hibernate/resume regression

hibernate/resume regression

[CAI Qian]

The latest Linus tree had a regression that prevent hibernate/resume from working.

It is now failing to resume every time tried after hibernate.

PM: Starting manual resume from disk
Freezing user space processes ... 
EXT4-fs (dm-0): INFO: recovery required on readonly filesystem
EXT4-fs (dm-0): write access will be enabled during recovery
EXT4-fs (dm-0): recovery complete
EXT4-fs (dm-0): mounted filesystem with ordered data mode. Opts: (null)
(elapsed 0.18 seconds) done.
Freezing remaining freezable tasks ... (elapsed 0.01 seconds) done.
PM: Loading and decompressing image data (301765 pages) ... done
PM: Read 1207060 kbytes in 6.47 seconds (186.56 MB/s)
serial 00:08: disabled

I have not had a chance to track down the offensive patch(es) yet.

CAI Qian

Re: hibernate/resume regression

[CAI Qian]

OK, reverted the following patch fixed the problem.

CAI Qian

commit e5f15b45ddf3afa2bbbb10c7ea34fb32b6de0a0e
Author: Yinghai Lu <yinghai@kernel.org>
Date:   Fri Feb 18 11:30:30 2011 +0000

    x86: Cleanup highmap after brk is concluded
    
    Now cleanup_highmap actually is in two steps: one is early in head64.c
    and only clears above _end; a second one is in init_memory_mapping() and
    tries to clean from _brk_end to _end.
    It should check if those boundaries are PMD_SIZE aligned but currently
    does not.
    Also init_memory_mapping() is called several times for numa or memory
    hotplug, so we really should not handle initial kernel mappings there.
    
    This patch moves cleanup_highmap() down after _brk_end is settled so
    we can do everything in one step.
    Also we honor max_pfn_mapped in the implementation of cleanup_highmap.
    
    Signed-off-by: Yinghai Lu <yinghai@kernel.org>
    Signed-off-by: Stefano Stabellini <stefano.stabellini@eu.citrix.com>
    LKML-Reference: <alpine.DEB.2.00.1103171739050.3382@kaball-desktop>
    Signed-off-by: H. Peter Anvin <hpa@zytor.com>

diff --git a/arch/x86/kernel/head64.c b/arch/x86/kernel/head64.c
index 2d2673c..5655c22 100644
--- a/arch/x86/kernel/head64.c
+++ b/arch/x86/kernel/head64.c
@@ -77,9 +77,6 @@ void __init x86_64_start_kernel(char * real_mode_data)
        /* Make NULL pointers segfault */
        zap_identity_mappings();
 
-       /* Cleanup the over mapped high alias */
-       cleanup_highmap();
-
        max_pfn_mapped = KERNEL_IMAGE_SIZE >> PAGE_SHIFT;
 
        for (i = 0; i < NUM_EXCEPTION_VECTORS; i++) {
diff --git a/arch/x86/kernel/setup.c b/arch/x86/kernel/setup.c
index b176f2b..4a52a5f 100644
--- a/arch/x86/kernel/setup.c
+++ b/arch/x86/kernel/setup.c
@@ -294,30 +294,11 @@ static void __init init_gbpages(void)
        else
                direct_gbpages = 0;
 }
-
-static void __init cleanup_highmap_brk_end(void)
-{
-       pud_t *pud;
-       pmd_t *pmd;
-
-       mmu_cr4_features = read_cr4();
-
-       /*
-        * _brk_end cannot change anymore, but it and _end may be
-        * located on different 2M pages. cleanup_highmap(), however,
-        * can only consider _end when it runs, so destroy any
-        * mappings beyond _brk_end here.
-        */
-       pud = pud_offset(pgd_offset_k(_brk_end), _brk_end);
-       pmd = pmd_offset(pud, _brk_end - 1);
-       while (++pmd <= pmd_offset(pud, (unsigned long)_end - 1))
-               pmd_clear(pmd);
-}
 #else
 static inline void init_gbpages(void)
 {
 }
-static inline void cleanup_highmap_brk_end(void)
+static void __init cleanup_highmap(void)
 {
 }
 #endif
@@ -330,8 +311,6 @@ static void __init reserve_brk(void)
        /* Mark brk area as locked down and no longer taking any
           new allocations */
        _brk_start = 0;
-
-       cleanup_highmap_brk_end();
 }
 
 #ifdef CONFIG_BLK_DEV_INITRD
@@ -950,6 +929,8 @@ void __init setup_arch(char **cmdline_p)
         */
        reserve_brk();
 
