When we excute the following commands, we got oops rmmod ipmi_si cat /proc/ioports
[ 1623.482380] Unable to handle kernel paging request at virtual address ffff00000901d478 [ 1623.482382] Mem abort info: [ 1623.482383] ESR = 0x96000007 [ 1623.482385] Exception class = DABT (current EL), IL = 32 bits [ 1623.482386] SET = 0, FnV = 0 [ 1623.482387] EA = 0, S1PTW = 0 [ 1623.482388] Data abort info: [ 1623.482389] ISV = 0, ISS = 0x00000007 [ 1623.482390] CM = 0, WnR = 0 [ 1623.482393] swapper pgtable: 4k pages, 48-bit VAs, pgdp = 00000000d7d94a66 [ 1623.482395] [ffff00000901d478] pgd=000000dffbfff003, pud=000000dffbffe003, pmd=0000003f5d06e003, pte=0000000000000000 [ 1623.482399] Internal error: Oops: 96000007 [#1] SMP [ 1623.487407] Modules linked in: ipmi_si(E) nls_utf8 isofs rpcrdma ib_iser ib_srpt target_core_mod ib_srp scsi_transport_srp ib_ipoib rdma_ucm ib_umad rdma_cm ib_cm dm_mirror dm_region_hash dm_log iw_cm dm_mod aes_ce_blk crypto_simd cryptd aes_ce_cipher ses ghash_ce sha2_ce enclosure sha256_arm64 sg sha1_ce hisi_sas_v2_hw hibmc_drm sbsa_gwdt hisi_sas_main ip_tables mlx5_ib ib_uverbs marvell ib_core mlx5_core ixgbe mdio hns_dsaf ipmi_devintf hns_enet_drv ipmi_msghandler hns_mdio [last unloaded: ipmi_si] [ 1623.532410] CPU: 30 PID: 11438 Comm: cat Kdump: loaded Tainted: G E 5.0.0-rc3+ #168 [ 1623.541498] Hardware name: Huawei TaiShan 2280 /BC11SPCD, BIOS 1.37 11/21/2017 [ 1623.548822] pstate: a0000005 (NzCv daif -PAN -UAO) [ 1623.553684] pc : string+0x28/0x98 [ 1623.557040] lr : vsnprintf+0x368/0x5e8 [ 1623.560837] sp : ffff000013213a80 [ 1623.564191] x29: ffff000013213a80 x28: ffff00001138abb5 [ 1623.569577] x27: ffff000013213c18 x26: ffff805f67d06049 [ 1623.574963] x25: 0000000000000000 x24: ffff00001138abb5 [ 1623.580349] x23: 0000000000000fb7 x22: ffff0000117ed000 [ 1623.585734] x21: ffff000011188fd8 x20: ffff805f67d07000 [ 1623.591119] x19: ffff805f67d06061 x18: ffffffffffffffff [ 1623.596505] x17: 0000000000000200 x16: 0000000000000000 [ 1623.601890] x15: ffff0000117ed748 x14: ffff805f67d07000 [ 1623.607276] x13: ffff805f67d0605e x12: 0000000000000000 [ 1623.612661] x11: 0000000000000000 x10: 0000000000000000 [ 1623.618046] x9 : 0000000000000000 x8 : 000000000000000f [ 1623.623432] x7 : ffff805f67d06061 x6 : fffffffffffffffe [ 1623.628817] x5 : 0000000000000012 x4 : ffff00000901d478 [ 1623.634203] x3 : ffff0a00ffffff04 x2 : ffff805f67d07000 [ 1623.639588] x1 : ffff805f67d07000 x0 : ffffffffffffffff [ 1623.644974] Process cat (pid: 11438, stack limit = 0x000000008d4cbc10) [ 1623.651592] Call trace: [ 1623.654068] string+0x28/0x98 [ 1623.657071] vsnprintf+0x368/0x5e8 [ 1623.660517] seq_vprintf+0x70/0x98 [ 1623.668009] seq_printf+0x7c/0xa0 [ 1623.675530] r_show+0xc8/0xf8 [ 1623.682558] seq_read+0x330/0x440 [ 1623.689877] proc_reg_read+0x78/0xd0 [ 1623.697346] __vfs_read+0x60/0x1a0 [ 1623.704564] vfs_read+0x94/0x150 [ 1623.711339] ksys_read+0x6c/0xd8 [ 1623.717939] __arm64_sys_read+0x24/0x30 [ 1623.725077] el0_svc_common+0x120/0x148 [ 1623.732035] el0_svc_handler+0x30/0x40 [ 1623.738757] el0_svc+0x8/0xc [ 1623.744520] Code: d1000406 aa0103e2 54000149 b4000080 (39400085) [ 1623.753441] ---[ end trace f91b6a4937de9835 ]--- [ 1623.760871] Kernel panic - not syncing: Fatal exception [ 1623.768935] SMP: stopping secondary CPUs [ 1623.775718] Kernel Offset: disabled [ 1623.781998] CPU features: 0x002,21006008 [ 1623.788777] Memory Limit: none [ 1623.798329] Starting crashdump kernel... [ 1623.805202] Bye!
