[PATCH] swsusp: low level interface

Introduce the low level interface that can be used for handling the
snapshot of the system memory by the in-kernel swap-writing/reading code of
swsusp and the userland interface code (to be introduced shortly).

Also change the way in which swsusp records the allocated swap pages and,
consequently, simplifies the in-kernel swap-writing/reading code (this is
necessary for the userland interface too).  To this end, it introduces two
helper functions in mm/swapfile.c, so that the swsusp code does not refer
directly to the swap internals.

Signed-off-by: Rafael J. Wysocki <rjw@sisk.pl>
Acked-by: Pavel Machek <pavel@ucw.cz>
Signed-off-by: Adrian Bunk <bunk@stusta.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

1 files changed, 321 insertions, 5 deletions

diff --git a/kernel/power/snapshot.c b/kernel/power/snapshot.c
index 8d5a5986d621..cc349437fb72 100644
--- a/kernel/power/snapshot.c
+++ b/kernel/power/snapshot.c
@@ -10,6 +10,7 @@
  */
 
 
+#include <linux/version.h>
 #include <linux/module.h>
 #include <linux/mm.h>
 #include <linux/suspend.h>
@@ -34,7 +35,8 @@
 #include "power.h"
 
 struct pbe *pagedir_nosave;
-unsigned int nr_copy_pages;
+static unsigned int nr_copy_pages;
+static unsigned int nr_meta_pages;
 
 #ifdef CONFIG_HIGHMEM
 unsigned int count_highmem_pages(void)
@@ -235,7 +237,7 @@ static void copy_data_pages(struct pbe *pblist)
  *      free_pagedir - free pages allocated with alloc_pagedir()
  */
 
-void free_pagedir(struct pbe *pblist)
+static void free_pagedir(struct pbe *pblist)
 {
         struct pbe *pbe;
 
@@ -301,7 +303,7 @@ struct eaten_page {
 
 static struct eaten_page *eaten_pages = NULL;
 
-void release_eaten_pages(void)
+static void release_eaten_pages(void)
 {
         struct eaten_page *p, *q;
 
@@ -376,7 +378,6 @@ struct pbe *alloc_pagedir(unsigned int nr_pages, gfp_t gfp_mask, int safe_needed
         if (!nr_pages)
                 return NULL;
 
-        pr_debug("alloc_pagedir(): nr_pages = %d\n", nr_pages);
         pblist = alloc_image_page(gfp_mask, safe_needed);
         /* FIXME: rewrite this ugly loop */
         for (pbe = pblist, num = PBES_PER_PAGE; pbe && num < nr_pages;
@@ -414,6 +415,9 @@ void swsusp_free(void)
                                 }
                         }
         }
+        nr_copy_pages = 0;
+        nr_meta_pages = 0;
+        pagedir_nosave = NULL;
 }
 
 
@@ -437,7 +441,7 @@ static int enough_free_mem(unsigned int nr_pages)
                 (nr_pages + PBES_PER_PAGE - 1) / PBES_PER_PAGE);
 }
 
-int alloc_data_pages(struct pbe *pblist, gfp_t gfp_mask, int safe_needed)
+static int alloc_data_pages(struct pbe *pblist, gfp_t gfp_mask, int safe_needed)
 {
         struct pbe *p;
 
@@ -504,7 +508,319 @@ asmlinkage int swsusp_save(void)
          */
 
         nr_copy_pages = nr_pages;
+        nr_meta_pages = (nr_pages * sizeof(long) + PAGE_SIZE - 1) >> PAGE_SHIFT;
 
