don't exist.  The actual symbols are
+         *     __kernel_sigtramp_rt
+         *     __kernel_restart_syscall
+         * which means that nothing is exported at all.
+         * QEMU handles syscall restart internally, so we don't
+         * need to implement __kernel_restart_syscall at all.
+         */
+        LINUX_5.18 {
+        local: *;
+        };
+}
+
+
+PHDRS {
+        phdr            PT_PHDR         FLAGS(4) PHDRS;
+        load            PT_LOAD         FLAGS(7) FILEHDR PHDRS;
+        dynamic         PT_DYNAMIC      FLAGS(4);
+        note            PT_NOTE         FLAGS(4);
+        eh_frame_hdr    PT_GNU_EH_FRAME;
+}
+
+SECTIONS {
+        . = SIZEOF_HEADERS;
+
+        /* The following, including the FILEHDRS and PHDRS, are modified
+           when we relocate the binary.  We want them to be initially
+           writable for the relocation; we'll force them read-only after.  */
+        .note           : { *(.note*) }         :load :note
+        .dynamic        : { *(.dynamic) }       :load :dynamic
+        .dynsym         : { *(.dynsym) }        :load
+        .data           : {
+                /* There ought not be any real read-write data.
+                   But since we manipulated the segment layout,
+                   we have to put these sections somewhere.  */
+                *(.data*)
+                *(.sdata*)
+                *(.got.plt) *(.got)
+                *(.gnu.linkonce.d.*)
+                *(.bss*)
+                *(.dynbss*)
+                *(.gnu.linkonce.b.*)
+        }
+
+        .rodata         : { *(.rodata) }
+        .hash           : { *(.hash) }
+        .gnu.hash       : { *(.gnu.hash) }
+        .dynstr         : { *(.dynstr) }
+        .gnu.version    : { *(.gnu.version) }
+        .gnu.version_d  : { *(.gnu.version_d) }
+        .gnu.version_r  : { *(.gnu.version_r) }
+        .eh_frame_hdr   : { *(.eh_frame_hdr) }  :load :eh_frame_hdr
+        .eh_frame       : { *(.eh_frame) }      :load
+
+        .text           : { *(.text*) }         :load
+}

linux-user/i386/Makefile.vdso

@@ -0,0 +1,11 @@
+include $(BUILD_DIR)/tests/tcg/i386-linux-user/config-target.mak
+
+SUBDIR = $(SRC_PATH)/linux-user/i386
+VPATH += $(SUBDIR)
+
+all: $(SUBDIR)/vdso.so
+
+$(SUBDIR)/vdso.so: vdso.S vdso.ld vdso-asmoffset.h
+	$(CC) -o $@ -m32 -nostdlib -shared -Wl,-h,linux-gate.so.1 \
+	  -Wl,--build-id=sha1 -Wl,--hash-style=both \
+	  -Wl,-T,$(SUBDIR)/vdso.ld $<

linux-user/i386/meson.build

@@ -3,3 +3,10 @@ syscall_nr_generators += {
                     arguments: [ meson.current_source_dir() / 'syscallhdr.sh', '@INPUT@', '@OUTPUT@', '@EXTRA_ARGS@' ],
                     output: '@BASENAME@_nr.h')
 }
+
+vdso_inc = gen_vdso.process('vdso.so', extra_args: [
+                                '-s', '__kernel_sigreturn',
+                                '-r', '__kernel_rt_sigreturn'
+                            ])
+
+linux_user_ss.add(when: 'TARGET_I386', if_true: vdso_inc)

linux-user/i386/signal.c

@@ -214,6 +214,17 @@ struct rt_sigframe {
 };
 #define TARGET_RT_SIGFRAME_FXSAVE_OFFSET (                                 \
     offsetof(struct rt_sigframe, fpstate) + TARGET_FPSTATE_FXSAVE_OFFSET)
+
+/*
+ * Verify that vdso-asmoffset.h constants match.
+ */
+#include "i386/vdso-asmoffset.h"
+
+QEMU_BUILD_BUG_ON(offsetof(struct sigframe, sc.eip)
+                  != SIGFRAME_SIGCONTEXT_eip);
+QEMU_BUILD_BUG_ON(offsetof(struct rt_sigframe, uc.tuc_mcontext.eip)
+                  != RT_SIGFRAME_SIGCONTEXT_eip);
+
 #else
 
 struct rt_sigframe {

linux-user/i386/vdso-asmoffset.h

@@ -0,0 +1,6 @@
+/*
+ * offsetof(struct sigframe, sc.eip)
+ * offsetof(struct rt_sigframe, uc.tuc_mcontext.eip)
+ */
+#define SIGFRAME_SIGCONTEXT_eip      64
+#define RT_SIGFRAME_SIGCONTEXT_eip  220

linux-user/i386/vdso.S

@@ -0,0 +1,143 @@
+/*
+ * i386 linux replacement vdso.
+ *
+ * Copyright 2023 Linaro, Ltd.
+ *
+ * SPDX-License-Identifier: GPL-2.0-or-later
+ */
+
+#include <asm/unistd.h>
+#include "vdso-asmoffset.h"
+
+.macro endf name
+	.globl	\name
+	.type	\name, @function
+	.size	\name, . - \name
+.endm
+
+.macro vdso_syscall1 name, nr
+\name:
+	.cfi_startproc
+	mov	%ebx, %edx
+	.cfi_register %ebx, %edx
+	mov	4(%esp), %ebx
+	mov	$\nr, %eax
+	int	$0x80
+	mov	%edx, %ebx
+	ret
+	.cfi_endproc
+endf	\name
+.endm
+
+.macro vdso_syscall2 name, nr
+\name:
+	.cfi_startproc
+	mov	%ebx, %edx
+	.cfi_register %ebx, %edx
+	mov	4(%esp), %ebx
+	mov	8(%esp), %ecx
+	mov	$\nr, %eax
+	int	$0x80
+	mov	%edx, %ebx
+	ret
+	.cfi_endproc
+endf	\name
+.endm
+
+.macro vdso_syscall3 name, nr
+\name:
+	.cfi_startproc
+	push	%ebx
+	.cfi_adjust_cfa_offset 4
+	.cfi_rel_offset %ebx, 0
+	mov	8(%esp), %ebx
+	mov	12(%esp), %ecx
+	mov	16(%esp), %edx
+	mov	$\nr, %eax
+	int	$0x80
+	pop	%ebx
+	.cfi_adjust_cfa_offset -4
+	.cfi_restore %ebx
+	ret
+	.cfi_endproc
+endf	\name
+.endm
+
+__kernel_vsyscall:
+	.cfi_startproc
+	int	$0x80
+	ret
+	.cfi_endproc
+endf	__kernel_vsyscall
+
+vdso_syscall2 __vdso_clock_gettime, __NR_clock_gettime
+vdso_syscall2 __vdso_clock_gettime64, __NR_clock_gettime64
+vdso_syscall2 __vdso_clock_getres, __NR_clock_getres
+vdso_syscall2 __vdso_gettimeofday, __NR_gettimeofday
+vdso_syscall1 __vdso_time, __NR_time
+vdso_syscall3 __vdso_getcpu, __NR_gettimeofday
+
+/*
+ * Signal return handlers.
+ */
+
+	.cfi_startproc simple
+	.cfi_signal_frame
+
+/*
+ * For convenience, put the cfa just above eip in sigcontext, and count
+ * offsets backward from there.  Re-compute the cfa in the two contexts
+ * we have for signal unwinding.  This is far simpler than the
+ * DW_CFA_expression form that the kernel uses, and is equally correct.
+ */
+
+	.cfi_def_cfa	%esp, SIGFRAME_SIGCONTEXT_eip + 4
+
+	.cfi_offset	%eip, -4
+			/* err, -8 */
+			/* trapno, -12 */
+	.cfi_offset	%eax, -16
+	.cfi_offset	%ecx, -20
+	.cfi_offset	%edx, -24
+	.cfi_offset	%ebx, -28
+	.cfi_offset	%esp, -32
+	.cfi_offset	%ebp, -36
+	.cfi_offset	%esi, -40
+	.cfi_offset	%edi, -44
+
+/*
+ * While this frame is marked as a signal frame, that only applies to how
+ * the return address is handled for the outer frame.  The return address
+ * that arrived here, from the inner frame, is not marked as a signal frame
+ * and so the unwinder still tries to subtract 1 to examine the presumed
+ * call insn.  Thus we must extend the unwind info to a nop before the start.
+ */
+	nop
+
+__kernel_sigreturn:
+	popl	%eax	/* pop sig */
+	.cfi_adjust_cfa_offset -4
+	movl	$__NR_sigreturn, %eax
+	int	$0x80
+endf	__kernel_sigreturn
+
+	.cfi_def_cfa_offset RT_SIGFRAME_SIGCONTEXT_eip + 4
+	nop
+
+__kernel_rt_sigreturn:
+	movl	$__NR_rt_sigreturn, %eax
+	int	$0x80
+endf	__kernel_rt_sigreturn
+
+	.cfi_endproc
+
+/*
+ * TODO: Add elf notes.  E.g.
+ *
+ * #include <linux/elfnote.h>
+ * ELFNOTE_START(Linux, 0, "a")
+ *   .long LINUX_VERSION_CODE
+ * ELFNOTE_END
+ *
+ * but what version number would we set for QEMU?
+ */

linux-user/i386/vdso.ld

@@ -0,0 +1,76 @@
+/*
+ * Linker script for linux i386 replacement vdso.
+ *
+ * Copyright 2023 Linaro, Ltd.
+ *
+ * SPDX-License-Identifier: GPL-2.0-or-later
+ */
+
+ENTRY(__kernel_vsyscall)
+
+VERSION {
+        LINUX_2.6 {
+        global:
+                __vdso_clock_gettime;
+                __vdso_gettimeofday;
+                __vdso_time;
+                __vdso_clock_getres;
+                __vdso_clock_gettime64;
+                __vdso_getcpu;
+        };
+
+        LINUX_2.5 {
+        global:
+                __kernel_vsyscall;
+                __kernel_sigreturn;
+                __kernel_rt_sigreturn;
+        local: *;
+        };
+}
+
+PHDRS {
+        phdr            PT_PHDR         FLAGS(4) PHDRS;
+        load            PT_LOAD         FLAGS(7) FILEHDR PHDRS; /* FLAGS=RWX */
+        dynamic         PT_DYNAMIC      FLAGS(4);
+        eh_frame_hdr    PT_GNU_EH_FRAME;
+        note            PT_NOTE         FLAGS(4);
+}
+
+SECTIONS {
+        . = SIZEOF_HEADERS;
+
+        /*
+         * The following, including the FILEHDRS and PHDRS, are modified
+         * when we relocate the binary.  We want them to be initially
+         * writable for the relocation; we'll force them read-only after.
+         */
+        .note           : { *(.note*) }         :load :note
+        .dynamic        : { *(.dynamic) }       :load :dynamic
+        .dynsym         : { *(.dynsym) }        :load
+        .data           : {
+                /*
+                 * There ought not be any real read-write data.
+                 * But since we manipulated the segment layout,
+                 * we have to put these sections somewhere.
+                 */
+                *(.data*)
+                *(.sdata*)
+                *(.got.plt) *(.got)
+                *(.gnu.linkonce.d.*)
+                *(.bss*)
+                *(.dynbss*)
+                *(.gnu.linkonce.b.*)
+        }
+
+        .rodata         : { *(.rodata*) }
+        .hash           : { *(.hash) }
+        .gnu.hash       : { *(.gnu.hash) }
+        .dynstr         : { *(.dynstr) }
+        .gnu.version    : { *(.gnu.version) }
+        .gnu.version_d  : { *(.gnu.version_d) }
+        .gnu.version_r  : { *(.gnu.version_r) }
+        .eh_frame_hdr   : { *(.eh_frame_hdr) }  :load :eh_frame_hdr
+        .eh_frame       : { *(.eh_frame) }      :load
+
+        .text           : { *(.text*) }         :load   =0x90909090
+}

linux-user/linuxload.c

@@ -3,7 +3,9 @@
 #include "qemu/osdep.h"
 #include "qemu.h"
 #include "user-internals.h"
+#include "user-mmap.h"
 #include "loader.h"
+#include "qapi/error.h"
 