If io_setup is called successful in try_smi_init() but try_smi_init() goes out_err before calling ipmi_register_smi(), so ipmi_unregister_smi() will not be called while removing module. It leads to the resource that allocated in io_setup() can not be freed, but the name(DEVICE_NAME) of resource is freed while removing the module. It causes use-after-free when cat /proc/ioports.
Fix this by calling shutdown_smi() while removing the module and don't call release_region() if request_region() is not called to avoid error prints.
Fixes: 93c303d2045b ("ipmi_si: Clean up shutdown a bit") Cc: stable@vger.kernel.org Reported-by: NuoHan Qiao qiaonuohan@huawei.com Signed-off-by: Yang Yingliang yangyingliang@huawei.com --- drivers/char/ipmi/ipmi_si_intf.c | 2 ++ drivers/char/ipmi/ipmi_si_port_io.c | 3 +++ 2 files changed, 5 insertions(+)
diff --git a/drivers/char/ipmi/ipmi_si_intf.c b/drivers/char/ipmi/ipmi_si_intf.c index dc8603d..635e98a 100644 --- a/drivers/char/ipmi/ipmi_si_intf.c +++ b/drivers/char/ipmi/ipmi_si_intf.c @@ -2235,6 +2235,8 @@ static void cleanup_one_si(struct smi_info *smi_info)
if (smi_info->intf) ipmi_unregister_smi(smi_info->intf); + else + shutdown_smi(smi_info);
if (smi_info->pdev) { if (smi_info->pdev_registered) diff --git a/drivers/char/ipmi/ipmi_si_port_io.c b/drivers/char/ipmi/ipmi_si_port_io.c index ef6dffc..0c46a3f 100644 --- a/drivers/char/ipmi/ipmi_si_port_io.c +++ b/drivers/char/ipmi/ipmi_si_port_io.c @@ -53,6 +53,9 @@ static void port_cleanup(struct si_sm_io *io) unsigned int addr = io->addr_data; int idx;
+ if (io->regsize != 1 && io->regsize != 2 && io->regsize != 4) + return; + if (addr) { for (idx = 0; idx < io->io_size; idx++) release_region(addr + idx * io->regspacing,
On Fri, Jan 25, 2019 at 10:30:59AM +0800, Yang Yingliang wrote:
When we excute the following commands, we got oops rmmod ipmi_si cat /proc/ioports
snip...
If io_setup is called successful in try_smi_init() but try_smi_init() goes out_err before calling ipmi_register_smi(), so ipmi_unregister_smi() will not be called while removing module. It leads to the resource that allocated in io_setup() can not be freed, but the name(DEVICE_NAME) of resource is freed while removing the module. It causes use-after-free when cat /proc/ioports.
Fix this by calling shutdown_smi() while removing the module and don't call release_region() if request_region() is not called to avoid error prints.
Fixes: 93c303d2045b ("ipmi_si: Clean up shutdown a bit") Cc: stable@vger.kernel.org Reported-by: NuoHan Qiao qiaonuohan@huawei.com Signed-off-by: Yang Yingliang yangyingliang@huawei.com
drivers/char/ipmi/ipmi_si_intf.c | 2 ++ drivers/char/ipmi/ipmi_si_port_io.c | 3 +++ 2 files changed, 5 insertions(+)
diff --git a/drivers/char/ipmi/ipmi_si_intf.c b/drivers/char/ipmi/ipmi_si_intf.c index dc8603d..635e98a 100644 --- a/drivers/char/ipmi/ipmi_si_intf.c +++ b/drivers/char/ipmi/ipmi_si_intf.c @@ -2235,6 +2235,8 @@ static void cleanup_one_si(struct smi_info *smi_info) if (smi_info->intf) ipmi_unregister_smi(smi_info->intf);
- else
shutdown_smi(smi_info);
This is completely the wrong way to fix this. The general principle is that a function cleans up for itself if it returns an error. If you add hacks other places for a function failing you end up with a mess.
