Hi Huang Shijie,
-----Original Message-----
> 1.) The bitmap of vmcore file can be very big in the server kernel panic,
> such as over 256M.
>
> This patch uses mmap/madvise to improve the read speed
> of the bitmap in the non-FLAT_FORMAT code path.
>
> Use MADV_SEQUENTIAL for madvise, it will trigger aggressively
> readahead for reading the bitmap.
I'm not familiar with the detailed behavior of madvise() and a little
concerned about the sentence "may be freed soon after they are accessed"
below:
MADV_SEQUENTIAL
Expect page references in sequential order. (Hence, pages in
the given range can be aggressively read ahead, and may be freed
soon after they are accessed.)
MADV_WILLNEED
Expect access in the near future. (Hence, it might be a good
idea to read some pages ahead.)
dd->bitmap is often used after the initialization process, is there
no side-effect?
So if MADV_WILLNEED has same or some good effect, I think it will be
easier to accept. Any benchmark and information?
Thanks,
Kazu
>
> 2.) Test.
>
> The files:
> vmlinux: 272M
> vmcore : 23G (bitmap is 281018368 bytes)
>
> #cat ./commands.txt
> quit
>
> Before this patch:
> #echo 3 > /proc/sys/vm/drop_caches;
> #time ./crash -i ./commands.txt /root/t/vmlinux /root/t/vmcore > /dev/null 2>&1
> ............................
> real 0m55.217s
> user 0m15.114s
> sys 0m3.560s
> ............................
>
> After this patch:
> #echo 3 > /proc/sys/vm/drop_caches;
> #time ./crash -i ./commands.txt /root/t/vmlinux /root/t/vmcore > /dev/null 2>&1
> ............................
> real 0m46.781s
> user 0m19.693s
> sys 0m1.642s
> ............................
>
> Signed-off-by: Huang Shijie <shijie@xxxxxxxxxxxxxxxxxxxxxx>
> ---
> This patch is based on patch:
> diskdump: Optimize the boot time
>
> ---
> diskdump.c | 39 ++++++++++++++++-----------------------
> 1 file changed, 16 insertions(+), 23 deletions(-)
>
> diff --git a/diskdump.c b/diskdump.c
> index d30db9d..c60b283 100644
> --- a/diskdump.c
> +++ b/diskdump.c
> @@ -724,10 +724,6 @@ restart:
>
> offset = (off_t)block_size * (1 + header->sub_hdr_size);
>
> - if ((dd->bitmap = malloc(bitmap_len)) == NULL)
> - error(FATAL, "%s: cannot malloc bitmap buffer\n",
> - DISKDUMP_VALID() ? "diskdump" : "compressed kdump");
> -
> dd->dumpable_bitmap = calloc(bitmap_len, 1);
>
> if (CRASHDEBUG(8))
> @@ -736,30 +732,23 @@ restart:
> (ulonglong)offset);
>
> if (FLAT_FORMAT()) {
> + if ((dd->bitmap = malloc(bitmap_len)) == NULL)
> + error(FATAL, "%s: cannot malloc bitmap buffer\n",
> + DISKDUMP_VALID() ? "diskdump" : "compressed kdump");
> +
> if (!read_flattened_format(dd->dfd, offset, dd->bitmap, bitmap_len)) {
> error(INFO, "%s: cannot read memory bitmap\n",
> DISKDUMP_VALID() ? "diskdump" : "compressed kdump");
> goto err;
> }
> } else {
> - if (lseek(dd->dfd, offset, SEEK_SET) == failed) {
> - error(INFO, "%s: cannot lseek memory bitmap\n",
> + dd->bitmap = mmap(NULL, bitmap_len, PROT_READ,
> + MAP_SHARED, dd->dfd, offset);
> + if (dd->bitmap == MAP_FAILED)
> + error(FATAL, "%s: cannot mmap bitmap buffer\n",
> DISKDUMP_VALID() ? "diskdump" : "compressed kdump");
> - goto err;
> - }
> - bufptr = dd->bitmap;
> - len = bitmap_len;
> - while (len) {
> - bytes_read = read(dd->dfd, bufptr, len);
> - if (bytes_read <= 0) {
> - error(INFO, "%s: cannot read memory bitmap\n",
> - DISKDUMP_VALID() ? "diskdump"
> - : "compressed kdump");
> - goto err;
> - }
> - len -= bytes_read;
> - bufptr += bytes_read;
> - }
> +
> + madvise(dd->bitmap, bitmap_len, MADV_SEQUENTIAL);
> }
>
> if (dump_is_partial(header))
> @@ -920,8 +909,12 @@ err:
> free(sub_header);
> if (sub_header_kdump)
> free(sub_header_kdump);
> - if (dd->bitmap)
> - free(dd->bitmap);
> + if (dd->bitmap) {
> + if (FLAT_FORMAT())
> + free(dd->bitmap);
> + else
> + munmap(dd->bitmap, dd->bitmap_len);
> + }
> if (dd->dumpable_bitmap)
> free(dd->dumpable_bitmap);
> if (dd->notes_buf)
> --
> 2.30.2
--
Crash-utility mailing list
Crash-utility@xxxxxxxxxx
https://listman.redhat.com/mailman/listinfo/crash-utility
Contribution Guidelines: https://github.com/crash-utility/crash/wiki