         printk("swsusp: critical section/: done (%d pages copied)\n", nr_pages);
         return 0;
 }
+
+static void init_header(struct swsusp_info *info)
+{
+        memset(info, 0, sizeof(struct swsusp_info));
+        info->version_code = LINUX_VERSION_CODE;
+        info->num_physpages = num_physpages;
+        memcpy(&info->uts, &system_utsname, sizeof(system_utsname));
+        info->cpus = num_online_cpus();
+        info->image_pages = nr_copy_pages;
+        info->pages = nr_copy_pages + nr_meta_pages + 1;
+}
+
+/**
+ *      pack_orig_addresses - the .orig_address fields of the PBEs from the
+ *      list starting at @pbe are stored in the array @buf[] (1 page)
+ */
+
+static inline struct pbe *pack_orig_addresses(unsigned long *buf, struct pbe *pbe)
+{
+        int j;
+
+        for (j = 0; j < PAGE_SIZE / sizeof(long) && pbe; j++) {
+                buf[j] = pbe->orig_address;
+                pbe = pbe->next;
+        }
+        if (!pbe)
+                for (; j < PAGE_SIZE / sizeof(long); j++)
+                        buf[j] = 0;
+        return pbe;
+}
+
+/**
+ *      snapshot_read_next - used for reading the system memory snapshot.
+ *
+ *      On the first call to it @handle should point to a zeroed
+ *      snapshot_handle structure.  The structure gets updated and a pointer
+ *      to it should be passed to this function every next time.
+ *
+ *      The @count parameter should contain the number of bytes the caller
+ *      wants to read from the snapshot.  It must not be zero.
+ *
+ *      On success the function returns a positive number.  Then, the caller
+ *      is allowed to read up to the returned number of bytes from the memory
+ *      location computed by the data_of() macro.  The number returned
+ *      may be smaller than @count, but this only happens if the read would
+ *      cross a page boundary otherwise.
+ *
+ *      The function returns 0 to indicate the end of data stream condition,
+ *      and a negative number is returned on error.  In such cases the
+ *      structure pointed to by @handle is not updated and should not be used
+ *      any more.
+ */
+
+int snapshot_read_next(struct snapshot_handle *handle, size_t count)
+{
+        static unsigned long *buffer;
+
+        if (handle->page > nr_meta_pages + nr_copy_pages)
+                return 0;
+        if (!buffer) {
+                /* This makes the buffer be freed by swsusp_free() */
+                buffer = alloc_image_page(GFP_ATOMIC, 0);
+                if (!buffer)
+                        return -ENOMEM;
+        }
+        if (!handle->offset) {
+                init_header((struct swsusp_info *)buffer);
+                handle->buffer = buffer;
+                handle->pbe = pagedir_nosave;
+        }
+        if (handle->prev < handle->page) {
+                if (handle->page <= nr_meta_pages) {
+                        handle->pbe = pack_orig_addresses(buffer, handle->pbe);
+                        if (!handle->pbe)
+                                handle->pbe = pagedir_nosave;
+                } else {
+                        handle->buffer = (void *)handle->pbe->address;
+                        handle->pbe = handle->pbe->next;
+                }
+                handle->prev = handle->page;
+        }
+        handle->buf_offset = handle->page_offset;
+        if (handle->page_offset + count >= PAGE_SIZE) {
+                count = PAGE_SIZE - handle->page_offset;
+                handle->page_offset = 0;
+                handle->page++;
+        } else {
+                handle->page_offset += count;
+        }
+        handle->offset += count;
+        return count;
+}
+
+/**
+ *      mark_unsafe_pages - mark the pages that cannot be used for storing
+ *      the image during resume, because they conflict with the pages that
+ *      had been used before suspend
+ */
+
+static int mark_unsafe_pages(struct pbe *pblist)
+{
+        struct zone *zone;
+        unsigned long zone_pfn;
+        struct pbe *p;
+
+        if (!pblist) /* a sanity check */
+                return -EINVAL;
+
+        /* Clear page flags */
+        for_each_zone (zone) {
+                for (zone_pfn = 0; zone_pfn < zone->spanned_pages; ++zone_pfn)
+                        if (pfn_valid(zone_pfn + zone->zone_start_pfn))
+                                ClearPageNosaveFree(pfn_to_page(zone_pfn +
+                                        zone->zone_start_pfn));
+        }
+
+        /* Mark orig addresses */
+        for_each_pbe (p, pblist) {
+                if (virt_addr_valid(p->orig_address))
+                        SetPageNosaveFree(virt_to_page(p->orig_address));
+                else
+                        return -EFAULT;
+        }
+
+        return 0;
+}
+
+static void copy_page_backup_list(struct pbe *dst, struct pbe *src)
+{
+        /* We assume both lists contain the same number of elements */
+        while (src) {
+                dst->orig_address = src->orig_address;
+                dst = dst->next;
+                src = src->next;
+        }
+}
+
+static int check_header(struct swsusp_info *info)
+{
+        char *reason = NULL;
+
+        if (info->version_code != LINUX_VERSION_CODE)
+                reason = "kernel version";
+        if (info->num_physpages != num_physpages)
+                reason = "memory size";
+        if (strcmp(info->uts.sysname,system_utsname.sysname))
+                reason = "system type";
+        if (strcmp(info->uts.release,system_utsname.release))
+                reason = "kernel release";
+        if (strcmp(info->uts.version,system_utsname.version))
+                reason = "version";
+        if (strcmp(info->uts.machine,system_utsname.machine))
