]> git.kernelconcepts.de Git - karo-tx-linux.git/commitdiff
metag: Module support
authorJames Hogan <james.hogan@imgtec.com>
Tue, 9 Oct 2012 09:55:00 +0000 (10:55 +0100)
committerJames Hogan <james.hogan@imgtec.com>
Sat, 2 Mar 2013 20:09:49 +0000 (20:09 +0000)
Signed-off-by: James Hogan <james.hogan@imgtec.com>
arch/metag/include/asm/module.h [new file with mode: 0644]
arch/metag/kernel/module.c [new file with mode: 0644]

diff --git a/arch/metag/include/asm/module.h b/arch/metag/include/asm/module.h
new file mode 100644 (file)
index 0000000..e47e609
--- /dev/null
@@ -0,0 +1,37 @@
+#ifndef _ASM_METAG_MODULE_H
+#define _ASM_METAG_MODULE_H
+
+#include <asm-generic/module.h>
+
+struct metag_plt_entry {
+       /* Indirect jump instruction sequence. */
+       unsigned long tramp[2];
+};
+
+struct mod_arch_specific {
+       /* Indices of PLT sections within module. */
+       unsigned int core_plt_section, init_plt_section;
+};
+
+#if defined CONFIG_METAG_META12
+#define MODULE_PROC_FAMILY "META 1.2 "
+#elif defined CONFIG_METAG_META21
+#define MODULE_PROC_FAMILY "META 2.1 "
+#else
+#define MODULE_PROC_FAMILY ""
+#endif
+
+#ifdef CONFIG_4KSTACKS
+#define MODULE_STACKSIZE "4KSTACKS "
+#else
+#define MODULE_STACKSIZE ""
+#endif
+
+#define MODULE_ARCH_VERMAGIC MODULE_PROC_FAMILY MODULE_STACKSIZE
+
+#ifdef MODULE
+asm(".section .plt,\"ax\",@progbits; .balign 8; .previous");
+asm(".section .init.plt,\"ax\",@progbits; .balign 8; .previous");
+#endif
+
+#endif /* _ASM_METAG_MODULE_H */
diff --git a/arch/metag/kernel/module.c b/arch/metag/kernel/module.c
new file mode 100644 (file)
index 0000000..986331c
--- /dev/null
@@ -0,0 +1,284 @@
+/*  Kernel module help for Meta.
+
+    This program is free software; you can redistribute it and/or modify
+    it under the terms of the GNU General Public License as published by
+    the Free Software Foundation; either version 2 of the License, or
+    (at your option) any later version.
+
+    This program 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 General Public License for more details.
+*/
+#include <linux/moduleloader.h>
+#include <linux/elf.h>
+#include <linux/vmalloc.h>
+#include <linux/fs.h>
+#include <linux/string.h>
+#include <linux/kernel.h>
+#include <linux/sort.h>
+
+#include <asm/unaligned.h>
+
+/* Count how many different relocations (different symbol, different
+   addend) */
+static unsigned int count_relocs(const Elf32_Rela *rela, unsigned int num)
+{
+       unsigned int i, r_info, r_addend, _count_relocs;
+
+       _count_relocs = 0;
+       r_info = 0;
+       r_addend = 0;
+       for (i = 0; i < num; i++)
+               /* Only count relbranch relocs, others don't need stubs */
+               if (ELF32_R_TYPE(rela[i].r_info) == R_METAG_RELBRANCH &&
+                   (r_info != ELF32_R_SYM(rela[i].r_info) ||
+                    r_addend != rela[i].r_addend)) {
+                       _count_relocs++;
+                       r_info = ELF32_R_SYM(rela[i].r_info);
+                       r_addend = rela[i].r_addend;
+               }
+
+       return _count_relocs;
+}
+
+static int relacmp(const void *_x, const void *_y)
+{
+       const Elf32_Rela *x, *y;
+
+       y = (Elf32_Rela *)_x;
+       x = (Elf32_Rela *)_y;
+
+       /* Compare the entire r_info (as opposed to ELF32_R_SYM(r_info) only) to
+        * make the comparison cheaper/faster. It won't affect the sorting or
+        * the counting algorithms' performance
+        */
+       if (x->r_info < y->r_info)
+               return -1;
+       else if (x->r_info > y->r_info)
+               return 1;
+       else if (x->r_addend < y->r_addend)
+               return -1;
+       else if (x->r_addend > y->r_addend)
+               return 1;
+       else
+               return 0;
+}
+
+static void relaswap(void *_x, void *_y, int size)
+{
+       uint32_t *x, *y, tmp;
+       int i;
+
+       y = (uint32_t *)_x;
+       x = (uint32_t *)_y;
+
+       for (i = 0; i < sizeof(Elf32_Rela) / sizeof(uint32_t); i++) {
+               tmp = x[i];
+               x[i] = y[i];
+               y[i] = tmp;
+       }
+}
+
+/* Get the potential trampolines size required of the init and
+   non-init sections */
+static unsigned long get_plt_size(const Elf32_Ehdr *hdr,