 #define NGROUPS 32
 
@@ -37,7 +39,7 @@ static int prepare_binprm(struct linux_binprm *bprm)
     int mode;
     int retval;
 
-    if (fstat(bprm->fd, &st) < 0) {
+    if (fstat(bprm->src.fd, &st) < 0) {
         return -errno;
     }
 
@@ -67,7 +69,7 @@ static int prepare_binprm(struct linux_binprm *bprm)
         bprm->e_gid = st.st_gid;
     }
 
-    retval = read(bprm->fd, bprm->buf, BPRM_BUF_SIZE);
+    retval = read(bprm->src.fd, bprm->buf, BPRM_BUF_SIZE);
     if (retval < 0) {
         perror("prepare_binprm");
         exit(-1);
@@ -76,6 +78,10 @@ static int prepare_binprm(struct linux_binprm *bprm)
         /* Make sure the rest of the loader won't read garbage.  */
         memset(bprm->buf + retval, 0, BPRM_BUF_SIZE - retval);
     }
+
+    bprm->src.cache = bprm->buf;
+    bprm->src.cache_size = retval;
+
     return retval;
 }
 
@@ -138,7 +144,7 @@ int loader_exec(int fdexec, const char *filename, char **argv, char **envp,
 {
     int retval;
 
-    bprm->fd = fdexec;
+    bprm->src.fd = fdexec;
     bprm->filename = (char *)filename;
     bprm->argc = count(argv);
     bprm->argv = argv;
@@ -147,29 +153,112 @@ int loader_exec(int fdexec, const char *filename, char **argv, char **envp,
 
     retval = prepare_binprm(bprm);
 
-    if (retval >= 0) {
+    if (retval < 4) {
-        if (bprm->buf[0] == 0x7f
+        return -ENOEXEC;
-                && bprm->buf[1] == 'E'
-                && bprm->buf[2] == 'L'
-                && bprm->buf[3] == 'F') {
-            retval = load_elf_binary(bprm, infop);
-#if defined(TARGET_HAS_BFLT)
-        } else if (bprm->buf[0] == 'b'
-                && bprm->buf[1] == 'F'
-                && bprm->buf[2] == 'L'
-                && bprm->buf[3] == 'T') {
-            retval = load_flt_binary(bprm, infop);
-#endif
-        } else {
-            return -ENOEXEC;
-        }
     }
-
+    if (bprm->buf[0] == 0x7f
-    if (retval >= 0) {
+        && bprm->buf[1] == 'E'
-        /* success.  Initialize important registers */
+        && bprm->buf[2] == 'L'
-        do_init_thread(regs, infop);
+        && bprm->buf[3] == 'F') {
+        retval = load_elf_binary(bprm, infop);
+#if defined(TARGET_HAS_BFLT)
+    } else if (bprm->buf[0] == 'b'
+               && bprm->buf[1] == 'F'
+               && bprm->buf[2] == 'L'
+               && bprm->buf[3] == 'T') {
+        retval = load_flt_binary(bprm, infop);
+#endif
+    } else {
+        return -ENOEXEC;
+    }
+    if (retval < 0) {
         return retval;
     }
 
-    return retval;
+    /* Success.  Initialize important registers. */
+    do_init_thread(regs, infop);
+    return 0;
+}
+
+bool imgsrc_read(void *dst, off_t offset, size_t len,
+                 const ImageSource *img, Error **errp)
+{
+    ssize_t ret;
+
+    if (offset + len <= img->cache_size) {
+        memcpy(dst, img->cache + offset, len);
+        return true;
+    }
+
+    if (img->fd < 0) {
+        error_setg(errp, "read past end of buffer");
+        return false;
+    }
+
+    ret = pread(img->fd, dst, len, offset);
+    if (ret == len) {
+        return true;
+    }
+    if (ret < 0) {
+        error_setg_errno(errp, errno, "Error reading file header");
+    } else {
+        error_setg(errp, "Incomplete read of file header");
+    }
+    return false;
+}
+
+void *imgsrc_read_alloc(off_t offset, size_t len,
+                        const ImageSource *img, Error **errp)
+{
+    void *alloc = g_malloc(len);
+    bool ok = imgsrc_read(alloc, offset, len, img, errp);
+
+    if (!ok) {
+        g_free(alloc);
+        alloc = NULL;
+    }
+    return alloc;
+}
+
+abi_long imgsrc_mmap(abi_ulong start, abi_ulong len, int prot,
+                     int flags, const ImageSource *src, abi_ulong offset)
+{
+    const int prot_write = PROT_READ | PROT_WRITE;
+    abi_long ret;
+    void *haddr;
+
+    assert(flags == (MAP_PRIVATE | MAP_FIXED));
+
+    if (src->fd >= 0) {
+        return target_mmap(start, len, prot, flags, src->fd, offset);
+    }
+
+    /*
+     * This case is for the vdso; we don't expect bad images.
+     * The mmap may extend beyond the end of the image, especially
+     * to the end of the page.  Zero fill.
+     */
+    assert(offset < src->cache_size);
+
+    ret = target_mmap(start, len, prot_write, flags | MAP_ANON, -1, 0);
+    if (ret == -1) {
+        return ret;
+    }
+
+    haddr = lock_user(VERIFY_WRITE, start, len, 0);
+    assert(haddr != NULL);
+    if (offset + len <= src->cache_size) {
+        memcpy(haddr, src->cache + offset, len);
+    } else {
+        size_t rest = src->cache_size - offset;
+        memcpy(haddr, src->cache + offset, rest);
+        memset(haddr + rest, 0, len - rest);
+    }
+    unlock_user(haddr, start, len);
+
+    if (prot != prot_write) {
+        target_mprotect(start, len, prot);
+    }
+
+    return ret;
 }

linux-user/loader.h

@@ -18,6 +18,48 @@
 #ifndef LINUX_USER_LOADER_H
 #define LINUX_USER_LOADER_H
 
+typedef struct {
+    const void *cache;
+    unsigned int cache_size;
+    int fd;
+} ImageSource;
+
+/**
+ * imgsrc_read: Read from ImageSource
+ * @dst: destination for read
+ * @offset: offset within file for read
+ * @len: size of the read
+ * @img: ImageSource to read from
+ * @errp: Error details.
+ *
+ * Read into @dst, using the cache when possible.
+ */
+bool imgsrc_read(void *dst, off_t offset, size_t len,
+                 const ImageSource *img, Error **errp);
+
+/**
+ * imgsrc_read_alloc: Read from ImageSource
+ * @offset: offset within file for read
+ * @size: size of the read
+ * @img: ImageSource to read from
+ * @errp: Error details.
+ *
+ * Read into newly allocated memory, using the cache when possible.
+ */
+void *imgsrc_read_alloc(off_t offset, size_t len,
+                        const ImageSource *img, Error **errp);
+
+/**
+ * imgsrc_mmap: Map from ImageSource
+ *
+ * If @src has a file descriptor, pass on to target_mmap.  Otherwise,
+ * this is "mapping" from a host buffer, which resolves to memcpy.
+ * Therefore, flags must be MAP_PRIVATE | MAP_FIXED; the argument is
+ * retained for clarity.
+ */
+abi_long imgsrc_mmap(abi_ulong start, abi_ulong len, int prot,
+                     int flags, const ImageSource *src, abi_ulong offset);
+
 /*
  * Read a good amount of data initially, to hopefully get all the
  * program headers loaded.
@@ -29,15 +71,15 @@
  * used when loading binaries.
  */
 struct linux_binprm {
-        char buf[BPRM_BUF_SIZE] __attribute__((aligned));
+    char buf[BPRM_BUF_SIZE] __attribute__((aligned));
-        abi_ulong p;
+    ImageSource src;
-        int fd;
+    abi_ulong p;
-        int e_uid, e_gid;
+    int e_uid, e_gid;
-        int argc, envc;
+    int argc, envc;
-        char **argv;
+    char **argv;
-        char **envp;
+    char **envp;
-        char *filename;        /* Name of binary */
+    char *filename;        /* Name of binary */
-        int (*core_dump)(int, const CPUArchState *); /* coredump routine */
+    int (*core_dump)(int, const CPUArchState *); /* coredump routine */
 };
 
 void do_init_thread(struct target_pt_regs *regs, struct image_info *infop);

linux-user/loongarch64/Makefile.vdso

@@ -0,0 +1,11 @@
+include $(BUILD_DIR)/tests/tcg/loongarch64-linux-user/config-target.mak
+
+SUBDIR = $(SRC_PATH)/linux-user/loongarch64
+VPATH += $(SUBDIR)
+
+all: $(SUBDIR)/vdso.so
+
+$(SUBDIR)/vdso.so: vdso.S vdso.ld vdso-asmoffset.h
+	$(CC) -o $@ -nostdlib -shared -fpic -Wl,-h,linux-vdso.so.1 \
+	  -Wl,--build-id=sha1 -Wl,--hash-style=both \
+	  -Wl,--no-warn-rwx-segments -Wl,-T,$(SUBDIR)/vdso.ld $<

linux-user/loongarch64/meson.build

@@ -0,0 +1,4 @@
+vdso_inc = gen_vdso.process('vdso.so',
+                            extra_args: ['-r', '__vdso_rt_sigreturn'])
+
+linux_user_ss.add(when: 'TARGET_LOONGARCH64', if_true: vdso_inc)

linux-user/loongarch64/signal.c

@@ -10,9 +10,9 @@
 #include "user-internals.h"
 #include "signal-common.h"
 #include "linux-user/trace.h"
-
 #include "target/loongarch/internals.h"
 #include "target/loongarch/vec.h"
+#include "vdso-asmoffset.h"
 
