| Message ID | 20220124145110.442335-1-jmeneghi@redhat.com (mailing list archive) |
|---|---|
| State | Accepted |
| Headers | show |
| Series | scsi: bnx2fc: make bnx2fc_recv_frame mp safe | expand |
Can I please get a review and approval for this patch? I'd like this to get into v5.17-rc2 Thanks, /John On 1/24/22 09:51, John Meneghini wrote: > Running tests with a debug kernel shows that bnx2fc_recv_frame is > modifying the per_cpu lport stats counters in a non-mpsafe way. > Just boot a debug kernel and run the bnx2fc driver with the hardware > enabled. > > [ 1391.699147] BUG: using smp_processor_id() in preemptible [00000000] code: bnx2fc_ > [ 1391.699160] caller is bnx2fc_recv_frame+0xbf9/0x1760 [bnx2fc] > [ 1391.699174] CPU: 2 PID: 4355 Comm: bnx2fc_l2_threa Kdump: loaded Tainted: G B > [ 1391.699180] Hardware name: HP ProLiant DL120 G7, BIOS J01 07/01/2013 > [ 1391.699183] Call Trace: > [ 1391.699188] dump_stack_lvl+0x57/0x7d > [ 1391.699198] check_preemption_disabled+0xc8/0xd0 > [ 1391.699205] bnx2fc_recv_frame+0xbf9/0x1760 [bnx2fc] > [ 1391.699215] ? do_raw_spin_trylock+0xb5/0x180 > [ 1391.699221] ? bnx2fc_npiv_create_vports.isra.0+0x4e0/0x4e0 [bnx2fc] > [ 1391.699229] ? bnx2fc_l2_rcv_thread+0xb7/0x3a0 [bnx2fc] > [ 1391.699240] bnx2fc_l2_rcv_thread+0x1af/0x3a0 [bnx2fc] > [ 1391.699250] ? bnx2fc_ulp_init+0xc0/0xc0 [bnx2fc] > [ 1391.699258] kthread+0x364/0x420 > [ 1391.699263] ? _raw_spin_unlock_irq+0x24/0x50 > [ 1391.699268] ? set_kthread_struct+0x100/0x100 > [ 1391.699273] ret_from_fork+0x22/0x30 > > To fix the problem: restore the old get_cpu/put_cpu code with some > modifications to reduce the size of the critical section. > > Fixes: d576a5e80cd0 ("bnx2fc: Improve stats update mechanism") > Tested-by: Guangwu Zhang <guazhang@redhat.com> > Signed-off-by: John Meneghini <jmeneghi@redhat.com> > --- > drivers/scsi/bnx2fc/bnx2fc_fcoe.c | 21 +++++++++++++-------- > 1 file changed, 13 insertions(+), 8 deletions(-) > > diff --git a/drivers/scsi/bnx2fc/bnx2fc_fcoe.c b/drivers/scsi/bnx2fc/bnx2fc_fcoe.c > index 71fa62bd3083..e41a94dc2d1f 100644 > --- a/drivers/scsi/bnx2fc/bnx2fc_fcoe.c > +++ b/drivers/scsi/bnx2fc/bnx2fc_fcoe.c > @@ -508,7 +508,8 @@ static int bnx2fc_l2_rcv_thread(void *arg) > > static void bnx2fc_recv_frame(struct sk_buff *skb) > { > - u32 fr_len; > + u64 crc_err; > + u32 fr_len, fr_crc; > struct fc_lport *lport; > struct fcoe_rcv_info *fr; > struct fc_stats *stats; > @@ -542,6 +543,11 @@ static void bnx2fc_recv_frame(struct sk_buff *skb) > skb_pull(skb, sizeof(struct fcoe_hdr)); > fr_len = skb->len - sizeof(struct fcoe_crc_eof); > > + stats = per_cpu_ptr(lport->stats, get_cpu()); > + stats->RxFrames++; > + stats->RxWords += fr_len / FCOE_WORD_TO_BYTE; > + put_cpu(); > + > fp = (struct fc_frame *)skb; > fc_frame_init(fp); > fr_dev(fp) = lport; > @@ -624,16 +630,15 @@ static void bnx2fc_recv_frame(struct