| lists.openwall.net | lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC | |
|
Open Source and information security mailing list archives
| ||
|
Message-Id: <20240416061410.352-1-zhangyang@nfschina.com>
Date: Tue, 16 Apr 2024 14:14:11 +0800
From: zhangyang <zhangyang@...china.com>
To: rafael@...nel.org,
pavel@....cz,
len.brown@...el.com
Cc: linux-pm@...r.kernel.org,
linux-kernel@...r.kernel.org,
zhangyang <zhangyang@...china.com>
Subject: [PATCH] fix hib resume pages cache clean called in_irq (option hibernate=nocompress)
When kernel cmd option hibernate=nocompress,load_image
call swap_read_page. after every page read finished hib_end_io call page
cache clean, in aarch6464 flush_icache_range not allow call in interrupt.
This patch fix this problem by record all pages read in load_image,after all
pages read finished call flush_icache_range clean all reags in
load_image,flush_icache_range called in task.
call trace info:
[ 12.603468] WARNING: CPU: 6 PID: 0 at kernel/smp.c:424 smp_call_function_many+0x2fc/0x390
[ 12.612730] Modules linked in: raid10(E) raid456(E) libcrc32c(E) async_raid6_recov(E) async_memcpy(E) async_pq(E) async_xor(E) xor(E) xor_neon(E) async_tx(E) raid6_pq(E) raid1(E) raid0(E) multipath(E) linear(E) md_mod(E) hid_generic(E) usbhid(E) hid(E) uhci_hcd(E) ehci_hcd(E) arise_pro(OE) drm_kms_helper(E) evdev(E) efivars(E) clk_scpi(E) sd_mod(E) usbcore(E) drm(E) arm_scpi(E)
[ 12.650675] CPU: 6 PID: 0 Comm: swapper/6 Tainted: G W OE 5.4.0-100-generic #100.1+m38+21nfs5
[ 12.661512] Hardware name: LENOVO INVALID/FD2000ZX200MB1, BIOS W0AKT19B 08/16/2022
[ 12.670085] pstate: 20000085 (nzCv daIf -PAN -UAO)
[ 12.675506] pc : smp_call_function_many+0x2fc/0x390
[ 12.681026] lr : kick_all_cpus_sync+0x34/0x3c
[ 12.685954] sp : ffff800010033c40
[ 12.689700] x29: ffff800010033c40 x28: ffff80001152a1b8
[ 12.695712] x27: 0000000000000006 x26: 0000000000000001
[ 12.701723] x25: ffff0000780b8000 x24: 0000000000000000
[ 12.707735] x23: 0000000000000000 x22: ffff8000101b6ad0
[ 12.713746] x21: 0000000000000006 x20: ffff80001152a1b8
[ 12.719758] x19: ffff0026d3e08100 x18: 0000000000000000
[ 12.725769] x17: 000000004143b1b7 x16: 0000000000000008
[ 12.731781] x15: 0000000000004446 x14: 000000000000ba7e
[ 12.737792] x13: 000042cf7b84cf9c x12: 000042cf7b84cf9c
[ 12.743804] x11: 00000000000042cf x10: 0000000000000040
[ 12.749815] x9 : ffff80001154e748 x8 : ffff80001154e740
[ 12.755826] x7 : ffff0026d8400af8 x6 : ffff0026d3e08180
[ 12.761837] x5 : 0000000000001000 x4 : ffff0026d0b2f000
[ 12.767849] x3 : 0000000000000001 x2 : ffff800011159018
[ 12.773860] x1 : 0000000000000080 x0 : 0000000000000000
[ 12.779871] Call trace:
[ 12.782634] smp_call_function_many+0x2fc/0x390
[ 12.787759] kick_all_cpus_sync+0x34/0x3c
[ 12.792295] hib_end_io+0x104/0x180
[ 12.796240] bio_endio+0x148/0x1f0
[ 12.800086] blk_update_request+0xd8/0x3b0
[ 12.804718] blk_mq_end_request+0x34/0x150
[ 12.809352] nvme_complete_rq+0x74/0x230
[ 12.813788] nvme_pci_complete_rq+0x5c/0xd0
[ 12.818520] blk_mq_complete_request+0x10c/0x14c
[ 12.823743] nvme_complete_cqes+0xbc/0x1bc
[ 12.828375] nvme_irq+0x13c/0x15c
[ 12.832123] __handle_irq_event_percpu+0x68/0x240
[ 12.837445] handle_irq_event+0x68/0x1ac
[ 12.841881] handle_fasteoi_irq+0xc8/0x23c
[ 12.846514] __handle_domain_irq+0x80/0xe0
[ 12.851147] gic_handle_irq+0xd8/0x180
[ 12.855385] el1_irq+0xb8/0x140
[ 12.858934] arch_cpu_idle+0x40/0x1d0
[ 12.863075] do_idle+0x230/0x2dc
[ 12.866722] cpu_startup_entry+0x30/0xc0
[ 12.871159] secondary_start_kernel+0x138/0x184
[ 12.876283] ---[ end trace 9cf7b6db3165264d ]---
kernel/power/swap.c | 59 +++++++++++++++++++++++++++++++++++++++++----
1 file changed, 54 insertions(+), 5 deletions(-)
diff --git a/kernel/power/swap.c b/kernel/power/swap.c
index 5bc04bfe2..257622dcc 100644
--- a/kernel/power/swap.c
+++ b/kernel/power/swap.c
@@ -41,7 +41,6 @@ u32 swsusp_hardware_signature;
* in which case some architectures need these pages cleaning before they
* can be executed. We don't know which pages these may be, so clean the lot.