 /* FP context was used */
 #define SC_USED_FP              (1 << 0)
@@ -24,6 +24,11 @@ struct target_sigcontext {
     uint64_t sc_extcontext[0]   QEMU_ALIGNED(16);
 };
 
+QEMU_BUILD_BUG_ON(sizeof(struct target_sigcontext) != sizeof_sigcontext);
+QEMU_BUILD_BUG_ON(offsetof(struct target_sigcontext, sc_pc)
+                  != offsetof_sigcontext_pc);
+QEMU_BUILD_BUG_ON(offsetof(struct target_sigcontext, sc_regs)
+                  != offsetof_sigcontext_gr);
 
 #define FPU_CTX_MAGIC           0x46505501
 #define FPU_CTX_ALIGN           8
@@ -33,6 +38,9 @@ struct target_fpu_context {
     uint32_t fcsr;
 } QEMU_ALIGNED(FPU_CTX_ALIGN);
 
+QEMU_BUILD_BUG_ON(offsetof(struct target_fpu_context, regs)
+                  != offsetof_fpucontext_fr);
+
 #define CONTEXT_INFO_ALIGN      16
 struct target_sctx_info {
     uint32_t magic;
@@ -40,6 +48,8 @@ struct target_sctx_info {
     uint64_t padding;
 } QEMU_ALIGNED(CONTEXT_INFO_ALIGN);
 
+QEMU_BUILD_BUG_ON(sizeof(struct target_sctx_info) != sizeof_sctx_info);
+
 struct target_ucontext {
     abi_ulong tuc_flags;
     abi_ptr tuc_link;
@@ -54,6 +64,11 @@ struct target_rt_sigframe {
     struct target_ucontext       rs_uc;
 };
 
+QEMU_BUILD_BUG_ON(sizeof(struct target_rt_sigframe)
+                  != sizeof_rt_sigframe);
+QEMU_BUILD_BUG_ON(offsetof(struct target_rt_sigframe, rs_uc.tuc_mcontext)
+                  != offsetof_sigcontext);
+
 /*
  * These two structures are not present in guest memory, are private
  * to the signal implementation, but are largely copied from the

linux-user/loongarch64/vdso-asmoffset.h

@@ -0,0 +1,8 @@
+#define sizeof_rt_sigframe         0x240
+#define sizeof_sigcontext          0x110
+#define sizeof_sctx_info           0x10
+
+#define offsetof_sigcontext        0x130
+#define offsetof_sigcontext_pc     0
+#define offsetof_sigcontext_gr     8
+#define offsetof_fpucontext_fr     0

linux-user/loongarch64/vdso.S

@@ -0,0 +1,130 @@
+/*
+ * Loongarch64 linux replacement vdso.
+ *
+ * Copyright 2023 Linaro, Ltd.
+ *
+ * SPDX-License-Identifier: GPL-2.0-or-later
+ */
+
+#include <asm/unistd.h>
+#include <asm/errno.h>
+#include "vdso-asmoffset.h"
+
+
+	.text
+
+.macro endf name
+	.globl	\name
+	.type	\name, @function
+	.size	\name, . - \name
+.endm
+
+.macro vdso_syscall name, nr
+\name:
+	li.w	$a7, \nr
+	syscall	0
+	jr	$ra
+endf	\name
+.endm
+
+	.cfi_startproc
+
+vdso_syscall __vdso_gettimeofday, __NR_gettimeofday
+vdso_syscall __vdso_clock_gettime, __NR_clock_gettime
+vdso_syscall __vdso_clock_getres, __NR_clock_getres
+vdso_syscall __vdso_getcpu, __NR_getcpu
+
+	.cfi_endproc
+
+/*
+ * Start the unwind info at least one instruction before the signal
+ * trampoline, because the unwinder will assume we are returning
+ * after a call site.
+ */
+
+	.cfi_startproc simple
+	.cfi_signal_frame
+
+#define B_GR	offsetof_sigcontext_gr
+#define B_FR	sizeof_sigcontext + sizeof_sctx_info + offsetof_fpucontext_fr
+
+	.cfi_def_cfa	2, offsetof_sigcontext
+
+	/* Return address */
+	.cfi_return_column 64
+	.cfi_offset	64, offsetof_sigcontext_pc	/* pc */
+
+	/* Integer registers */
+	.cfi_offset	1, B_GR + 1 * 8
+	.cfi_offset	2, B_GR + 2 * 8
+	.cfi_offset	3, B_GR + 3 * 8
+	.cfi_offset	4, B_GR + 4 * 8
+	.cfi_offset	5, B_GR + 5 * 8
+	.cfi_offset	6, B_GR + 6 * 8
+	.cfi_offset	7, B_GR + 7 * 8
+	.cfi_offset	8, B_GR + 8 * 8
+	.cfi_offset	9, B_GR + 9 * 8
+	.cfi_offset	10, B_GR + 10 * 8
+	.cfi_offset	11, B_GR + 11 * 8
+	.cfi_offset	12, B_GR + 12 * 8
+	.cfi_offset	13, B_GR + 13 * 8
+	.cfi_offset	14, B_GR + 14 * 8
+	.cfi_offset	15, B_GR + 15 * 8
+	.cfi_offset	16, B_GR + 16 * 8
+	.cfi_offset	17, B_GR + 17 * 8
+	.cfi_offset	18, B_GR + 18 * 8
+	.cfi_offset	19, B_GR + 19 * 8
+	.cfi_offset	20, B_GR + 20 * 8
+	.cfi_offset	21, B_GR + 21 * 8
+	.cfi_offset	22, B_GR + 22 * 8
+	.cfi_offset	23, B_GR + 23 * 8
+	.cfi_offset	24, B_GR + 24 * 8
+	.cfi_offset	25, B_GR + 25 * 8
+	.cfi_offset	26, B_GR + 26 * 8
+	.cfi_offset	27, B_GR + 27 * 8
+	.cfi_offset	28, B_GR + 28 * 8
+	.cfi_offset	29, B_GR + 29 * 8
+	.cfi_offset	30, B_GR + 30 * 8
+	.cfi_offset	31, B_GR + 31 * 8
+
+	/* Floating point registers */
+	.cfi_offset	32, B_FR + 0
+	.cfi_offset	33, B_FR + 1 * 8
+	.cfi_offset	34, B_FR + 2 * 8
+	.cfi_offset	35, B_FR + 3 * 8
+	.cfi_offset	36, B_FR + 4 * 8
+	.cfi_offset	37, B_FR + 5 * 8
+	.cfi_offset	38, B_FR + 6 * 8
+	.cfi_offset	39, B_FR + 7 * 8
+	.cfi_offset	40, B_FR + 8 * 8
+	.cfi_offset	41, B_FR + 9 * 8
+	.cfi_offset	42, B_FR + 10 * 8
+	.cfi_offset	43, B_FR + 11 * 8
+	.cfi_offset	44, B_FR + 12 * 8
+	.cfi_offset	45, B_FR + 13 * 8
+	.cfi_offset	46, B_FR + 14 * 8
+	.cfi_offset	47, B_FR + 15 * 8
+	.cfi_offset	48, B_FR + 16 * 8
+	.cfi_offset	49, B_FR + 17 * 8
+	.cfi_offset	50, B_FR + 18 * 8
+	.cfi_offset	51, B_FR + 19 * 8
+	.cfi_offset	52, B_FR + 20 * 8
+	.cfi_offset	53, B_FR + 21 * 8
+	.cfi_offset	54, B_FR + 22 * 8
+	.cfi_offset	55, B_FR + 23 * 8
+	.cfi_offset	56, B_FR + 24 * 8
+	.cfi_offset	57, B_FR + 25 * 8
+	.cfi_offset	58, B_FR + 26 * 8
+	.cfi_offset	59, B_FR + 27 * 8
+	.cfi_offset	60, B_FR + 28 * 8
+	.cfi_offset	61, B_FR + 29 * 8
+	.cfi_offset	62, B_FR + 30 * 8
+	.cfi_offset	63, B_FR + 31 * 8
+
+	nop
+
+__vdso_rt_sigreturn:
+	li.w	$a7, __NR_rt_sigreturn
+	syscall	0
+	.cfi_endproc
+endf __vdso_rt_sigreturn

linux-user/loongarch64/vdso.ld

@@ -0,0 +1,73 @@
+/*
+ * Linker script for linux loongarch64 replacement vdso.
+ *
+ * Copyright 2023 Linaro, Ltd.
+ *
+ * SPDX-License-Identifier: GPL-2.0-or-later
+ */
+
+VERSION {
+        LINUX_5.10 {
+        global:
+                __vdso_getcpu;
+                __vdso_clock_getres;
+                __vdso_clock_gettime;
+                __vdso_gettimeofday;
+                __vdso_rt_sigreturn;
+
+        local: *;
+        };
+}
+
+
+PHDRS {
+        phdr            PT_PHDR         FLAGS(4) PHDRS;
+        load            PT_LOAD         FLAGS(7) FILEHDR PHDRS;
+        dynamic         PT_DYNAMIC      FLAGS(4);
+        eh_frame_hdr    PT_GNU_EH_FRAME;
+        note            PT_NOTE         FLAGS(4);
+}
+
+SECTIONS {
+        /*
+         * We can't prelink to any address without knowing something about
+         * the virtual memory space of the host, since that leaks over into
+         * the available memory space of the guest.
+         */
+        . = SIZEOF_HEADERS;
+
+        /*
+         * The following, including the FILEHDRS and PHDRS, are modified
+         * when we relocate the binary.  We want them to be initially
+         * writable for the relocation; we'll force them read-only after.
+         */
+        .note           : { *(.note*) }         :load :note
+        .dynamic        : { *(.dynamic) }       :load :dynamic
+        .dynsym         : { *(.dynsym) }        :load
+        /*
+         * There ought not be any real read-write data.
+         * But since we manipulated the segment layout,
+         * we have to put these sections somewhere.
+         */
+        .data           : {
+                *(.data*)
+                *(.sdata*)
+                *(.got.plt) *(.got)
+                *(.gnu.linkonce.d.*)
+                *(.bss*)
+                *(.dynbss*)
+                *(.gnu.linkonce.b.*)
+        }
+
+        .rodata         : { *(.rodata*) }
+        .hash           : { *(.hash) }
+        .gnu.hash       : { *(.gnu.hash) }
+        .dynstr         : { *(.dynstr) }
+        .gnu.version    : { *(.gnu.version) }
+        .gnu.version_d  : { *(.gnu.version_d) }
+        .gnu.version_r  : { *(.gnu.version_r) }
+        .eh_frame_hdr   : { *(.eh_frame_hdr) }  :load :eh_frame_hdr
+        .eh_frame       : { *(.eh_frame) }      :load
+
+        .text           : { *(.text*) }         :load   =0xd503201f
+}

linux-user/meson.build

@@ -28,18 +28,25 @@ linux_user_ss.add(when: 'TARGET_HAS_BFLT', if_true: files('flatload.c'))
 linux_user_ss.add(when: 'TARGET_I386', if_true: files('vm86.c'))
 linux_user_ss.add(when: 'CONFIG_ARM_COMPATIBLE_SEMIHOSTING', if_true: files('semihost.c'))
 