I think the right way to fix this is to add something like:
if (rv && new_smi->io.io_size && smi_info->io.io_cleanup) { smi_info->io.io_cleanup(&smi_info->io); smi_info->io.io_cleanup = NULL; }
at the end of try_smi_init().
-corey
if (smi_info->pdev) { if (smi_info->pdev_registered) diff --git a/drivers/char/ipmi/ipmi_si_port_io.c b/drivers/char/ipmi/ipmi_si_port_io.c index ef6dffc..0c46a3f 100644 --- a/drivers/char/ipmi/ipmi_si_port_io.c +++ b/drivers/char/ipmi/ipmi_si_port_io.c @@ -53,6 +53,9 @@ static void port_cleanup(struct si_sm_io *io) unsigned int addr = io->addr_data; int idx;
- if (io->regsize != 1 && io->regsize != 2 && io->regsize != 4)
return;- if (addr) { for (idx = 0; idx < io->io_size; idx++) release_region(addr + idx * io->regspacing,
-- 1.8.3
On 2019/1/26 4:35, Corey Minyard wrote:
On Fri, Jan 25, 2019 at 10:30:59AM +0800, Yang Yingliang wrote:
When we excute the following commands, we got oops rmmod ipmi_si cat /proc/ioports
snip...
If io_setup is called successful in try_smi_init() but try_smi_init() goes out_err before calling ipmi_register_smi(), so ipmi_unregister_smi() will not be called while removing module. It leads to the resource that allocated in io_setup() can not be freed, but the name(DEVICE_NAME) of resource is freed while removing the module. It causes use-after-free when cat /proc/ioports.
Fix this by calling shutdown_smi() while removing the module and don't call release_region() if request_region() is not called to avoid error prints.
Fixes: 93c303d2045b ("ipmi_si: Clean up shutdown a bit") Cc: stable@vger.kernel.org Reported-by: NuoHan Qiao qiaonuohan@huawei.com Signed-off-by: Yang Yingliang yangyingliang@huawei.com
drivers/char/ipmi/ipmi_si_intf.c | 2 ++ drivers/char/ipmi/ipmi_si_port_io.c | 3 +++ 2 files changed, 5 insertions(+)
diff --git a/drivers/char/ipmi/ipmi_si_intf.c b/drivers/char/ipmi/ipmi_si_intf.c index dc8603d..635e98a 100644 --- a/drivers/char/ipmi/ipmi_si_intf.c +++ b/drivers/char/ipmi/ipmi_si_intf.c @@ -2235,6 +2235,8 @@ static void cleanup_one_si(struct smi_info *smi_info) if (smi_info->intf) ipmi_unregister_smi(smi_info->intf);
- else
shutdown_smi(smi_info);This is completely the wrong way to fix this. The general principle is that a function cleans up for itself if it returns an error. If you add hacks other places for a function failing you end up with a mess.
I think the right way to fix this is to add something like:
if (rv && new_smi->io.io_size && smi_info->io.io_cleanup) { smi_info->io.io_cleanup(&smi_info->io); smi_info->io.io_cleanup = NULL; }
at the end of try_smi_init().
-corey
OK, I will do some test, and send v2 later.
Thanks, Yang
if (smi_info->pdev) { if (smi_info->pdev_registered) diff --git a/drivers/char/ipmi/ipmi_si_port_io.c b/drivers/char/ipmi/ipmi_si_port_io.c index ef6dffc..0c46a3f 100644 --- a/drivers/char/ipmi/ipmi_si_port_io.c +++ b/drivers/char/ipmi/ipmi_si_port_io.c @@ -53,6 +53,9 @@ static void port_cleanup(struct si_sm_io *io) unsigned int addr = io->addr_data; int idx;
- if (io->regsize != 1 && io->regsize != 2 && io->regsize != 4)
return;- if (addr) { for (idx = 0; idx < io->io_size; idx++) release_region(addr + idx * io->regspacing,
-- 1.8.3
.
linux-stable-mirror@lists.linaro.org
-
Corey Minyard -
Yang Yingliang