*/
-static bool clean_pages_on_read;
static bool clean_pages_on_decompress;
/*
@@ -256,9 +255,6 @@ static void hib_end_io(struct bio *bio)
if (bio_data_dir(bio) == WRITE)
put_page(page);
- else if (clean_pages_on_read)
- flush_icache_range((unsigned long)page_address(page),
- (unsigned long)page_address(page) + PAGE_SIZE);
if (bio->bi_status && !hb->error)
hb->error = bio->bi_status;
@@ -1084,6 +1080,50 @@ static int swap_reader_finish(struct swap_map_handle *handle)
return 0;
}
+struct swsusp_readpages {
+ unsigned long size;
+ unsigned long cursor;
+ struct page **pages;
+};
+
+static int swsusp_init_readpages(struct swsusp_readpages *read_pages, unsigned int nr_to_read)
+{
+ read_pages->pages = (struct page **)vzalloc(sizeof(struct page *) * nr_to_read);
+ if (!read_pages->pages) {
+ return -ENOMEM;
+ }
+
+ read_pages->size = nr_to_read;
+ read_pages->cursor = 0;
+
+ return 0;
+}
+
+
+static int swsusp_add_readpage(struct swsusp_readpages *read_pages, void *page_addr)
+{
+ if (read_pages->cursor >= read_pages->size) {
+ return -ENOMEM;
+ }
+
+ read_pages->pages[read_pages->cursor++] = virt_to_page(page_addr);
+ return 0;
+}
+
+static void swsusp_clean_readedpages(struct swsusp_readpages *read_pages, bool flush)
+{
+ unsigned long idx;
+
+ for (idx = 0; idx < read_pages->cursor; idx++) {
+ if (flush && read_pages->pages[idx])
+ flush_icache_range((unsigned long)page_address(read_pages->pages[idx]),
+ (unsigned long)page_address(read_pages->pages[idx] + PAGE_SIZE));
+ }
+
+ vfree(read_pages->pages);
+ return;
+}
+
/**
* load_image - load the image using the swap map handle
* @handle and the snapshot handle @snapshot
@@ -1101,10 +1141,14 @@ static int load_image(struct swap_map_handle *handle,
struct hib_bio_batch hb;
int err2;
unsigned nr_pages;
+ struct swsusp_readpages pages_to_clean;
hib_init_batch(&hb);
+ ret = swsusp_init_readpages(&pages_to_clean, nr_to_read);
+ if (!ret) {
+ return ret;
+ }
- clean_pages_on_read = true;
pr_info("Loading image data pages (%u pages)...\n", nr_to_read);
m = nr_to_read / 10;
if (!m)
@@ -1122,6 +1166,10 @@ static int load_image(struct swap_map_handle *handle,
ret = hib_wait_io(&hb);
if (ret)
break;
+ ret = swsusp_add_readpage(&pages_to_clean, data_of(*snapshot));
+ if (ret)
+ break;
+
if (!(nr_pages % m))
pr_info("Image loading progress: %3d%%\n",
nr_pages / m * 10);
@@ -1132,6 +1180,7 @@ static int load_image(struct swap_map_handle *handle,
stop = ktime_get();
if (!ret)
ret = err2;
+ swsusp_clean_readedpages(&pages_to_clean, !ret);
if (!ret) {
pr_info("Image loading done\n");
ret = snapshot_write_finalize(snapshot);
--
2.25.1
Powered by blists - more mailing lists