-
 syscall_nr_generators = {}
 
+gen_vdso_exe = executable('gen-vdso', 'gen-vdso.c',
+                          native: true, build_by_default: false)
+gen_vdso = generator(gen_vdso_exe, output: '@BASENAME@.c.inc',
+                     arguments: ['-o', '@OUTPUT@', '@EXTRA_ARGS@', '@INPUT@'])
+
+subdir('aarch64')
 subdir('alpha')
 subdir('arm')
 subdir('hppa')
 subdir('i386')
+subdir('loongarch64')
 subdir('m68k')
 subdir('microblaze')
 subdir('mips64')
 subdir('mips')
 subdir('ppc')
+subdir('riscv')
 subdir('s390x')
 subdir('sh4')
 subdir('sparc')

linux-user/ppc/Makefile.vdso

@@ -0,0 +1,20 @@
+include $(BUILD_DIR)/tests/tcg/ppc64-linux-user/config-target.mak
+
+SUBDIR = $(SRC_PATH)/linux-user/ppc
+VPATH += $(SUBDIR)
+
+all: $(SUBDIR)/vdso-32.so $(SUBDIR)/vdso-64.so $(SUBDIR)/vdso-64le.so
+
+LDFLAGS32 = -nostdlib -shared -Wl,-T,$(SUBDIR)/vdso-32.ld \
+            -Wl,-h,linux-vdso32.so.1 -Wl,--hash-style=both -Wl,--build-id=sha1
+LDFLAGS64 = -nostdlib -shared -Wl,-T,$(SUBDIR)/vdso-64.ld \
+            -Wl,-h,linux-vdso64.so.1 -Wl,--hash-style=both -Wl,--build-id=sha1
+
+$(SUBDIR)/vdso-32.so: vdso.S vdso-32.ld vdso-asmoffset.h
+	$(CC) -o $@ $(LDFLAGS32) -m32 $<
+
+$(SUBDIR)/vdso-64.so: vdso.S vdso-64.ld vdso-asmoffset.h
+	$(CC) -o $@ $(LDFLAGS64) -mbig-endian $<
+
+$(SUBDIR)/vdso-64le.so: vdso.S vdso-64.ld vdso-asmoffset.h
+	$(CC) -o $@ $(LDFLAGS64) -mlittle-endian $<

linux-user/ppc/meson.build

@@ -3,3 +3,15 @@ syscall_nr_generators += {
                    arguments: [ meson.current_source_dir() / 'syscallhdr.sh', '@INPUT@', '@OUTPUT@', '@EXTRA_ARGS@' ],
                    output: '@BASENAME@_nr.h')
 }
+
+vdso_32_inc = gen_vdso.process('vdso-32.so', extra_args: [
+                                   '-s', '__kernel_sigtramp32',
+                                   '-r', '__kernel_sigtramp_rt32'
+                               ])
+linux_user_ss.add(when: 'TARGET_PPC', if_true: vdso_32_inc)
+
+vdso_64_inc = gen_vdso.process('vdso-64.so',
+                               extra_args: ['-r', '__kernel_sigtramp_rt64'])
+vdso_64le_inc = gen_vdso.process('vdso-64le.so',
+                                 extra_args: ['-r', '__kernel_sigtramp_rt64'])
+linux_user_ss.add(when: 'TARGET_PPC64', if_true: [vdso_64_inc, vdso_64le_inc])

linux-user/ppc/signal.c

@@ -21,14 +21,7 @@
 #include "user-internals.h"
 #include "signal-common.h"
 #include "linux-user/trace.h"
-
+#include "vdso-asmoffset.h"
-/* Size of dummy stack frame allocated when calling signal handler.
-   See arch/powerpc/include/asm/ptrace.h.  */
-#if defined(TARGET_PPC64)
-#define SIGNAL_FRAMESIZE 128
-#else
-#define SIGNAL_FRAMESIZE 64
-#endif
 
 /* See arch/powerpc/include/asm/ucontext.h.  Only used for 32-bit PPC;
    on 64-bit PPC, sigcontext and mcontext are one and the same.  */
@@ -73,6 +66,16 @@ struct target_mcontext {
 #endif
 };
 
+QEMU_BUILD_BUG_ON(offsetof(struct target_mcontext, mc_fregs)
+                  != offsetof_mcontext_fregs);
+#if defined(TARGET_PPC64)
+QEMU_BUILD_BUG_ON(offsetof(struct target_mcontext, v_regs)
+                  != offsetof_mcontext_vregs_ptr);
+#else
+QEMU_BUILD_BUG_ON(offsetof(struct target_mcontext, mc_vregs)
+                  != offsetof_mcontext_vregs);
+#endif
+
 /* See arch/powerpc/include/asm/sigcontext.h.  */
 struct target_sigcontext {
     target_ulong _unused[4];
@@ -161,6 +164,7 @@ struct target_ucontext {
 #endif
 };
 
+#if !defined(TARGET_PPC64)
 /* See arch/powerpc/kernel/signal_32.c.  */
 struct target_sigframe {
     struct target_sigcontext sctx;
@@ -168,6 +172,10 @@ struct target_sigframe {
     int32_t abigap[56];
 };
 
+QEMU_BUILD_BUG_ON(offsetof(struct target_sigframe, mctx)
+                  != offsetof_sigframe_mcontext);
+#endif
+
 #if defined(TARGET_PPC64)
 
 #define TARGET_TRAMP_SIZE 6
@@ -184,6 +192,10 @@ struct target_rt_sigframe {
     char abigap[288];
 } __attribute__((aligned(16)));
 
+QEMU_BUILD_BUG_ON(offsetof(struct target_rt_sigframe,
+                           uc.tuc_sigcontext.mcontext)
+                  != offsetof_rt_sigframe_mcontext);
+
 #else
 
 struct target_rt_sigframe {
@@ -192,6 +204,9 @@ struct target_rt_sigframe {
     int32_t abigap[56];
 };
 
+QEMU_BUILD_BUG_ON(offsetof(struct target_rt_sigframe, uc.tuc_mcontext)
+                  != offsetof_rt_sigframe_mcontext);
+
 #endif
 
 #if defined(TARGET_PPC64)

linux-user/ppc/vdso-32.ld

@@ -0,0 +1,70 @@
+/*
+ * Linker script for linux powerpc64 replacement vdso.
+ *
+ * Copyright 2023 Linaro, Ltd.
+ *
+ * SPDX-License-Identifier: GPL-2.0-or-later
+ */
+
+VERSION {
+        LINUX_2.6.15 {
+        global:
+                __kernel_gettimeofday;
+                __kernel_clock_gettime;
+                __kernel_clock_gettime64;
+                __kernel_clock_getres;
+                __kernel_time;
+                __kernel_sync_dicache;
+                __kernel_sigtramp32;
+                __kernel_sigtramp_rt32;
+                __kernel_getcpu;
+        local: *;
+        };
+}
+
+PHDRS {
+        phdr            PT_PHDR         FLAGS(4) PHDRS;
+        load            PT_LOAD         FLAGS(7) FILEHDR PHDRS; /* FLAGS=RWX */
+        dynamic         PT_DYNAMIC      FLAGS(4);
+        eh_frame_hdr    PT_GNU_EH_FRAME;
+        note            PT_NOTE         FLAGS(4);
+}
+
+SECTIONS {
+        . = SIZEOF_HEADERS;
+
+        /*
+         * The following, including the FILEHDRS and PHDRS, are modified
+         * when we relocate the binary.  We want them to be initially
+         * writable for the relocation; we'll force them read-only after.
+         */
+        .note           : { *(.note*) }         :load :note
+        .dynamic        : { *(.dynamic) }       :load :dynamic
+        .dynsym         : { *(.dynsym) }        :load
+        .data           : {
+                /*
+                 * There ought not be any real read-write data.
+                 * But since we manipulated the segment layout,
+                 * we have to put these sections somewhere.
+                 */
+                *(.data*)
+                *(.sdata*)
+                *(.got.plt) *(.got)
+                *(.gnu.linkonce.d.*)
+                *(.bss*)
+                *(.dynbss*)
+                *(.gnu.linkonce.b.*)
+        }
+
+        .rodata         : { *(.rodata*) }
+        .hash           : { *(.hash) }
+        .gnu.hash       : { *(.gnu.hash) }
+        .dynstr         : { *(.dynstr) }
+        .gnu.version    : { *(.gnu.version) }
+        .gnu.version_d  : { *(.gnu.version_d) }
+        .gnu.version_r  : { *(.gnu.version_r) }
+        .eh_frame_hdr   : { *(.eh_frame_hdr) }  :load :eh_frame_hdr
+        .eh_frame       : { *(.eh_frame) }      :load
+
+        .text           : { *(.text*) }         :load
+}

