Skip to contentSkip to conversation

makedumpfile: elf_info: check for invalid physical address when finding max_paddr

PATCH In review Coiby Xu 2 participants 3 messages Jun 15, 2021 – Jun 21, 2021

Subsystems (2)
lore
Share & tools
lore

Coiby Xu

author
#0 lore raw

PATCH 1/1 makedumpfile: elf_info: check for invalid physical address when finding max_paddr

Kernel commit 464920104bf7adac12722035bfefb3d772eb04d8 "/proc/kcore: update physical address for kcore ram and text" sets an invalid paddr (-1=0xffffffffffffffff) for PT_LOAD segments of not direct mapped regions,

    $ readelf -l /proc/kcore
    Elf file type is CORE (Core file)
    Entry point 0x0
    There are 4 program headers, starting at offset 64

    Program Headers:
      Type           Offset             VirtAddr           PhysAddr
                     FileSiz            MemSiz              Flags  Align
      NOTE           0x0000000000000120 0x0000000000000000 0x0000000000000000
                     0x0000000000002320 0x0000000000000000         0x0
      LOAD           0x1000000000010000 0xd000000000000000 0xffffffffffffffff
                                                           ^^^^^^^^^^^^^^^^^^
                     0x0001f80000000000 0x0001f80000000000  RWE    0x10000
      LOAD           0x0000000000010000 0xc000000000000000 0x0000000000000000
                     0x00000003f0000000 0x00000003f0000000  RWE    0x10000
      LOAD           0x3000000000010000 0xf000000000000000 0xffffffffffffffff
                                                           ^^^^^^^^^^^^^^^^^^
                     0x0000000000fc0000 0x0000000000fc0000  RWE    0x10000

makedumple uses max_paddr to calculate the number of sections for sparse memory model thus wrong number is obtained based on max_paddr=-1. This error could lead to the failure of copying /proc/kcore for RHEL-8.5 on ppc64le machine [1],

    $ uname -r
    4.18.0-312.el8.ppc64le
    $ makedumpfile /proc/kcore vmcore1
    get_mem_section: Could not validate mem_section.
    get_mm_sparsemem: Can't get the address of mem_section.

makedumpfile Failed.

Let's check if the phys_start of the segment is a valid physical address to fix this problem.

[1] https://bugzilla.redhat.com/show_bug.cgi?id=1965267

---
 elf_info.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
elf_info.c +1 -1
index e8affb7..9444847 100644
@@ -628,7 +628,7 @@ get_max_paddr(void)
 
 	for (i = 0; i < num_pt_loads; i++) {
 		pls = &pt_loads[i];
-		if (max_paddr < pls->phys_end)
+		if (pls->phys_start != NOT_PADDR  && max_paddr < pls->phys_end)
 			max_paddr = pls->phys_end;
 	}
 	return max_paddr;
Signature · 7 lines
2.31.1


_______________________________________________
kexec mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/kexec
#1 lore raw

-----Original Message-----

Quoted 29 lines

Kernel commit 464920104bf7adac12722035bfefb3d772eb04d8 "/proc/kcore: update physical address for kcore ram and text" sets an invalid paddr (-1=0xffffffffffffffff) for PT_LOAD segments of not direct mapped regions,

    $ readelf -l /proc/kcore
    Elf file type is CORE (Core file)
    Entry point 0x0
    There are 4 program headers, starting at offset 64

    Program Headers:
      Type           Offset             VirtAddr           PhysAddr
                     FileSiz            MemSiz              Flags  Align
      NOTE           0x0000000000000120 0x0000000000000000 0x0000000000000000
                     0x0000000000002320 0x0000000000000000         0x0
      LOAD           0x1000000000010000 0xd000000000000000 0xffffffffffffffff
                                                           ^^^^^^^^^^^^^^^^^^
                     0x0001f80000000000 0x0001f80000000000  RWE    0x10000
      LOAD           0x0000000000010000 0xc000000000000000 0x0000000000000000
                     0x00000003f0000000 0x00000003f0000000  RWE    0x10000
      LOAD           0x3000000000010000 0xf000000000000000 0xffffffffffffffff
                                                           ^^^^^^^^^^^^^^^^^^
                     0x0000000000fc0000 0x0000000000fc0000  RWE    0x10000

makedumple uses max_paddr to calculate the number of sections for sparse memory model thus wrong number is obtained based on max_paddr=-1. This error could lead to the failure of copying /proc/kcore for RHEL-8.5 on ppc64le machine [1],

Quoted 7 lines
    $ uname -r
    4.18.0-312.el8.ppc64le
    $ makedumpfile /proc/kcore vmcore1
    get_mem_section: Could not validate mem_section.
    get_mm_sparsemem: Can't get the address of mem_section.

makedumpfile Failed.