+                reason = "machine";
+        if (reason) {
+                printk(KERN_ERR "swsusp: Resume mismatch: %s\n", reason);
+                return -EPERM;
+        }
+        return 0;
+}
+
+/**
+ *      load header - check the image header and copy data from it
+ */
+
+static int load_header(struct snapshot_handle *handle,
+                              struct swsusp_info *info)
+{
+        int error;
+        struct pbe *pblist;
+
+        error = check_header(info);
+        if (!error) {
+                pblist = alloc_pagedir(info->image_pages, GFP_ATOMIC, 0);
+                if (!pblist)
+                        return -ENOMEM;
+                pagedir_nosave = pblist;
+                handle->pbe = pblist;
+                nr_copy_pages = info->image_pages;
+                nr_meta_pages = info->pages - info->image_pages - 1;
+        }
+        return error;
+}
+
+/**
+ *      unpack_orig_addresses - copy the elements of @buf[] (1 page) to
+ *      the PBEs in the list starting at @pbe
+ */
+
+static inline struct pbe *unpack_orig_addresses(unsigned long *buf,
+                                                struct pbe *pbe)
+{
+        int j;
+
+        for (j = 0; j < PAGE_SIZE / sizeof(long) && pbe; j++) {
+                pbe->orig_address = buf[j];
+                pbe = pbe->next;
+        }
+        return pbe;
+}
+
+/**
+ *      create_image - use metadata contained in the PBE list
+ *      pointed to by pagedir_nosave to mark the pages that will
+ *      be overwritten in the process of restoring the system
+ *      memory state from the image and allocate memory for
+ *      the image avoiding these pages
+ */
+
+static int create_image(struct snapshot_handle *handle)
+{
+        int error = 0;
+        struct pbe *p, *pblist;
+
+        p = pagedir_nosave;
+        error = mark_unsafe_pages(p);
+        if (!error) {
+                pblist = alloc_pagedir(nr_copy_pages, GFP_ATOMIC, 1);
+                if (pblist)
+                        copy_page_backup_list(pblist, p);
+                free_pagedir(p);
+                if (!pblist)
+                        error = -ENOMEM;
+        }
+        if (!error)
+                error = alloc_data_pages(pblist, GFP_ATOMIC, 1);
+        if (!error) {
+                release_eaten_pages();
+                pagedir_nosave = pblist;
+        } else {
+                pagedir_nosave = NULL;
+                handle->pbe = NULL;
+                nr_copy_pages = 0;
+                nr_meta_pages = 0;
+        }
+        return error;
+}
+
+/**
+ *      snapshot_write_next - used for writing the system memory snapshot.
+ *
+ *      On the first call to it @handle should point to a zeroed
+ *      snapshot_handle structure.  The structure gets updated and a pointer
+ *      to it should be passed to this function every next time.
+ *
+ *      The @count parameter should contain the number of bytes the caller
+ *      wants to write to the image.  It must not be zero.
+ *
+ *      On success the function returns a positive number.  Then, the caller
+ *      is allowed to write up to the returned number of bytes to the memory
+ *      location computed by the data_of() macro.  The number returned
+ *      may be smaller than @count, but this only happens if the write would
+ *      cross a page boundary otherwise.
+ *
+ *      The function returns 0 to indicate the "end of file" condition,
+ *      and a negative number is returned on error.  In such cases the
+ *      structure pointed to by @handle is not updated and should not be used
+ *      any more.
+ */
+
+int snapshot_write_next(struct snapshot_handle *handle, size_t count)
+{
+        static unsigned long *buffer;
+        int error = 0;
+
+        if (handle->prev && handle->page > nr_meta_pages + nr_copy_pages)
+                return 0;
+        if (!buffer) {
+                /* This makes the buffer be freed by swsusp_free() */
+                buffer = alloc_image_page(GFP_ATOMIC, 0);
+                if (!buffer)
+                        return -ENOMEM;
+        }
+        if (!handle->offset)
+                handle->buffer = buffer;
+        if (handle->prev < handle->page) {
+                if (!handle->prev) {
+                        error = load_header(handle, (struct swsusp_info *)buffer);
+                        if (error)
+                                return error;
+                } else if (handle->prev <= nr_meta_pages) {
+                        handle->pbe = unpack_orig_addresses(buffer, handle->pbe);
+                        if (!handle->pbe) {
+                                error = create_image(handle);
+                                if (error)
+                                        return error;
+                                handle->pbe = pagedir_nosave;
+                                handle->buffer = (void *)handle->pbe->address;
+                        }
+                } else {
+                        handle->pbe = handle->pbe->next;
+                        handle->buffer = (void *)handle->pbe->address;
+                }
+                handle->prev = handle->page;
+        }
+        handle->buf_offset = handle->page_offset;
+        if (handle->page_offset + count >= PAGE_SIZE) {
+                count = PAGE_SIZE - handle->page_offset;
+                handle->page_offset = 0;
+                handle->page++;
+        } else {
+                handle->page_offset += count;
+        }
+        handle->offset += count;
+        return count;
+}
+
+int snapshot_image_loaded(struct snapshot_handle *handle)
+{
+        return !(!handle->pbe || handle->pbe->next || !nr_copy_pages ||
+                handle->page <= nr_meta_pages + nr_copy_pages);
+}