+                                 const Elf32_Shdr *sechdrs,
+                                 const char *secstrings,
+                                 int is_init)
+{
+       unsigned long ret = 0;
+       unsigned i;
+
+       /* Everything marked ALLOC (this includes the exported
+          symbols) */
+       for (i = 1; i < hdr->e_shnum; i++) {
+               /* If it's called *.init*, and we're not init, we're
+                  not interested */
+               if ((strstr(secstrings + sechdrs[i].sh_name, ".init") != NULL)
+                   != is_init)
+                       continue;
+
+               /* We don't want to look at debug sections. */
+               if (strstr(secstrings + sechdrs[i].sh_name, ".debug") != NULL)
+                       continue;
+
+               if (sechdrs[i].sh_type == SHT_RELA) {
+                       pr_debug("Found relocations in section %u\n", i);
+                       pr_debug("Ptr: %p.  Number: %u\n",
+                                (void *)hdr + sechdrs[i].sh_offset,
+                                sechdrs[i].sh_size / sizeof(Elf32_Rela));
+
+                       /* Sort the relocation information based on a symbol and
+                        * addend key. This is a stable O(n*log n) complexity
+                        * alogrithm but it will reduce the complexity of
+                        * count_relocs() to linear complexity O(n)
+                        */
+                       sort((void *)hdr + sechdrs[i].sh_offset,
+                            sechdrs[i].sh_size / sizeof(Elf32_Rela),
+                            sizeof(Elf32_Rela), relacmp, relaswap);
+
+                       ret += count_relocs((void *)hdr
+                                            + sechdrs[i].sh_offset,
+                                            sechdrs[i].sh_size
+                                            / sizeof(Elf32_Rela))
+                               * sizeof(struct metag_plt_entry);
+               }
+       }
+
+       return ret;
+}
+
+int module_frob_arch_sections(Elf32_Ehdr *hdr,
+                             Elf32_Shdr *sechdrs,
+                             char *secstrings,
+                             struct module *me)
+{
+       unsigned int i;
+
+       /* Find .plt and .init.plt sections */
+       for (i = 0; i < hdr->e_shnum; i++) {
+               if (strcmp(secstrings + sechdrs[i].sh_name, ".init.plt") == 0)
+                       me->arch.init_plt_section = i;
+               else if (strcmp(secstrings + sechdrs[i].sh_name, ".plt") == 0)
+                       me->arch.core_plt_section = i;
+       }
+       if (!me->arch.core_plt_section || !me->arch.init_plt_section) {
+               pr_err("Module doesn't contain .plt or .init.plt sections.\n");
+               return -ENOEXEC;
+       }
+
+       /* Override their sizes */
+       sechdrs[me->arch.core_plt_section].sh_size
+               = get_plt_size(hdr, sechdrs, secstrings, 0);
+       sechdrs[me->arch.core_plt_section].sh_type = SHT_NOBITS;
+       sechdrs[me->arch.init_plt_section].sh_size
+               = get_plt_size(hdr, sechdrs, secstrings, 1);
+       sechdrs[me->arch.init_plt_section].sh_type = SHT_NOBITS;
+       return 0;
+}
+
+/* Set up a trampoline in the PLT to bounce us to the distant function */
+static uint32_t do_plt_call(void *location, Elf32_Addr val,
+                           Elf32_Shdr *sechdrs, struct module *mod)
+{
+       struct metag_plt_entry *entry;
+       /* Instructions used to do the indirect jump.  */
+       uint32_t tramp[2];
+
+       /* We have to trash a register, so we assume that any control
+          transfer more than 21-bits away must be a function call
+          (so we can use a call-clobbered register).  */
+
+       /* MOVT D0Re0,#HI(v) */
+       tramp[0] = 0x02000005 | (((val & 0xffff0000) >> 16) << 3);
+       /* JUMP D0Re0,#LO(v) */
+       tramp[1] = 0xac000001 | ((val & 0x0000ffff) << 3);
+
+       /* Init, or core PLT? */
+       if (location >= mod->module_core
+           && location < mod->module_core + mod->core_size)
+               entry = (void *)sechdrs[mod->arch.core_plt_section].sh_addr;
+       else
