Messages in this thread Patch in this message | | | From | Fan Ni <> | Subject | Re: [PATCH] cxl/hdm: Fix hdm decoder init by adding COMMIT field check | Date | Fri, 3 Mar 2023 21:54:54 +0000 |
| |
On Fri, Mar 03, 2023 at 12:58:10PM -0800, Dan Williams wrote:
> Fan Ni wrote: > > Add COMMIT field check aside with existing COMMITTED field check during > > hdm decoder initialization to avoid a system crash during module removal > > after destroying a region which leaves the COMMIT field being reset while > > the COMMITTED field still being set. > > > > In current kernel implementation, when destroying a region (cxl > > destroy-region),the decoders associated to the region will be reset > > as that in cxl_decoder_reset, where the COMMIT field will be reset. > > However, resetting COMMIT field will not automatically reset the > > COMMITTED field, causing a situation where COMMIT is reset (0) while > > COMMITTED is set (1) after the region is destroyed. Later, when > > init_hdm_decoder is called (during modprobe), current code only check > > the COMMITTED to decide whether the decoder is enabled or not. Since > > the COMMITTED will be 1 and the code treats the decoder as enabled, > > which will cause unexpected behaviour. > > > > Before the fix, a system crash was observed when performing following > > steps: > > 1. modprobe -a cxl_acpi cxl_core cxl_pci cxl_port cxl_mem > > 2. cxl create-region -m -d decoder0.0 -w 1 mem0 -s 256M > > 3. cxl destroy-region region0 -f > > 4. rmmod cxl_acpi cxl_pci cxl_port cxl_mem cxl_pmem cxl_core > > 5. modprobe -a cxl_acpi cxl_core cxl_pci cxl_port cxl_mem (showing > > "no CXL window for range 0x0:0xffffffffffffffff" error message) > > 6. rmmod cxl_acpi cxl_pci cxl_port cxl_mem cxl_pmem cxl_core (kernel > > crash at cxl_dpa_release due to dpa_res has been freed when destroying > > the region). > > I think a separate fix for that crash is needed, can you send the > backtrace? I.e. I worry that crash can be triggered by other means. Hi Dan, See backtrace below.
[ 130.299394] BUG: kernel NULL pointer dereference, address: 0000000000000008 [ 130.299907] #PF: supervisor read access in kernel mode [ 130.299907] #PF: error_code(0x0000) - not-present page [ 130.299907] PGD 0 P4D 0 [ 130.299907] Oops: 0000 [#1] PREEMPT SMP NOPTI [ 130.299907] CPU: 13 PID: 467 Comm: rmmod Not tainted 6.2.0-rc6-00024-g3ea761ec9dd5 #58 [ 130.299907] Hardware name: QEMU Standard PC (Q35 + ICH9, 2009), BIOS rel-1.16.1-0-g3208b098f51a-prebuilt.qemu.org 04/01/2014 [ 130.299907] RIP: 0010:__cxl_dpa_release+0x3c/0xb0 [cxl_core] [ 130.299907] Code: ff ff 48 8b 7d 40 4c 8b a8 d8 02 00 00 e8 5c a6 ff ff 4c 8b a5 28 03 00 00 48 89 c3 48 8b 85 20 03 00 00 4d 8b ad 40 03 00 00 <48> 8b 50 08 4c 8b 30 49 81 c5 90 00 00 00 4c 89 ef 48 83 c2 01 4c [ 130.299907] RSP: 0018:ffffc9000075fae0 EFLAGS: 00000246 [ 130.299907] RAX: 0000000000000000 RBX: ffff88810250cc00 RCX: 0000000000000000 [ 130.299907] RDX: 0000000000000001 RSI: ffff8881008d25e8 RDI: ffff88810250cc00 [ 130.299907] RBP: ffff88810250d000 R08: 0000000000000001 R09: ffffffff8182b400 [ 130.299907] R10: ffff888101fd7238 R11: ffff888201c1f406 R12: 0000000000000000 [ 130.299907] R13: 0000000000000000 R14: ffff88810250ce90 R15: ffff88810250ce8c [ 130.299907] FS: 00007f53b3884c40(0000) GS:ffff888277d40000(0000) knlGS:0000000000000000 [ 130.299907] CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 [ 130.299907] CR2: 0000000000000008 CR3: 000000010285c000 CR4: 00000000000006e0 [ 130.299907] Call Trace: [ 130.299907] <TASK> [ 130.299907] cxl_dpa_release+0x18/0x30 [cxl_core] [ 130.299907] release_nodes+0x40/0x70 [ 130.299907] devres_release_all+0x86/0xc0 [ 130.299907] device_unbind_cleanup+0x9/0x70 [ 130.299907] device_release_driver_internal+0xe9/0x160 [ 130.299907] bus_remove_device+0xd3/0x140 [ 130.299907] device_del+0x186/0x3d0 [ 130.299907] ? _raw_spin_unlock_irqrestore+0x16/0x30 [ 130.299907] ? devres_remove+0xcb/0xf0 [ 130.299907] device_unregister+0xe/0x60 [ 130.299907] ? __pfx_devm_action_release+0x10/0x10 [ 130.299907] devres_release+0x22/0x50 [ 130.299907] devm_release_action+0x33/0x60 [ 130.299907] ? __pfx_unregister_port+0x10/0x10 [cxl_core] [ 130.299907] delete_endpoint+0x7a/0x80 [cxl_core] [ 130.299907] release_nodes+0x40/0x70 [ 130.299907] devres_release_all+0x86/0xc0 [ 130.299907] device_unbind_cleanup+0x9/0x70 [ 130.299907] device_release_driver_internal+0xe9/0x160 [ 130.299907] bus_remove_device+0xd3/0x140 [ 130.299907] device_del+0x186/0x3d0 [ 130.299907] cdev_device_del+0x10/0x30 [ 130.299907] cxl_memdev_unregister+0x36/0x40 [cxl_core] [ 130.299907] release_nodes+0x40/0x70 [ 130.299907] devres_release_all+0x86/0xc0 [ 130.299907] device_unbind_cleanup+0x9/0x70 [ 130.299907] device_release_driver_internal+0xe9/0x160 [ 130.299907] driver_detach+0x3f/0x80 [ 130.299907] bus_remove_driver+0x50/0xd0 [ 130.299907] pci_unregister_driver+0x36/0x80 [ 130.299907] __x64_sys_delete_module+0x191/0x270 [ 130.299907] ? fpregs_assert_state_consistent+0x1d/0x50 [ 130.299907] ? exit_to_user_mode_prepare+0x36/0x120 [ 130.299907] do_syscall_64+0x3b/0x90 [ 130.299907] entry_SYSCALL_64_after_hwframe+0x72/0xdc [ 130.299907] RIP: 0033:0x7f53b3126c9b [ 130.299907] Code: 73 01 c3 48 8b 0d 95 21 0f 00 f7 d8 64 89 01 48 83 c8 ff c3 66 2e 0f 1f 84 00 00 00 00 00 90 f3 0f 1e fa b8 b0 00 00 00 0f 05 <48> 3d 01 f0 ff ff 73 01 c3 48 8b 0d 65 21 0f 00 f7 d8 64 89 01 48 [ 130.299907] RSP: 002b:00007fff5a72c558 EFLAGS: 00000206 ORIG_RAX: 00000000000000b0 [ 130.299907] RAX: ffffffffffffffda RBX: 000056037a16e790 RCX: 00007f53b3126c9b [ 130.299907] RDX: 000000000000000a RSI: 0000000000000800 RDI: 000056037a16e7f8 [ 130.299907] RBP: 0000000000000000 R08: 0000000000000000 R09: 0000000000000000 [ 130.299907] R10: 00007f53b31beac0 R11: 0000000000000206 