+       cleanup_highmap();
+
        memblock.current_limit = get_max_mapped();
        memblock_x86_fill();
 
diff --git a/arch/x86/mm/init_64.c b/arch/x86/mm/init_64.c
index a08a62c..7026505 100644
--- a/arch/x86/mm/init_64.c
+++ b/arch/x86/mm/init_64.c
@@ -51,6 +51,7 @@
 #include <asm/numa.h>
 #include <asm/cacheflush.h>
 #include <asm/init.h>
+#include <asm/setup.h>
 
 static int __init parse_direct_gbpages_off(char *arg)
 {
@@ -293,18 +294,18 @@ void __init init_extra_mapping_uc(unsigned long phys, unsigned long size)
  * to the compile time generated pmds. This results in invalid pmds up
  * to the point where we hit the physaddr 0 mapping.
  *
- * We limit the mappings to the region from _text to _end.  _end is
- * rounded up to the 2MB boundary. This catches the invalid pmds as
+ * We limit the mappings to the region from _text to _brk_end.  _brk_end
+ * is rounded up to the 2MB boundary. This catches the invalid pmds as
  * well, as they are located before _text:
  */
 void __init cleanup_highmap(void)
 {
        unsigned long vaddr = __START_KERNEL_map;
-       unsigned long end = roundup((unsigned long)_end, PMD_SIZE) - 1;
+       unsigned long vaddr_end = __START_KERNEL_map + (max_pfn_mapped << PAGE_SHIFT);
+       unsigned long end = roundup((unsigned long)_brk_end, PMD_SIZE) - 1;
        pmd_t *pmd = level2_kernel_pgt;
-       pmd_t *last_pmd = pmd + PTRS_PER_PMD;
 
-       for (; pmd < last_pmd; pmd++, vaddr += PMD_SIZE) {
+       for (; vaddr + PMD_SIZE - 1 < vaddr_end; pmd++, vaddr += PMD_SIZE) {
                if (pmd_none(*pmd))
                        continue;
                if (vaddr < (unsigned long) _text || vaddr > end)

----- Original Message -----
> The latest Linus tree had a regression that prevent hibernate/resume
> from working.
> 
> It is now failing to resume every time tried after hibernate.
> 
> PM: Starting manual resume from disk
> Freezing user space processes ...
> EXT4-fs (dm-0): INFO: recovery required on readonly filesystem
> EXT4-fs (dm-0): write access will be enabled during recovery
> EXT4-fs (dm-0): recovery complete
> EXT4-fs (dm-0): mounted filesystem with ordered data mode. Opts:
> (null)
> (elapsed 0.18 seconds) done.
> Freezing remaining freezable tasks ... (elapsed 0.01 seconds) done.
> PM: Loading and decompressing image data (301765 pages) ... done
> PM: Read 1207060 kbytes in 6.47 seconds (186.56 MB/s)
> serial 00:08: disabled
> 
> I have not had a chance to track down the offensive patch(es) yet.
> 
> CAI Qian

Re: hibernate/resume regression

[CAI Qian]

----- Original Message -----
> Yes we know; does the fix currently in mainstream work for you?
Yes, it worked for me.