linux-user/ppc/vdso-64.ld

@@ -0,0 +1,68 @@
+/*
+ * Linker script for linux powerpc64 replacement vdso.
+ *
+ * Copyright 2023 Linaro, Ltd.
+ *
+ * SPDX-License-Identifier: GPL-2.0-or-later
+ */
+
+VERSION {
+        LINUX_2.6.15 {
+        global:
+                __kernel_gettimeofday;
+                __kernel_clock_gettime;
+                __kernel_clock_getres;
+                __kernel_sync_dicache;
+                __kernel_sigtramp_rt64;
+                __kernel_getcpu;
+                __kernel_time;
+        local: *;
+        };
+}
+
+PHDRS {
+        phdr            PT_PHDR         FLAGS(4) PHDRS;
+        load            PT_LOAD         FLAGS(7) FILEHDR PHDRS; /* FLAGS=RWX */
+        dynamic         PT_DYNAMIC      FLAGS(4);
+        eh_frame_hdr    PT_GNU_EH_FRAME;
+        note            PT_NOTE         FLAGS(4);
+}
+
+SECTIONS {
+        . = SIZEOF_HEADERS;
+
+        /*
+         * The following, including the FILEHDRS and PHDRS, are modified
+         * when we relocate the binary.  We want them to be initially
+         * writable for the relocation; we'll force them read-only after.
+         */
+        .note           : { *(.note*) }         :load :note
+        .dynamic        : { *(.dynamic) }       :load :dynamic
+        .dynsym         : { *(.dynsym) }        :load
+        .data           : {
+                /*
+                 * There ought not be any real read-write data.
+                 * But since we manipulated the segment layout,
+                 * we have to put these sections somewhere.
+                 */
+                *(.data*)
+                *(.sdata*)
+                *(.got.plt) *(.got)
+                *(.gnu.linkonce.d.*)
+                *(.bss*)
+                *(.dynbss*)
+                *(.gnu.linkonce.b.*)
+        }
+
+        .rodata         : { *(.rodata*) }
+        .hash           : { *(.hash) }
+        .gnu.hash       : { *(.gnu.hash) }
+        .dynstr         : { *(.dynstr) }
+        .gnu.version    : { *(.gnu.version) }
+        .gnu.version_d  : { *(.gnu.version_d) }
+        .gnu.version_r  : { *(.gnu.version_r) }
+        .eh_frame_hdr   : { *(.eh_frame_hdr) }  :load :eh_frame_hdr
+        .eh_frame       : { *(.eh_frame) }      :load
+
+        .text           : { *(.text*) }         :load
+}

linux-user/ppc/vdso-asmoffset.h

@@ -0,0 +1,20 @@
+/*
+ * Size of dummy stack frame allocated when calling signal handler.
+ * See arch/powerpc/include/asm/ptrace.h.
+ */
+#ifdef TARGET_ABI32
+# define SIGNAL_FRAMESIZE                   64
+#else
+# define SIGNAL_FRAMESIZE                   128
+#endif
+
+#ifdef TARGET_ABI32
+# define offsetof_sigframe_mcontext         0x20
+# define offsetof_rt_sigframe_mcontext      0x140
+# define offsetof_mcontext_fregs            0xc0
+# define offsetof_mcontext_vregs            0x1d0
+#else
+# define offsetof_rt_sigframe_mcontext      0xe8
+# define offsetof_mcontext_fregs            0x180
+# define offsetof_mcontext_vregs_ptr        0x288
+#endif

linux-user/ppc/vdso.S

@@ -0,0 +1,239 @@
+/*
+ * PowerPC linux replacement vdso.
+ *
+ * Copyright 2023 Linaro, Ltd.
+ *
+ * SPDX-License-Identifier: GPL-2.0-or-later
+ */
+
+#include <asm/unistd.h>
+#include <asm/errno.h>
+
+#ifndef _ARCH_PPC64
+# define TARGET_ABI32
+#endif
+#include "vdso-asmoffset.h"
+
+
+	.text
+
+.macro endf name
+	.globl	\name
+	.size	\name, .-\name
+	/* For PPC64, functions have special linkage; we export pointers. */
+#ifndef _ARCH_PPC64
+	.type	\name, @function
+#endif
+.endm
+
+.macro raw_syscall nr
+	addi	0, 0, \nr
+	sc
+.endm
+
+.macro vdso_syscall name, nr
+\name:
+	raw_syscall \nr
+	blr
+endf	\name
+.endm
+
+	.cfi_startproc
+
+vdso_syscall __kernel_gettimeofday, __NR_gettimeofday
+vdso_syscall __kernel_clock_gettime, __NR_clock_gettime
+vdso_syscall __kernel_clock_getres, __NR_clock_getres
+vdso_syscall __kernel_getcpu, __NR_getcpu
+vdso_syscall __kernel_time, __NR_time
+
+#ifdef __NR_clock_gettime64
+vdso_syscall __kernel_clock_gettime64, __NR_clock_gettime64
+#endif
+
+__kernel_sync_dicache:
+	/* qemu does not need to flush caches */
+	blr
+endf	__kernel_sync_dicache
+
+	.cfi_endproc
+
+/*
+ * TODO: __kernel_get_tbfreq
+ * This is probably a constant for QEMU.
+ */
+
+/*
+ * Start the unwind info at least one instruction before the signal
+ * trampoline, because the unwinder will assume we are returning
+ * after a call site.
+ */
+
+	.cfi_startproc simple
+	.cfi_signal_frame
+
+#ifdef _ARCH_PPC64
+# define __kernel_sigtramp_rt  __kernel_sigtramp_rt64
+# define sizeof_reg	8
+#else
+# define __kernel_sigtramp_rt  __kernel_sigtramp_rt32
+# define sizeof_reg	4
+#endif
+#define sizeof_freg	8
+#define sizeof_vreg	16
+
+	.cfi_def_cfa	1, SIGNAL_FRAMESIZE + offsetof_rt_sigframe_mcontext
+
+	/* Return address */
+	.cfi_return_column 67
+	.cfi_offset	67, 32 * sizeof_reg		/* nip */
+
+	/* Integer registers */
+	.cfi_offset	0, 0 * sizeof_reg
+	.cfi_offset	1, 1 * sizeof_reg
+	.cfi_offset	2, 2 * sizeof_reg
+	.cfi_offset	3, 3 * sizeof_reg
+	.cfi_offset	4, 4 * sizeof_reg
+	.cfi_offset	5, 5 * sizeof_reg
+	.cfi_offset	6, 6 * sizeof_reg
+	.cfi_offset	7, 7 * sizeof_reg
+	.cfi_offset	8, 8 * sizeof_reg
+	.cfi_offset	9, 9 * sizeof_reg
+	.cfi_offset	10, 10 * sizeof_reg
+	.cfi_offset	11, 11 * sizeof_reg
+	.cfi_offset	12, 12 * sizeof_reg
+	.cfi_offset	13, 13 * sizeof_reg
+	.cfi_offset	14, 14 * sizeof_reg
+	.cfi_offset	15, 15 * sizeof_reg
+	.cfi_offset	16, 16 * sizeof_reg
+	.cfi_offset	17, 17 * sizeof_reg
+	.cfi_offset	18, 18 * sizeof_reg
+	.cfi_offset	19, 19 * sizeof_reg
+	.cfi_offset	20, 20 * sizeof_reg
+	.cfi_offset	21, 21 * sizeof_reg
+	.cfi_offset	22, 22 * sizeof_reg
+	.cfi_offset	23, 23 * sizeof_reg
+	.cfi_offset	24, 24 * sizeof_reg
+	.cfi_offset	25, 25 * sizeof_reg
+	.cfi_offset	26, 26 * sizeof_reg
+	.cfi_offset	27, 27 * sizeof_reg
+	.cfi_offset	28, 28 * sizeof_reg
+	.cfi_offset	29, 29 * sizeof_reg
+	.cfi_offset	30, 30 * sizeof_reg
+	.cfi_offset	31, 31 * sizeof_reg
+	.cfi_offset	65, 36 * sizeof_reg		/* lr */
+	.cfi_offset	70, 38 * sizeof_reg		/* ccr */
+
+	/* Floating point registers */
+	.cfi_offset	32, offsetof_mcontext_fregs
+	.cfi_offset	33, offsetof_mcontext_fregs + 1 * sizeof_freg
+	.cfi_offset	34, offsetof_mcontext_fregs + 2 * sizeof_freg
+	.cfi_offset	35, offsetof_mcontext_fregs + 3 * sizeof_freg
+	.cfi_offset	36, offsetof_mcontext_fregs + 4 * sizeof_freg
+	.cfi_offset	37, offsetof_mcontext_fregs + 5 * sizeof_freg
+	.cfi_offset	38, offsetof_mcontext_fregs + 6 * sizeof_freg
+	.cfi_offset	39, offsetof_mcontext_fregs + 7 * sizeof_freg
+	.cfi_offset	40, offsetof_mcontext_fregs + 8 * sizeof_freg
+	.cfi_offset	41, offsetof_mcontext_fregs + 9 * sizeof_freg
+	.cfi_offset	42, offsetof_mcontext_fregs + 10 * sizeof_freg
+	.cfi_offset	43, offsetof_mcontext_fregs + 11 * sizeof_freg
+	.cfi_offset	44, offsetof_mcontext_fregs + 12 * sizeof_freg
+	.cfi_offset	45, offsetof_mcontext_fregs + 13 * sizeof_freg
+	.cfi_offset	46, offsetof_mcontext_fregs + 14 * sizeof_freg
+	.cfi_offset	47, offsetof_mcontext_fregs + 15 * sizeof_freg
+	.cfi_offset	48, offsetof_mcontext_fregs + 16 * sizeof_freg
+	.cfi_offset	49, offsetof_mcontext_fregs + 17 * sizeof_freg
+	.cfi_offset	50, offsetof_mcontext_fregs + 18 * sizeof_freg
+	.cfi_offset	51, offsetof_mcontext_fregs + 19 * sizeof_freg
+	.cfi_offset	52, offsetof_mcontext_fregs + 20 * sizeof_freg
+	.cfi_offset	53, offsetof_mcontext_fregs + 21 * sizeof_freg
+	.cfi_offset	54, offsetof_mcontext_fregs + 22 * sizeof_freg
+	.cfi_offset	55, offsetof_mcontext_fregs + 23 * sizeof_freg
+	.cfi_offset	56, offsetof_mcontext_fregs + 24 * sizeof_freg
+	.cfi_offset	57, offsetof_mcontext_fregs + 25 * sizeof_freg
+	.cfi_offset	58, offsetof_mcontext_fregs + 26 * sizeof_freg
+	.cfi_offset	59, offsetof_mcontext_fregs + 27 * sizeof_freg
+	.cfi_offset	60, offsetof_mcontext_fregs + 28 * sizeof_freg
+	.cfi_offset	61, offsetof_mcontext_fregs + 29 * sizeof_freg
+	.cfi_offset	62, offsetof_mcontext_fregs + 30 * sizeof_freg
+	.cfi_offset	63, offsetof_mcontext_fregs + 31 * sizeof_freg
+
+	/*
+	 * Unlike the kernel, unconditionally represent the Altivec/VSX regs.
+	 * The space within the stack frame is always available, and most of
+	 * our supported processors have them enabled.  The only complication
+	 * for PPC64 is the misalignment, so that we have to use indirection.
+	 */
+.macro	save_vreg_ofs reg, ofs
+#ifdef _ARCH_PPC64
+	/*
+	 * vreg = *(cfa + offsetof(v_regs)) + ofs
+         *
+         * The CFA is input to the expression on the stack, so:
+	 * DW_CFA_expression reg, length (7),
+         *   DW_OP_plus_uconst (0x23), vreg_ptr, DW_OP_deref (0x06),
+	 *   DW_OP_plus_uconst (0x23), ofs
+	 */
+	.cfi_escape 0x10, 77 + \reg, 7, 0x23, (offsetof_mcontext_vregs_ptr & 0x7f) + 0x80, offsetof_mcontext_vregs_ptr >> 7, 0x06, 0x23, (\ofs & 0x7f) | 0x80, \ofs >> 7
+#else
+	.cfi_offset 77 + \reg, offsetof_mcontext_vregs + \ofs
+#endif
+.endm
+
+.macro	save_vreg reg
+	save_vreg_ofs \reg, (\reg * sizeof_vreg)
+.endm
+
+	save_vreg   0
+	save_vreg   1
+	save_vreg   2
+	save_vreg   3
+	save_vreg   4
+	save_vreg   5
+	save_vreg   6
+	save_vreg   7
+	save_vreg   8
+	save_vreg   9
+	save_vreg  10
+	save_vreg  11
+	save_vreg  12
+	save_vreg  13
+	save_vreg  14
+	save_vreg  15
+	save_vreg  16
+	save_vreg  17
+	save_vreg  18
+	save_vreg  19
+	save_vreg  20
+	save_vreg  21
+	save_vreg  22
+	save_vreg  23
+	save_vreg  24
+	save_vreg  25
+	save_vreg  26
+	save_vreg  27
+	save_vreg  28
+	save_vreg  29
+	save_vreg  30
+	save_vreg  31
+	save_vreg  32
+	save_vreg_ofs 33, (32 * sizeof_vreg + 12)
+
+	nop
+
+__kernel_sigtramp_rt:
+	raw_syscall __NR_rt_sigreturn
+endf	__kernel_sigtramp_rt
+
+#ifndef _ARCH_PPC64
+	/*
+	 * The non-rt sigreturn has the same layout at a different offset.
+	 * Move the CFA and leave all othe other descriptions the same.
+	 */
+	.cfi_def_cfa	1, SIGNAL_FRAMESIZE + offsetof_sigframe_mcontext
+	nop
+__kernel_sigtramp32:
+	raw_syscall __NR_sigreturn
+endf	__kernel_sigtramp32
+#endif
+
+	.cfi_endproc