sk_buff *skb) > return; > } > > - stats = per_cpu_ptr(lport->stats, smp_processor_id()); > - stats->RxFrames++; > - stats->RxWords += fr_len / FCOE_WORD_TO_BYTE; > + fr_crc = le32_to_cpu(fr_crc(fp)); > > - if (le32_to_cpu(fr_crc(fp)) != > - ~crc32(~0, skb->data, fr_len)) { > - if (stats->InvalidCRCCount < 5) > + if (unlikely(fr_crc != ~crc32(~0, skb->data, fr_len))) { > + stats = per_cpu_ptr(lport->stats, get_cpu()); > + crc_err = (stats->InvalidCRCCount++); > + put_cpu(); > + if (crc_err < 5) > printk(KERN_WARNING PFX "dropping frame with " > "CRC error\n"); > - stats->InvalidCRCCount++; > kfree_skb(skb); > return; > }
Hi John, > -----Original Message----- > From: John Meneghini <jmeneghi@redhat.com> > Sent: Monday, January 24, 2022 8:21 PM > To: Saurav Kashyap <skashyap@marvell.com> > Cc: Nilesh Javali <njavali@marvell.com>; linux-scsi@vger.kernel.org; linux- > kernel@vger.kernel.org; GR-QLogic-Storage-Upstream <GR-QLogic-Storage- > Upstream@marvell.com>; mlombard@redhat.com; guazhang@redhat.com > Subject: [EXT] [PATCH] scsi: bnx2fc: make bnx2fc_recv_frame mp safe > > External Email > > ---------------------------------------------------------------------- > Running tests with a debug kernel shows that bnx2fc_recv_frame is > modifying the per_cpu lport stats counters in a non-mpsafe way. > Just boot a debug kernel and run the bnx2fc driver with the hardware > enabled. > > [ 1391.699147] BUG: using smp_processor_id() in preemptible [00000000] > code: bnx2fc_ > [ 1391.699160] caller is bnx2fc_recv_frame+0xbf9/0x1760 [bnx2fc] > [ 1391.699174] CPU: 2 PID: 4355 Comm: bnx2fc_l2_threa Kdump: loaded > Tainted: G B > [ 1391.699180] Hardware name: HP ProLiant DL120 G7, BIOS J01 > 07/01/2013 > [ 1391.699183] Call Trace: > [ 1391.699188] dump_stack_lvl+0x57/0x7d > [ 1391.699198] check_preemption_disabled+0xc8/0xd0 > [ 1391.699205] bnx2fc_recv_frame+0xbf9/0x1760 [bnx2fc] > [ 1391.699215] ? do_raw_spin_trylock+0xb5/0x180 > [ 1391.699221] ? bnx2fc_npiv_create_vports.isra.0+0x4e0/0x4e0 [bnx2fc] > [ 1391.699229] ? bnx2fc_l2_rcv_thread+0xb7/0x3a0 [bnx2fc] > [ 1391.699240] bnx2fc_l2_rcv_thread+0x1af/0x3a0 [bnx2fc] > [ 1391.699250] ? bnx2fc_ulp_init+0xc0/0xc0 [bnx2fc] > [ 1391.699258] kthread+0x364/0x420 > [ 1391.699263] ? _raw_spin_unlock_irq+0x24/0x50 > [ 1391.699268] ? set_kthread_struct+0x100/0x100 > [ 1391.699273] ret_from_fork+0x22/0x30 > > To fix the problem: restore the old get_cpu/put_cpu code with some > modifications to reduce the size of the critical section. > > Fixes: d576a5e80cd0 ("bnx2fc: Improve stats update mechanism") > Tested-by: Guangwu Zhang <guazhang@redhat.com> > Signed-off-by: John Meneghini <jmeneghi@redhat.com> > --- > drivers/scsi/bnx2fc/bnx2fc_fcoe.c | 21 +++++++++++++-------- > 1 file changed, 13 insertions(+), 8 deletions(-) > > diff --git a/drivers/scsi/bnx2fc/bnx2fc_fcoe.c > b/drivers/scsi/bnx2fc/bnx2fc_fcoe.c > index 71fa62bd3083..e41a94dc2d1f 100644 > --- a/drivers/scsi/bnx2fc/bnx2fc_fcoe.c > +++ b/drivers/scsi/bnx2fc/bnx2fc_fcoe.c > @@ -508,7 +508,8 @@ static int bnx2fc_l2_rcv_thread(void *arg) > > static void bnx2fc_recv_frame(struct sk_buff *skb) > { > - u32 fr_len; > + u64 crc_err; > + u32 fr_len, fr_crc; > struct fc_lport *lport; > struct fcoe_rcv_info *fr; > struct fc_stats *stats; > @@ -542,6 +543,11 @@ static void bnx2fc_recv_frame(struct sk_buff *skb) > skb_pull(skb, sizeof(struct fcoe_hdr)); > fr_len = skb->len - sizeof(struct fcoe_crc_eof); > > + stats = per_cpu_ptr(lport->stats, get_cpu()); > + stats->RxFrames++; > + stats->RxWords += fr_len / FCOE_WORD_TO_BYTE; > + put_cpu(); > + > fp = (struct fc_frame *)skb; > fc_frame_init(fp); > fr_dev(fp) = lport; > @@ -624,16 +630,15 @@ static void bnx2fc_recv_frame(struct sk_buff *skb) > return; > } > > - stats = per_cpu_ptr(lport->stats, smp_processor_id()); > - stats->RxFrames++; > - stats->RxWords += fr_len / FCOE_WORD_TO_BYTE; > + fr_crc = le32_to_cpu(fr_crc(fp)); > > - if (le32_to_cpu(fr_crc(fp)) != > - ~crc32(~0, skb->data, fr_len)) { > - if (stats->InvalidCRCCount < 5) > + if (unlikely(fr_crc != ~crc32(~0, skb->data, fr_len))) { > + stats = per_cpu_ptr(lport->stats, get_cpu()); > + crc_err = (stats->InvalidCRCCount++); > + put_cpu(); > + if (crc_err < 5) > printk(KERN_WARNING PFX "dropping frame with " > "CRC error\n"); > - stats->InvalidCRCCount++; > kfree_skb(skb); > return; > } > -- Thanks for the patch. Acked-by: Saurav Kashyap <skashyap@marvell.com> > 2.27.0
Thanks. Martin, is it too late to get this into staging for v5.17-rc2? Please merge this patch. /John On 1/27/22 00:13, Saurav Kashyap wrote: > Hi John, > >> -----Original Message----- >> From: John Meneghini <jmeneghi@redhat.com> >> Sent: Monday, January 24, 2022 8:21 PM >> To: Saurav Kashyap <skashyap@marvell.com> >> Cc: Nilesh Javali <njavali@marvell.com>; linux-scsi@vger.kernel.org; linux- >> kernel@vger.kernel.org; GR-QLogic-Storage-Upstream <GR-QLogic-Storage- >> Upstream@marvell.com>; mlombard@redhat.com; guazhang@redhat.com >> Subject: [EXT] [PATCH] scsi: bnx2fc: make bnx2fc_recv_frame mp safe >> >> External Email >> >> ---------------------------------------------------------------------- >> Running tests with a debug kernel shows that bnx2fc_recv_frame is >> modifying the per_cpu lport stats counters in a non-mpsafe way. >> Just boot a debug kernel and run the bnx2fc driver with the hardware >> enabled. >> >> [ 1391.699147] BUG: using smp_processor_id() in preemptible [00000000] >> code: bnx2fc_ >> [ 1391.699160] caller is bnx2fc_recv_frame+0xbf9/0x1760 [bnx2fc] >> [ 1391.699174] CPU: 2 PID: 4355 Comm: bnx2fc_l2_threa Kdump: loaded >> Tainted: G B >> [ 1391.699180] Hardware name: HP ProLiant DL120 G7, BIOS J01 >> 07/01/2013 >> [ 1391.699183] Call Trace: >> [ 