IIUC, makedumpfile has NOT been developed to filter /proc/kcore so far, except for the --mem-usage option, and this is the first patch for it.

It might be more handy than crash's snap.so extension, so I applied this:
https://github.com/makedumpfile/makedumpfile/commit/38d921a2ef50ebd36258097553626443ffe27496

But please note that maybe filtering /proc/kcore with makedumpfile (and also analyzing the dumpfile with crash) is still not mature enough for field or production use.

Thanks,
Kazu

Quoted 29 lines

Let's check if the phys_start of the segment is a valid physical address to fix this problem.

[1] https://bugzilla.redhat.com/show_bug.cgi?id=1965267

---
 elf_info.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
elf_info.c +1 -2
index e8affb7..9444847 100644
@@ -628,7 +628,7 @@ get_max_paddr(void)
 
 	for (i = 0; i < num_pt_loads; i++) {
 		pls = &pt_loads[i];
-		if (max_paddr < pls->phys_end)
+		if (pls->phys_start != NOT_PADDR  && max_paddr < pls->phys_end)
 			max_paddr = pls->phys_end;
 	}
 	return max_paddr;
--

2.31.1

Quoted 4 lines

_______________________________________________
kexec mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/kexec

_______________________________________________
kexec mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/kexec

author
#2 lore raw
Quoted 36 lines · HAGIO KAZUHITO(萩尾 一仁)

-----Original Message-----

Quoted 35 lines

Kernel commit 464920104bf7adac12722035bfefb3d772eb04d8 "/proc/kcore: update physical address for kcore ram and text" sets an invalid paddr (-1=0xffffffffffffffff) for PT_LOAD segments of not direct mapped regions,

    $ readelf -l /proc/kcore
    Elf file type is CORE (Core file)
    Entry point 0x0
    There are 4 program headers, starting at offset 64

    Program Headers:
      Type           Offset             VirtAddr           PhysAddr
                     FileSiz            MemSiz              Flags  Align
      NOTE           0x0000000000000120 0x0000000000000000 0x0000000000000000
                     0x0000000000002320 0x0000000000000000         0x0
      LOAD           0x1000000000010000 0xd000000000000000 0xffffffffffffffff
                                                           ^^^^^^^^^^^^^^^^^^
                     0x0001f80000000000 0x0001f80000000000  RWE    0x10000
      LOAD           0x0000000000010000 0xc000000000000000 0x0000000000000000
                     0x00000003f0000000 0x00000003f0000000  RWE    0x10000
      LOAD           0x3000000000010000 0xf000000000000000 0xffffffffffffffff
                                                           ^^^^^^^^^^^^^^^^^^
                     0x0000000000fc0000 0x0000000000fc0000  RWE    0x10000

makedumple uses max_paddr to calculate the number of sections for sparse memory model thus wrong number is obtained based on max_paddr=-1. This error could lead to the failure of copying /proc/kcore for RHEL-8.5 on ppc64le machine [1],

    $ uname -r
    4.18.0-312.el8.ppc64le
    $ makedumpfile /proc/kcore vmcore1
    get_mem_section: Could not validate mem_section.
    get_mm_sparsemem: Can't get the address of mem_section.
Quoted 7 lines
Quoted 1 line

makedumpfile Failed.

IIUC, makedumpfile has NOT been developed to filter /proc/kcore so far, except for the --mem-usage option, and this is the first patch for it.

It might be more handy than crash's snap.so extension, so I applied this:
https://github.com/makedumpfile/makedumpfile/commit/38d921a2ef50ebd36258097553626443ffe27496

Thanks for applying the patch!

Quoted 4 lines

But please note that maybe filtering /proc/kcore with makedumpfile (and also analyzing the dumpfile with crash) is still not mature enough for field or production use.

Thanks for the note!

Quoted 38 lines

Thanks,
Kazu

Quoted 33 lines

Let's check if the phys_start of the segment is a valid physical address to fix this problem.

[1] https://bugzilla.redhat.com/show_bug.cgi?id=1965267

---
 elf_info.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
elf_info.c +1 -2
index e8affb7..9444847 100644
@@ -628,7 +628,7 @@ get_max_paddr(void)
 
 	for (i = 0; i < num_pt_loads; i++) {
 		pls = &pt_loads[i];
-		if (max_paddr < pls->phys_end)
+		if (pls->phys_start != NOT_PADDR  && max_paddr < pls->phys_end)
 			max_paddr = pls->phys_end;
 	}
 	return max_paddr;
--

2.31.1

_______________________________________________
kexec mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/kexec

Signature · 8 lines
Best regards,
Coiby


_______________________________________________
kexec mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/kexec