+               entry = (void *)sechdrs[mod->arch.init_plt_section].sh_addr;
+
+       /* Find this entry, or if that fails, the next avail. entry */
+       while (entry->tramp[0])
+               if (entry->tramp[0] == tramp[0] && entry->tramp[1] == tramp[1])
+                       return (uint32_t)entry;
+               else
+                       entry++;
+
+       entry->tramp[0] = tramp[0];
+       entry->tramp[1] = tramp[1];
+
+       return (uint32_t)entry;
+}
+
+int apply_relocate_add(Elf32_Shdr *sechdrs,
+                  const char *strtab,
+                  unsigned int symindex,
+                  unsigned int relsec,
+                  struct module *me)
+{
+       unsigned int i;
+       Elf32_Rela *rel = (void *)sechdrs[relsec].sh_addr;
+       Elf32_Sym *sym;
+       Elf32_Addr relocation;
+       uint32_t *location;
+       int32_t value;
+
+       pr_debug("Applying relocate section %u to %u\n", relsec,
+                sechdrs[relsec].sh_info);
+       for (i = 0; i < sechdrs[relsec].sh_size / sizeof(*rel); i++) {
+               /* This is where to make the change */
+               location = (void *)sechdrs[sechdrs[relsec].sh_info].sh_addr
+                       + rel[i].r_offset;
+               /* This is the symbol it is referring to.  Note that all
+                  undefined symbols have been resolved.  */
+               sym = (Elf32_Sym *)sechdrs[symindex].sh_addr
+                       + ELF32_R_SYM(rel[i].r_info);
+               relocation = sym->st_value + rel[i].r_addend;
+
+               switch (ELF32_R_TYPE(rel[i].r_info)) {
+               case R_METAG_NONE:
+                       break;
+               case R_METAG_HIADDR16:
+                       relocation >>= 16;
+               case R_METAG_LOADDR16:
+                       *location = (*location & 0xfff80007) |
+                               ((relocation & 0xffff) << 3);
+                       break;
+               case R_METAG_ADDR32:
+                       /*
+                        * Packed data structures may cause a misaligned
+                        * R_METAG_ADDR32 to be emitted.
+                        */
+                       put_unaligned(relocation, location);
+                       break;
+               case R_METAG_GETSETOFF:
+                       *location += ((relocation & 0xfff) << 7);
+                       break;
+               case R_METAG_RELBRANCH:
+                       if (*location & (0x7ffff << 5)) {
+                               pr_err("bad relbranch relocation\n");
+                               break;
+                       }
+
+                       /* This jump is too big for the offset slot. Build
+                        * a PLT to jump through to get to where we want to go.
+                        * NB: 21bit check - not scaled to 19bit yet
+                        */
+                       if (((int32_t)(relocation -
+                                      (uint32_t)location) > 0xfffff) ||
+                           ((int32_t)(relocation -
+                                      (uint32_t)location) < -0xfffff)) {
+                               relocation = do_plt_call(location, relocation,
+                                                        sechdrs, me);
+                       }
+
+                       value = relocation - (uint32_t)location;
+
+                       /* branch instruction aligned */
+                       value /= 4;
+
+                       if ((value > 0x7ffff) || (value < -0x7ffff)) {
+                               /*
+                                * this should have been caught by the code
+                                * above!
+                                */
+                               pr_err("overflow of relbranch reloc\n");
+                       }
+
+                       *location = (*location & (~(0x7ffff << 5))) |
+                               ((value & 0x7ffff) << 5);
+                       break;
+
+               default:
+                       pr_err("module %s: Unknown relocation: %u\n",
+                              me->name, ELF32_R_TYPE(rel[i].r_info));
+                       return -ENOEXEC;
+               }
+       }
+       return 0;
+}