R12: 00007fff5a72c7b0 [ 130.299907] R13: 000056037a16d2a0 R14: 00007fff5a72cdb7 R15: 000056037a16e790 [ 130.299907] </TASK> [ 130.299907] Modules linked in: cxl_mem cxl_pmem cxl_port cxl_pci(-) cxl_acpi cxl_core dax_pmem nd_pmem nd_btt [last unloaded: cxl_core] [ 130.299907] CR2: 0000000000000008 [ 130.357813] ---[ end trace 0000000000000000 ]--- [ 130.358811] RIP: 0010:__cxl_dpa_release+0x3c/0xb0 [cxl_core] [ 130.360039] Code: ff ff 48 8b 7d 40 4c 8b a8 d8 02 00 00 e8 5c a6 ff ff 4c 8b a5 28 03 00 00 48 89 c3 48 8b 85 20 03 00 00 4d 8b ad 40 03 00 00 <48> 8b 50 08 4c 8b 30 49 81 c5 90 00 00 00 4c 89 ef 48 83 c2 01 4c [ 130.363227] RSP: 0018:ffffc9000075fae0 EFLAGS: 00000246 [ 130.364292] RAX: 0000000000000000 RBX: ffff88810250cc00 RCX: 0000000000000000 [ 130.365400] RDX: 0000000000000001 RSI: ffff8881008d25e8 RDI: ffff88810250cc00 [ 130.366645] RBP: ffff88810250d000 R08: 0000000000000001 R09: ffffffff8182b400 [ 130.368025] R10: ffff888101fd7238 R11: ffff888201c1f406 R12: 0000000000000000 [ 130.369337] R13: 0000000000000000 R14: ffff88810250ce90 R15: ffff88810250ce8c [ 130.370531] FS: 00007f53b3884c40(0000) GS:ffff888277d40000(0000) knlGS:0000000000000000 [ 130.372515] CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 [ 130.373567] CR2: 0000000000000008 CR3: 000000010285c000 CR4: 00000000000006e0
> > > > > The patch fixed the above issue, and is tested based on follow patch series: > > > > [PATCH 00/18] CXL RAM and the 'Soft Reserved' => 'System RAM' default > > Message-ID: 167601992097.1924368.18291887895351917895.stgit@dwillia2-xfh.jf.intel.com > > > > Signed-off-by: Fan Ni <fan.ni@samsung.com> > > --- > > drivers/cxl/core/hdm.c | 8 +++++--- > > 1 file changed, 5 insertions(+), 3 deletions(-) > > > > diff --git a/drivers/cxl/core/hdm.c b/drivers/cxl/core/hdm.c > > index 80eccae6ba9e..6cf854c949f0 100644 > > --- a/drivers/cxl/core/hdm.c > > +++ b/drivers/cxl/core/hdm.c > > @@ -695,6 +695,7 @@ static int init_hdm_decoder(struct cxl_port *port, struct cxl_decoder *cxld, > > struct cxl_endpoint_decoder *cxled = NULL; > > u64 size, base, skip, dpa_size; > > bool committed; > > + bool should_commit; > > u32 remainder; > > int i, rc; > > u32 ctrl; > > @@ -710,10 +711,11 @@ static int init_hdm_decoder(struct cxl_port *port, struct cxl_decoder *cxld, > > base = ioread64_hi_lo(hdm + CXL_HDM_DECODER0_BASE_LOW_OFFSET(which)); > > size = ioread64_hi_lo(hdm + CXL_HDM_DECODER0_SIZE_LOW_OFFSET(which)); > > committed = !!(ctrl & CXL_HDM_DECODER0_CTRL_COMMITTED); > > + should_commit = !!(ctrl & CXL_HDM_DECODER0_CTRL_COMMIT); > > This change looks like a good idea in general given the ambiguity of > 'committed'. However just combine the two checks into the @committed > variable with something like this: > > commit_mask = CXL_HDM_DECODER0_CTRL_COMMITTED|CXL_HDM_DECODER0_CTRL_COMMIT; > committed = (ctrl & commit_mask) == commit_mask;
| |