[Crash-utility] Re: PATCH] Fix "kmem -i" and "swap" commands on Linux 6.10-rc1 and later kernels

On 2024/06/12 15:34, lijiang wrote: > > --- a/memory.c > > +++ b/memory.c > > @@ -16075,6 +16075,8 @@ dump_swap_info(ulong swapflags, ulong *totalswap_pages, ulong *totalused_pages) > > char buf3[BUFSIZE]; > > char buf4[BUFSIZE]; > > char buf5[BUFSIZE]; > > + int swap_file_is_file = > > + STREQ(MEMBER_TYPE_NAME("swap_info_struct", "swap_file"), "file"); > > > > > > This patch looks good, but I still have one question: > > The 'swap_file' has been added into the struct swap_info_struct since linux v2.6.12-rc2, > > and until now it(its type) has been never changed, it indicates that the above checking is always true. > > > > STREQ(MEMBER_TYPE_NAME("swap_info_struct", "swap_file"), "file") > > > > Is that expected behavior? Just want to confirm with you. > > Yes. > > > Thank you for the confirmation, and explanation below, Kazu. > > As far as I checked, swap_file was changed from 'dentry *' to 'file *' > at somewhere between v2.4.0 and v2.6.0, so not always true. > > Seems no full history in linux github. I checked it from the linux initial git repo: > > commit 1da177e4c3f41524e886b7f1b8a0c1fc7321cac2 (tag: v2.6.12-rc2) > Author: Linus Torvalds <torvalds(a)ppc970.osdl.org <mailto: torvalds(a)ppc970.osdl.org&gt;&gt; > Date: Sat Apr 16 15:20:36 2005 -0700 > > Linux-2.6.12-rc2 > > Initial git repository build. I'm not bothering with the full history, > even though we have it. We can create a separate "historical" git > archive of that later if we want to, and in the meantime it's about > 3.2GB when imported into git - space that would just make the early > git days unnecessarily complicated, when we don't have a lot of good > infrastructure for it. > > Let it rip! > > > get_pathname()'s first arg is dentry, we have to choose the else-if block > without old_block_size, so I decided to use its type information directly > > } else if (VALID_MEMBER(swap_info_struct_old_block_size) || > swap_file_is_file) { > > The other if / else blocks are for very old kernels, so maybe they won't > be used, but if we don't remove these, I think this is a reasonable way. > > Sounds good to me. > > Does this answer your question? > > Yes. Thank you, Kazu. I have no other issues. For the patch: Ack. Good, thank you for your ack. btw, I forgot to add the From: line below to the commit log, it would be helpful if you would add this when merging: From: Kazuhito Hagio <k-hagio-ab(a)nec.com&gt;

Thanks, Kazu