Re: [RFC net] net: make page pool stall netdev unregistration to avoid IOMMU crashes

On 2024/8/10 11:57, Jakub Kicinski wrote:

On Fri, 9 Aug 2024 14:06:02 +0800 Yonglong Liu wrote:

quoted

[ 7724.272853] hns3 0000:7d:01.0: page_pool_release_retry(): eno1v0
stalled pool shutdown: id 553, 82 inflight 6706 sec (hold netdev: 1855491)

Alright :( You gotta look around for those 82 pages somehow with drgn.
bpftrace+kfunc the work that does the periodic print to get the address
of the page pool struct and then look around for pages from that pp.. :(

I spent some time to learn how to use the drgn, and found those page, 
but I think those page

is allocated by the hns3 driver, how to find out who own those page now?

The following is a page info:

*(struct page *)0xfffffe004359de80 = { .flags = (unsigned 
long)432345014471753728, .lru = (struct list_head){ .next = (struct 
list_head *)0xdead000000000040, .prev = (struct list_head 
*)0xffff002105c18000, }, .__filler = (void *)0xdead000000000040, 
.mlock_count = (unsigned int)96567296, .buddy_list = (struct list_head){ 
.next = (struct list_head *)0xdead000000000040, .prev = (struct 
list_head *)0xffff002105c18000, }, .pcp_list = (struct list_head){ .next 
= (struct list_head *)0xdead000000000040, .prev = (struct list_head 
*)0xffff002105c18000, }, .mapping = (struct address_space *)0x0, .index 
= (unsigned long)4223168512, .share = (unsigned long)4223168512, 
.private = (unsigned long)2, .pp_magic = (unsigned 
long)16045481047390945344, .pp = (struct page_pool *)0xffff002105c18000, 
._pp_mapping_pad = (unsigned long)0, .dma_addr = (unsigned 
long)4223168512, .pp_ref_count = (atomic_long_t){ .counter = (s64)2, }, 
.compound_head = (unsigned long)16045481047390945344, .pgmap = (struct 
dev_pagemap *)0xdead000000000040, .zone_device_data = (void 
*)0xffff002105c18000, .callback_head = (struct callback_head){ .next = 
(struct callback_head *)0xdead000000000040, .func = (void (*)(struct 
callback_head *))0xffff002105c18000, }, ._mapcount = (atomic_t){ 
.counter = (int)-1, }, .page_type = (unsigned int)4294967295, ._refcount 
= (atomic_t){ .counter = (int)1, }, .memcg_data = (unsigned long)0, }