linux-user/qemu.h

@@ -32,6 +32,7 @@ struct image_info {
         abi_ulong       brk;
         abi_ulong       start_stack;
         abi_ulong       stack_limit;
+        abi_ulong       vdso;
         abi_ulong       entry;
         abi_ulong       code_offset;
         abi_ulong       data_offset;

linux-user/riscv/Makefile.vdso

@@ -0,0 +1,15 @@
+include $(BUILD_DIR)/tests/tcg/riscv64-linux-user/config-target.mak
+
+SUBDIR = $(SRC_PATH)/linux-user/riscv
+VPATH += $(SUBDIR)
+
+all: $(SUBDIR)/vdso-32.so $(SUBDIR)/vdso-64.so
+
+LDFLAGS = -nostdlib -shared -fpic -Wl,-h,linux-vdso.so.1 -Wl,--build-id=sha1 \
+	  -Wl,--hash-style=both -Wl,-T,$(SUBDIR)/vdso.ld
+
+$(SUBDIR)/vdso-32.so: vdso.S vdso.ld vdso-asmoffset.h
+	$(CC) -o $@ $(LDFLAGS) -mabi=ilp32d -march=rv32g $<
+
+$(SUBDIR)/vdso-64.so: vdso.S vdso.ld vdso-asmoffset.h
+	$(CC) -o $@ $(LDFLAGS) -mabi=lp64d -march=rv64g $<

linux-user/riscv/meson.build

@@ -0,0 +1,7 @@
+vdso_32_inc = gen_vdso.process('vdso-32.so',
+                               extra_args: ['-r', '__vdso_rt_sigreturn'])
+vdso_64_inc = gen_vdso.process('vdso-64.so',
+                               extra_args: ['-r', '__vdso_rt_sigreturn'])
+
+linux_user_ss.add(when: 'TARGET_RISCV32', if_true: vdso_32_inc)
+linux_user_ss.add(when: 'TARGET_RISCV64', if_true: vdso_64_inc)

linux-user/riscv/signal.c

@@ -21,6 +21,7 @@
 #include "user-internals.h"
 #include "signal-common.h"
 #include "linux-user/trace.h"
+#include "vdso-asmoffset.h"
 
 /* Signal handler invocation must be transparent for the code being
    interrupted. Complete CPU (hart) state is saved on entry and restored
@@ -37,6 +38,8 @@ struct target_sigcontext {
     uint32_t fcsr;
 }; /* cf. riscv-linux:arch/riscv/include/uapi/asm/ptrace.h */
 
+QEMU_BUILD_BUG_ON(offsetof(struct target_sigcontext, fpr) != offsetof_freg0);
+
 struct target_ucontext {
     abi_ulong uc_flags;
     abi_ptr uc_link;
@@ -51,6 +54,11 @@ struct target_rt_sigframe {
     struct target_ucontext uc;
 };
 
+QEMU_BUILD_BUG_ON(sizeof(struct target_rt_sigframe)
+                  != sizeof_rt_sigframe);
+QEMU_BUILD_BUG_ON(offsetof(struct target_rt_sigframe, uc.uc_mcontext)
+                  != offsetof_uc_mcontext);
+
 static abi_ulong get_sigframe(struct target_sigaction *ka,
                               CPURISCVState *regs, size_t framesize)
 {

linux-user/riscv/vdso-asmoffset.h

@@ -0,0 +1,9 @@
+#ifdef TARGET_ABI32
+# define sizeof_rt_sigframe     0x2b0
+# define offsetof_uc_mcontext   0x120
+# define offsetof_freg0         0x80
+#else
+# define sizeof_rt_sigframe     0x340
+# define offsetof_uc_mcontext   0x130
+# define offsetof_freg0         0x100
+#endif

linux-user/riscv/vdso.S

@@ -0,0 +1,187 @@
+/*
+ * RISC-V linux replacement vdso.
+ *
+ * Copyright 2021 Linaro, Ltd.
+ *
+ * SPDX-License-Identifier: GPL-2.0-or-later
+ */
+
+#include <asm/unistd.h>
+#include <asm/errno.h>
+
+#if __riscv_xlen == 32
+# define TARGET_ABI32
+#endif
+#include "vdso-asmoffset.h"
+
+	.text
+
+.macro endf name
+	.globl	\name
+	.type	\name, @function
+	.size	\name, . - \name
+.endm
+
+.macro raw_syscall nr
+	li	a7, \nr
+	ecall
+.endm
+
+.macro vdso_syscall name, nr
+\name:
+	raw_syscall \nr
+	ret
+endf	\name
+.endm
+
+__vdso_gettimeofday:
+	.cfi_startproc
+#ifdef __NR_gettimeofday
+	raw_syscall __NR_gettimeofday
+	ret
+#else
+	/* No gettimeofday, fall back to clock_gettime64. */
+	beq	a1, zero, 1f
+	sw	zero, 0(a1)	/* tz->tz_minuteswest = 0 */
+	sw	zero, 4(a1)	/* tz->tz_dsttime = 0 */
+1:	addi	sp, sp, -32
+	.cfi_adjust_cfa_offset 32
+	sw	a0, 16(sp)	/* save tv */
+	mv	a0, sp
+	raw_syscall __NR_clock_gettime64
+	lw	t0, 0(sp)	/* timespec.tv_sec.low */
+	lw	t1, 4(sp)	/* timespec.tv_sec.high */
+	lw	t2, 8(sp)	/* timespec.tv_nsec.low */
+	lw	a1, 16(sp)	/* restore tv */
+	addi	sp, sp, 32
+	.cfi_adjust_cfa_offset -32
+	bne	a0, zero, 9f	/* syscall error? */
+	li	a0, -EOVERFLOW
+	bne	t1, zero, 9f	/* y2038? */
+	li	a0, 0
+	li	t3, 1000
+	divu	t2, t2, t3	/* nsec -> usec */
+	sw	t0, 0(a1)	/* tz->tv_sec */
+	sw	t2, 4(a1)	/* tz->tv_usec */
+9:	ret
+#endif
+	.cfi_endproc
+endf __vdso_gettimeofday
+
+	.cfi_startproc
+
+#ifdef __NR_clock_gettime
+vdso_syscall __vdso_clock_gettime, __NR_clock_gettime
+#else
+vdso_syscall __vdso_clock_gettime, __NR_clock_gettime64
+#endif
+
+#ifdef __NR_clock_getres
+vdso_syscall __vdso_clock_getres, __NR_clock_getres
+#else
+vdso_syscall __vdso_clock_getres, __NR_clock_getres_time64
+#endif
+
+vdso_syscall __vdso_getcpu, __NR_getcpu
+
+__vdso_flush_icache:
+	/* qemu does not need to flush the icache */
+	li	a0, 0
+	ret
+endf __vdso_flush_icache
+
+	.cfi_endproc
+
+/*
+ * Start the unwind info at least one instruction before the signal
+ * trampoline, because the unwinder will assume we are returning
+ * after a call site.
+ */
+
+	.cfi_startproc simple
+	.cfi_signal_frame
+
+#define sizeof_reg	(__riscv_xlen / 4)
+#define sizeof_freg	8
+#define B_GR	(offsetof_uc_mcontext - sizeof_rt_sigframe)
+#define B_FR	(offsetof_uc_mcontext - sizeof_rt_sigframe + offsetof_freg0)
+
+	.cfi_def_cfa	2, sizeof_rt_sigframe
+
+	/* Return address */
+	.cfi_return_column 64
+	.cfi_offset	64, B_GR + 0			/* pc */
+
+	/* Integer registers */
+	.cfi_offset	1, B_GR + 1 * sizeof_reg	/* r1 (ra) */
+	.cfi_offset	2, B_GR + 2 * sizeof_reg	/* r2 (sp) */
+	.cfi_offset	3, B_GR + 3 * sizeof_reg
+	.cfi_offset	4, B_GR + 4 * sizeof_reg
+	.cfi_offset	5, B_GR + 5 * sizeof_reg
+	.cfi_offset	6, B_GR + 6 * sizeof_reg
+	.cfi_offset	7, B_GR + 7 * sizeof_reg
+	.cfi_offset	8, B_GR + 8 * sizeof_reg
+	.cfi_offset	9, B_GR + 9 * sizeof_reg
+	.cfi_offset	10, B_GR + 10 * sizeof_reg
+	.cfi_offset	11, B_GR + 11 * sizeof_reg
+	.cfi_offset	12, B_GR + 12 * sizeof_reg
+	.cfi_offset	13, B_GR + 13 * sizeof_reg
+	.cfi_offset	14, B_GR + 14 * sizeof_reg
+	.cfi_offset	15, B_GR + 15 * sizeof_reg
+	.cfi_offset	16, B_GR + 16 * sizeof_reg
+	.cfi_offset	17, B_GR + 17 * sizeof_reg
+	.cfi_offset	18, B_GR + 18 * sizeof_reg
+	.cfi_offset	19, B_GR + 19 * sizeof_reg
+	.cfi_offset	20, B_GR + 20 * sizeof_reg
+	.cfi_offset	21, B_GR + 21 * sizeof_reg
+	.cfi_offset	22, B_GR + 22 * sizeof_reg
+	.cfi_offset	23, B_GR + 23 * sizeof_reg
+	.cfi_offset	24, B_GR + 24 * sizeof_reg
+	.cfi_offset	25, B_GR + 25 * sizeof_reg
+	.cfi_offset	26, B_GR + 26 * sizeof_reg
+	.cfi_offset	27, B_GR + 27 * sizeof_reg
+	.cfi_offset	28, B_GR + 28 * sizeof_reg
+	.cfi_offset	29, B_GR + 29 * sizeof_reg
+	.cfi_offset	30, B_GR + 30 * sizeof_reg
+	.cfi_offset	31, B_GR + 31 * sizeof_reg	/* r31 */
+
+	.cfi_offset	32, B_FR + 0			/* f0 */
+	.cfi_offset	33, B_FR + 1 * sizeof_freg	/* f1 */
+	.cfi_offset	34, B_FR + 2 * sizeof_freg
+	.cfi_offset	35, B_FR + 3 * sizeof_freg
+	.cfi_offset	36, B_FR + 4 * sizeof_freg
+	.cfi_offset	37, B_FR + 5 * sizeof_freg
+	.cfi_offset	38, B_FR + 6 * sizeof_freg
+	.cfi_offset	39, B_FR + 7 * sizeof_freg
+	.cfi_offset	40, B_FR + 8 * sizeof_freg
+	.cfi_offset	41, B_FR + 9 * sizeof_freg
+	.cfi_offset	42, B_FR + 10 * sizeof_freg
+	.cfi_offset	43, B_FR + 11 * sizeof_freg
+	.cfi_offset	44, B_FR + 12 * sizeof_freg
+	.cfi_offset	45, B_FR + 13 * sizeof_freg
+	.cfi_offset	46, B_FR + 14 * sizeof_freg
+	.cfi_offset	47, B_FR + 15 * sizeof_freg
+	.cfi_offset	48, B_FR + 16 * sizeof_freg
+	.cfi_offset	49, B_FR + 17 * sizeof_freg
+	.cfi_offset	50, B_FR + 18 * sizeof_freg
+	.cfi_offset	51, B_FR + 19 * sizeof_freg
+	.cfi_offset	52, B_FR + 20 * sizeof_freg
+	.cfi_offset	53, B_FR + 21 * sizeof_freg
+	.cfi_offset	54, B_FR + 22 * sizeof_freg
+	.cfi_offset	55, B_FR + 23 * sizeof_freg
+	.cfi_offset	56, B_FR + 24 * sizeof_freg
+	.cfi_offset	57, B_FR + 25 * sizeof_freg
+	.cfi_offset	58, B_FR + 26 * sizeof_freg
+	.cfi_offset	59, B_FR + 27 * sizeof_freg
+	.cfi_offset	60, B_FR + 28 * sizeof_freg
+	.cfi_offset	61, B_FR + 29 * sizeof_freg
+	.cfi_offset	62, B_FR + 30 * sizeof_freg
+	.cfi_offset	63, B_FR + 31 * sizeof_freg	/* f31 */
+
+	nop
+
+__vdso_rt_sigreturn:
+	raw_syscall __NR_rt_sigreturn
+endf __vdso_rt_sigreturn
+
+	.cfi_endproc