CAI Qian

> Sent from my mobile phone. Please pardon any lack of formatting.
> 
> 
> CAI Qian <caiqian@redhat.com> wrote:
> 
> 
> OK, reverted the following patch fixed the problem. CAI Qian commit
> e5f15b45ddf3afa2bbbb10c7ea34fb32b6de0a0e Author: Yinghai Lu
> <yinghai@kernel.org> Date: Fri Feb 18 11:30:30 2011 +0000 x86: Cleanup
> highmap after brk is concluded Now cleanup_highmap actually is in two
> steps: one is early in head64.c and only clears above _end; a second
> one is in init_memory_mapping() and tries to clean from _brk_end to
> _end. It should check if those boundaries are PMD_SIZE aligned but
> currently does not. Also init_memory_mapping() is called several times
> for numa or memory hotplug, so we really should not handle initial
> kernel mappings there. This patch moves cleanup_highmap() down after
> _brk_end is settled so we can do everything in one step. Also we honor
> max_pfn_mapped in the implementation of cleanup_highmap.
> Signed-off-by: Yinghai Lu <yinghai@kernel.org> Signed-off-by: Stefano
> Stabellini <stefano.stabellini@eu.citrix.com> LKML-Reference:
> <alpine.DEB.2.00.1103171739050.3382@kaball-desktop> Signed-off-by: H.
> Peter Anvin <hpa@zytor.com> diff --git a/arch/x86/kernel/head64.c
> b/arch/x86/kernel/head64.c index 2d2673c..5655c22 100644 ---
> a/arch/x86/kernel/head64.c +++ b/arch/x86/kernel/head64.c @@ -77,9
> +77,6 @@ void __init x86_64_start_kernel(char * real_mode_data) /*
> Make NULL pointers segfault */ zap_identity_mappings(); - /* Cleanup
> the over mapped high alias */ - cleanup_highmap(); - max_pfn_mapped =
> KERNEL_IMAGE_SIZE >> PAGE_SHIFT; for (i = 0; i <
> NUM_EXCEPTION_VECTORS; i++) { diff --git a/arch/x86/kernel/setup.c
> b/arch/x86/kernel/setup.c index b176f2b..4a52a5f 100644 ---
> a/arch/x86/kernel/setup.c +++ b/arch/x86/kernel/setup.c @@ -294,30
> +294,11 @@ static void __init init_gbpages(void) else direct_gbpages =
> 0; } - -static void __init cleanup_highmap_brk_end(void) -{ - pud_t
> *pud; - pmd_t *pmd; - - mmu_cr4_features = read_cr4(); - - /* - *
> _brk_end cannot change anymore, but it and _end may be - * located on
> different 2M pages. cleanup_highmap(), however, - * can only consider
> _end when it runs, so destroy any - * mappings beyond _brk_end here. -
> */ - pud = pud_offset(pgd_offset_k(_brk_end), _brk_end); - pmd =
> pmd_offset(pud, _brk_end - 1); - while (++pmd <= pmd_offset(pud,
> (unsigned long)_end - 1)) - pmd_clear(pmd); -} #else static inline
> void init_gbpages(void) { } -static inline void
> cleanup_highmap_brk_end(void) +static void __init
> cleanup_highmap(void) { } #endif @@ -330,8 +311,6 @@ static void
> __init reserve_brk(void) /* Mark brk area as locked down and no longer
> taking any new allocations */ _brk_start = 0; - -
> cleanup_highmap_brk_end(); } #ifdef CONFIG_BLK_DEV_INITRD @@ -950,6
> +929,8 @@ void __init setup_arch(char **cmdline_p) */ reserve_brk(); +
> cleanup_highmap(); + memblock.current_limit = get_max_mapped();
> memblock_x86_fill(); diff --git a/arch/x86/mm/init_64.c
> b/arch/x86/mm/init_64.c index a08a62c..7026505 100644 ---
> a/arch/x86/mm/init_64.c +++ b/arch/x86/mm/init_64.c @@ -51,6 +51,7 @@
> #include <asm/numa.h> #include <asm/cacheflush.h> #include
> <asm/init.h> +#include <asm/setup.h> static int __init
> parse_direct_gbpages_off(char *arg) { @@ -293,18 +294,18 @@ void
> __init init_extra_mapping_uc(unsigned long phys, unsigned long size) *
> to the compile time generated pmds. This results in invalid pmds up *
> to the point where we hit the physaddr 0 mapping. * - * We limit the
> mappings to the region from _text to _end. _end is - * rounded up to
> the 2MB boundary. This catches the invalid pmds as + * We limit the
> mappings to the region from _text to _brk_end. _brk_end + * is rounded
> up to the 2MB boundary. This catches the invalid pmds as * well, as
> they are located before _text: */ void __init cleanup_highmap(void) {
> unsigned long vaddr = __START_KERNEL_map; - unsigned long end =
> roundup((unsigned long)_end, PMD_SIZE) - 1; + unsigned long vaddr_end
> = __START_KERNEL_map + (max_pfn_mapped << PAGE_SHIFT); + unsigned long
> end = roundup((unsigned long)_brk_end, PMD_SIZE) - 1; pmd_t *pmd =
> level2_kernel_pgt; - pmd_t *last_pmd = pmd + PTRS_PER_PMD; - for (;
> pmd < last_pmd; pmd++, vaddr += PMD_SIZE) { + for (; vaddr + PMD_SIZE
> - 1 < vaddr_end; pmd++, vaddr += PMD_SIZE) { if (pmd_none(*pmd))
> continue; if (vaddr < (unsigned long) _text || vaddr > end) -----
> Original Message ----- > The latest Linus tree had a regression that
> prevent hibernate/resume > from working. > > It is now failing to
> resume every time tried after hibernate. > > PM: Starting manual
> resume from disk > Freezing user space processes ... > EXT4-fs (dm-0):
> INFO: recovery required on readonly filesystem > EXT4-fs (dm-0): write
> access will be enabled during recovery > EXT4-fs (dm-0): recovery
> complete > EXT4-fs (dm-0): mounted filesystem with ordered data mode.
> Opts: > (null) > (elapsed 0.18 seconds) done. > Freezing remaining
> freezable tasks ... (elapsed 0.01 seconds) done. > PM: Loading and
> decompressing image data (301765 pages) ... done > PM: Read 1207060
> kbytes in 6.47 seconds (186.56 MB/s) > serial 00:08: disabled > > I
> have not had a chance to track down the offensive patch(es) yet. > >
> CAI Qian