linux-user/riscv/vdso.ld

@@ -0,0 +1,74 @@
+/*
+ * Linker script for linux riscv replacement vdso.
+ *
+ * Copyright 2021 Linaro, Ltd.
+ *
+ * SPDX-License-Identifier: GPL-2.0-or-later
+ */
+
+VERSION {
+        LINUX_4.15 {
+        global:
+                __vdso_rt_sigreturn;
+                __vdso_gettimeofday;
+                __vdso_clock_gettime;
+                __vdso_clock_getres;
+                __vdso_getcpu;
+                __vdso_flush_icache;
+
+        local: *;
+        };
+}
+
+
+PHDRS {
+        phdr            PT_PHDR         FLAGS(4) PHDRS;
+        load            PT_LOAD         FLAGS(7) FILEHDR PHDRS;
+        dynamic         PT_DYNAMIC      FLAGS(4);
+        eh_frame_hdr    PT_GNU_EH_FRAME;
+        note            PT_NOTE         FLAGS(4);
+}
+
+SECTIONS {
+        /*
+         * We can't prelink to any address without knowing something about
+         * the virtual memory space of the host, since that leaks over into
+         * the available memory space of the guest.
+         */
+        . = SIZEOF_HEADERS;
+
+        /*
+         * The following, including the FILEHDRS and PHDRS, are modified
+         * when we relocate the binary.  We want them to be initially
+         * writable for the relocation; we'll force them read-only after.
+         */
+        .note           : { *(.note*) }         :load :note
+        .dynamic        : { *(.dynamic) }       :load :dynamic
+        .dynsym         : { *(.dynsym) }        :load
+        /*
+         * There ought not be any real read-write data.
+         * But since we manipulated the segment layout,
+         * we have to put these sections somewhere.
+         */
+        .data           : {
+                *(.data*)
+                *(.sdata*)
+                *(.got.plt) *(.got)
+                *(.gnu.linkonce.d.*)
+                *(.bss*)
+                *(.dynbss*)
+                *(.gnu.linkonce.b.*)
+        }
+
+        .rodata         : { *(.rodata*) }
+        .hash           : { *(.hash) }
+        .gnu.hash       : { *(.gnu.hash) }
+        .dynstr         : { *(.dynstr) }
+        .gnu.version    : { *(.gnu.version) }
+        .gnu.version_d  : { *(.gnu.version_d) }
+        .gnu.version_r  : { *(.gnu.version_r) }
+        .eh_frame_hdr   : { *(.eh_frame_hdr) }  :load :eh_frame_hdr
+        .eh_frame       : { *(.eh_frame) }      :load
+
+        .text           : { *(.text*) }         :load   =0xd503201f
+}

linux-user/s390x/Makefile.vdso

@@ -0,0 +1,11 @@
+include $(BUILD_DIR)/tests/tcg/s390x-linux-user/config-target.mak
+
+SUBDIR = $(SRC_PATH)/linux-user/s390x
+VPATH += $(SUBDIR)
+
+all: $(SUBDIR)/vdso.so
+
+$(SUBDIR)/vdso.so: vdso.S vdso.ld vdso-asmoffset.h
+	$(CC) -o $@ -nostdlib -shared -Wl,-h,linux-vdso64.so.1 \
+	  -Wl,--build-id=sha1 -Wl,--hash-style=both \
+	  -Wl,-T,$(SUBDIR)/vdso.ld $<

linux-user/s390x/meson.build

@@ -3,3 +3,9 @@ syscall_nr_generators += {
                      arguments: [ meson.current_source_dir() / 'syscallhdr.sh', '@INPUT@', '@OUTPUT@', '@EXTRA_ARGS@' ],
                      output: '@BASENAME@_nr.h')
 }
+
+vdso_inc = gen_vdso.process('vdso.so', extra_args: [
+                                '-s', '__kernel_sigreturn',
+                                '-r', '__kernel_rt_sigreturn'
+                            ])
+linux_user_ss.add(when: 'TARGET_S390X', if_true: vdso_inc)

linux-user/s390x/signal.c

@@ -21,13 +21,12 @@
 #include "user-internals.h"
 #include "signal-common.h"
 #include "linux-user/trace.h"
+#include "vdso-asmoffset.h"
 
 #define __NUM_GPRS 16
 #define __NUM_FPRS 16
 #define __NUM_ACRS 16
 
-#define __SIGNAL_FRAMESIZE      160 /* FIXME: 31-bit mode -> 96 */
-
 #define _SIGCONTEXT_NSIG        64
 #define _SIGCONTEXT_NSIG_BPW    64 /* FIXME: 31-bit mode -> 32 */
 #define _SIGCONTEXT_NSIG_WORDS  (_SIGCONTEXT_NSIG / _SIGCONTEXT_NSIG_BPW)
@@ -63,7 +62,7 @@ typedef struct {
 } target_sigcontext;
 
 typedef struct {
-    uint8_t callee_used_stack[__SIGNAL_FRAMESIZE];
+    uint8_t callee_used_stack[STACK_FRAME_OVERHEAD];
     target_sigcontext sc;
     target_sigregs sregs;
     int signo;
@@ -83,7 +82,7 @@ struct target_ucontext {
 };
 
 typedef struct {
-    uint8_t callee_used_stack[__SIGNAL_FRAMESIZE];
+    uint8_t callee_used_stack[STACK_FRAME_OVERHEAD];
     /*
      * This field is no longer initialized by the kernel, but it's still a part
      * of the ABI.

linux-user/s390x/vdso-asmoffset.h

@@ -0,0 +1,2 @@
+/* Minimum stack frame size */
+#define STACK_FRAME_OVERHEAD    160

linux-user/s390x/vdso.S

@@ -0,0 +1,61 @@
+/*
+ * s390x linux replacement vdso.
+ *
+ * Copyright 2023 Linaro, Ltd.
+ *
+ * SPDX-License-Identifier: GPL-2.0-or-later
+ */
+
+#include <asm/unistd.h>
+#include "vdso-asmoffset.h"
+
+.macro endf name
+	.globl	\name
+	.type	\name, @function
+	.size	\name, . - \name
+.endm
+
+.macro raw_syscall n
+        .ifne	\n < 0x100
+	svc	\n
+	.else
+	lghi	%r1, \n
+	svc	0
+        .endif
+.endm
+
+.macro vdso_syscall name, nr
+\name:
+	.cfi_startproc
+	aghi	%r15, -(STACK_FRAME_OVERHEAD + 16)
+	.cfi_adjust_cfa_offset STACK_FRAME_OVERHEAD + 16
+	stg	%r14, STACK_FRAME_OVERHEAD(%r15)
+	.cfi_rel_offset %r14, STACK_FRAME_OVERHEAD
+	raw_syscall \nr
+	lg	%r14, STACK_FRAME_OVERHEAD(%r15)
+	aghi	%r15, STACK_FRAME_OVERHEAD + 16
+	.cfi_restore %r14
+	.cfi_adjust_cfa_offset -(STACK_FRAME_OVERHEAD + 16)
+	br	%r14
+	.cfi_endproc
+endf	\name
+.endm
+
+vdso_syscall __kernel_gettimeofday, __NR_gettimeofday
+vdso_syscall __kernel_clock_gettime, __NR_clock_gettime
+vdso_syscall __kernel_clock_getres, __NR_clock_getres
+vdso_syscall __kernel_getcpu, __NR_getcpu
+
+/*
+ * TODO unwind info, though we're ok without it.
+ * The kernel supplies bogus empty unwind info, and it is likely ignored
+ * by all users.  Without it we get the fallback signal frame handling.
+ */
+
+__kernel_sigreturn:
+	raw_syscall __NR_sigreturn
+endf	__kernel_sigreturn
+
+__kernel_rt_sigreturn:
+	raw_syscall __NR_rt_sigreturn
+endf	__kernel_rt_sigreturn

linux-user/s390x/vdso.ld

@@ -0,0 +1,72 @@
+/*
+ * Linker script for linux s390x replacement vdso.
+ *
+ * Copyright 2023 Linaro, Ltd.
+ *
+ * SPDX-License-Identifier: GPL-2.0-or-later
+ */
+
+VERSION {
+        LINUX_2.6.29 {
+        global:
+                __kernel_gettimeofday;
+                __kernel_clock_gettime;
+                __kernel_clock_getres;
+                __kernel_getcpu;
+                __kernel_rt_sigreturn;
+                __kernel_sigreturn;
+                /*
+                 * QEMU handles syscall restart internally, so we don't
+                 * need the __kernel_restart_syscall entry point.
+                 */
+        local: *;
+        };
+}
+
+
+PHDRS {
+        phdr            PT_PHDR         FLAGS(4) PHDRS;
+        load            PT_LOAD         FLAGS(7) FILEHDR PHDRS; /* FLAGS=RWX */
+        dynamic         PT_DYNAMIC      FLAGS(4);
+        eh_frame_hdr    PT_GNU_EH_FRAME;
+        note            PT_NOTE         FLAGS(4);
+}
+
+SECTIONS {
+        . = SIZEOF_HEADERS;
+
+        /*
+         * The following, including the FILEHDRS and PHDRS, are modified
+         * when we relocate the binary.  We want them to be initially
+         * writable for the relocation; we'll force them read-only after.
+         */
+        .note           : { *(.note*) }         :load :note
+        .dynamic        : { *(.dynamic) }       :load :dynamic
+        .dynsym         : { *(.dynsym) }        :load
+        /*
+         * There ought not be any real read-write data.
+         * But since we manipulated the segment layout,
+         * we have to put these sections somewhere.
+         */
+        .data           : {
+                *(.data*)
+                *(.sdata*)
+                *(.got.plt) *(.got)
+                *(.gnu.linkonce.d.*)
+                *(.bss*)
+                *(.dynbss*)
+                *(.gnu.linkonce.b.*)
+        }
+
+        .rodata         : { *(.rodata*) }
+        .hash           : { *(.hash) }
+        .gnu.hash       : { *(.gnu.hash) }
+        .dynstr         : { *(.dynstr) }
+        .gnu.version    : { *(.gnu.version) }
+        .gnu.version_d  : { *(.gnu.version_d) }
+        .gnu.version_r  : { *(.gnu.version_r) }
+        .eh_frame_hdr   : { *(.eh_frame_hdr) }  :load :eh_frame_hdr
+        .eh_frame       : { *(.eh_frame) }      :load
+
+        .text           : { *(.text*) }         :load
+}

linux-user/signal.c

@@ -536,11 +536,10 @@ static void signal_table_init(void)
     host_to_target_signal_table[SIGABRT] = 0;
     host_to_target_signal_table[hsig++] = TARGET_SIGABRT;
 
-    for (; hsig <= SIGRTMAX; hsig++) {
+    for (tsig = TARGET_SIGRTMIN;
-        tsig = hsig - SIGRTMIN + TARGET_SIGRTMIN;
+         hsig <= SIGRTMAX && tsig <= TARGET_NSIG;
-        if (tsig <= TARGET_NSIG) {
+         hsig++, tsig++) {
-            host_to_target_signal_table[hsig] = tsig;
+        host_to_target_signal_table[hsig] = tsig;
-        }
     }
 
     /* Invert the mapping that has already been assigned. */

linux-user/syscall.c

@@ -7992,6 +7992,8 @@ static void open_self_maps_4(const struct open_self_maps_data *d,
         path = "[stack]";
     } else if (start == info->brk) {
         path = "[heap]";
+    } else if (start == info->vdso) {
+        path = "[vdso]";
     }
 
     /* Except null device (MAP_ANON), adjust offset for this fragment. */

linux-user/x86_64/Makefile.vdso

@@ -0,0 +1,11 @@
+include $(BUILD_DIR)/tests/tcg/x86_64-linux-user/config-target.mak
+
+SUBDIR = $(SRC_PATH)/linux-user/x86_64
+VPATH += $(SUBDIR)
+
+all: $(SUBDIR)/vdso.so
+
+$(SUBDIR)/vdso.so: vdso.S vdso.ld
+	$(CC) -o $@ -nostdlib -shared -Wl,-h,linux-vdso.so.1 \
+	  -Wl,--build-id=sha1 -Wl,--hash-style=both \
+	  -Wl,-T,$(SUBDIR)/vdso.ld $<

linux-user/x86_64/meson.build

@@ -3,3 +3,7 @@ syscall_nr_generators += {
                       arguments: [ meson.current_source_dir() / 'syscallhdr.sh', '@INPUT@', '@OUTPUT@', '@EXTRA_ARGS@' ],
                       output: '@BASENAME@_nr.h')
 }
+
+vdso_inc = gen_vdso.process('vdso.so')
+
+linux_user_ss.add(when: 'TARGET_X86_64', if_true: vdso_inc)

linux-user/x86_64/vdso.S

@@ -0,0 +1,78 @@
+/*
+ * x86-64 linux replacement vdso.
+ *
+ * Copyright 2023 Linaro, Ltd.
+ *
+ * SPDX-License-Identifier: GPL-2.0-or-later
+ */
+
+#include <asm/unistd.h>
+
+.macro endf name
+	.globl	\name
+	.type	\name, @function
+	.size	\name, . - \name
+.endm
+
+.macro weakalias name
+\name	= __vdso_\name
+	.weak	\name
+.endm
+
+.macro vdso_syscall name, nr
+__vdso_\name:
+	mov	$\nr, %eax
+	syscall
+	ret
+endf	__vdso_\name
+weakalias \name
+.endm
+
+	.cfi_startproc
+
+vdso_syscall clock_gettime, __NR_clock_gettime
+vdso_syscall clock_getres, __NR_clock_getres
+vdso_syscall gettimeofday, __NR_gettimeofday
+vdso_syscall time, __NR_time
+
+__vdso_getcpu:
+	/*
+         * There is no syscall number for this allocated on x64.
+	 * We can handle this several ways:
+         *
+	 * (1) Invent a syscall number for use within qemu.
+         *     It should be easy enough to pick a number that
+         *     is well out of the way of the kernel numbers.
+         *
+         * (2) Force the emulated cpu to support the rdtscp insn,
+	 *     and initialize the TSC_AUX value the appropriate value.
+         *
+	 * (3) Pretend that we're always running on cpu 0.
+         *
+	 * This last is the one that's implemented here, with the
+	 * tiny bit of extra code to support rdtscp in place.
+         */
+	xor	%ecx, %ecx		/* rdtscp w/ tsc_aux = 0 */
+
+	/* if (cpu != NULL) *cpu = (ecx & 0xfff); */
+	test	%rdi, %rdi
+	jz	1f
+	mov	%ecx, %eax
+	and	$0xfff, %eax
+	mov	%eax, (%rdi)
+
+	/* if (node != NULL) *node = (ecx >> 12); */
+1:	test	%rsi, %rsi
+	jz	2f
+	shr	$12, %ecx
+	mov	%ecx, (%rsi)
+
+2:	xor	%eax, %eax
+	ret
+endf	__vdso_getcpu
+
+weakalias getcpu
+
+	.cfi_endproc
+
+/* TODO: Add elf note for LINUX_VERSION_CODE */

linux-user/x86_64/vdso.ld

@@ -0,0 +1,73 @@
+/*
+ * Linker script for linux x86-64 replacement vdso.
+ *
+ * Copyright 2023 Linaro, Ltd.
+ *
+ * SPDX-License-Identifier: GPL-2.0-or-later
+ */
+
+VERSION {
+        LINUX_2.6 {
+        global:
+                clock_gettime;
+                __vdso_clock_gettime;
+                gettimeofday;
+                __vdso_gettimeofday;
+                getcpu;
+                __vdso_getcpu;
+                time;
+                __vdso_time;
+                clock_getres;
+                __vdso_clock_getres;
+
+        local: *;
+        };
+}
+
+
+PHDRS {
+        phdr            PT_PHDR         FLAGS(4) PHDRS;
+        load            PT_LOAD         FLAGS(7) FILEHDR PHDRS; /* FLAGS=RWX */
+        dynamic         PT_DYNAMIC      FLAGS(4);
+        eh_frame_hdr    PT_GNU_EH_FRAME;
+        note            PT_NOTE         FLAGS(4);
+}
+
+SECTIONS {
+        . = SIZEOF_HEADERS;
+
+        /*
+         * The following, including the FILEHDRS and PHDRS, are modified
+         * when we relocate the binary.  We want them to be initially
+         * writable for the relocation; we'll force them read-only after.
+         */
+        .note           : { *(.note*) }         :load :note
+        .dynamic        : { *(.dynamic) }       :load :dynamic
+        .dynsym         : { *(.dynsym) }        :load
+        .data           : {
+                /*
+                 * There ought not be any real read-write data.
+                 * But since we manipulated the segment layout,
+                 * we have to put these sections somewhere.
+                 */
+                *(.data*)
+                *(.sdata*)
+                *(.got.plt) *(.got)
+                *(.gnu.linkonce.d.*)
+                *(.bss*)
+                *(.dynbss*)
+                *(.gnu.linkonce.b.*)
+        }
+
+        .rodata         : { *(.rodata*) }
+        .hash           : { *(.hash) }
+        .gnu.hash       : { *(.gnu.hash) }
+        .dynstr         : { *(.dynstr) }
+        .gnu.version    : { *(.gnu.version) }
+        .gnu.version_d  : { *(.gnu.version_d) }
+        .gnu.version_r  : { *(.gnu.version_r) }
+        .eh_frame_hdr   : { *(.eh_frame_hdr) }  :load :eh_frame_hdr
+        .eh_frame       : { *(.eh_frame) }      :load
+
+        .text           : { *(.text*) }         :load   =0x90909090
+}