1391.699188] dump_stack_lvl+0x57/0x7d >> [ 1391.699198] check_preemption_disabled+0xc8/0xd0 >> [ 1391.699205] bnx2fc_recv_frame+0xbf9/0x1760 [bnx2fc] >> [ 1391.699215] ? do_raw_spin_trylock+0xb5/0x180 >> [ 1391.699221] ? bnx2fc_npiv_create_vports.isra.0+0x4e0/0x4e0 [bnx2fc] >> [ 1391.699229] ? bnx2fc_l2_rcv_thread+0xb7/0x3a0 [bnx2fc] >> [ 1391.699240] bnx2fc_l2_rcv_thread+0x1af/0x3a0 [bnx2fc] >> [ 1391.699250] ? bnx2fc_ulp_init+0xc0/0xc0 [bnx2fc] >> [ 1391.699258] kthread+0x364/0x420 >> [ 1391.699263] ? _raw_spin_unlock_irq+0x24/0x50 >> [ 1391.699268] ? set_kthread_struct+0x100/0x100 >> [ 1391.699273] ret_from_fork+0x22/0x30 >> >> To fix the problem: restore the old get_cpu/put_cpu code with some >> modifications to reduce the size of the critical section. >> >> Fixes: d576a5e80cd0 ("bnx2fc: Improve stats update mechanism") >> Tested-by: Guangwu Zhang <guazhang@redhat.com> >> Signed-off-by: John Meneghini <jmeneghi@redhat.com> >> --- >> drivers/scsi/bnx2fc/bnx2fc_fcoe.c | 21 +++++++++++++-------- >> 1 file changed, 13 insertions(+), 8 deletions(-) >> >> diff --git a/drivers/scsi/bnx2fc/bnx2fc_fcoe.c >> b/drivers/scsi/bnx2fc/bnx2fc_fcoe.c >> index 71fa62bd3083..e41a94dc2d1f 100644 >> --- a/drivers/scsi/bnx2fc/bnx2fc_fcoe.c >> +++ b/drivers/scsi/bnx2fc/bnx2fc_fcoe.c >> @@ -508,7 +508,8 @@ static int bnx2fc_l2_rcv_thread(void *arg) >> >> static void bnx2fc_recv_frame(struct sk_buff *skb) >> { >> - u32 fr_len; >> + u64 crc_err; >> + u32 fr_len, fr_crc; >> struct fc_lport *lport; >> struct fcoe_rcv_info *fr; >> struct fc_stats *stats; >> @@ -542,6 +543,11 @@ static void bnx2fc_recv_frame(struct sk_buff *skb) >> skb_pull(skb, sizeof(struct fcoe_hdr)); >> fr_len = skb->len - sizeof(struct fcoe_crc_eof); >> >> + stats = per_cpu_ptr(lport->stats, get_cpu()); >> + stats->RxFrames++; >> + stats->RxWords += fr_len / FCOE_WORD_TO_BYTE; >> + put_cpu(); >> + >> fp = (struct fc_frame *)skb; >> fc_frame_init(fp); >> fr_dev(fp) = lport; >> @@ -624,16 +630,15 @@ static void bnx2fc_recv_frame(struct sk_buff *skb) >> return; >> } >> >> - stats = per_cpu_ptr(lport->stats, smp_processor_id()); >> - stats->RxFrames++; >> - stats->RxWords += fr_len / FCOE_WORD_TO_BYTE; >> + fr_crc = le32_to_cpu(fr_crc(fp)); >> >> - if (le32_to_cpu(fr_crc(fp)) != >> - ~crc32(~0, skb->data, fr_len)) { >> - if (stats->InvalidCRCCount < 5) >> + if (unlikely(fr_crc != ~crc32(~0, skb->data, fr_len))) { >> + stats = per_cpu_ptr(lport->stats, get_cpu()); >> + crc_err = (stats->InvalidCRCCount++); >> + put_cpu(); >> + if (crc_err < 5) >> printk(KERN_WARNING PFX "dropping frame with " >> "CRC error\n"); >> - stats->InvalidCRCCount++; >> kfree_skb(skb); >> return; >> } >> -- > > Thanks for the patch. > > Acked-by: Saurav Kashyap <skashyap@marvell.com> > >> 2.27.0 >
John, Applied to 5.18/scsi-staging, thanks! Please make sure your commit descriptions are left aligned: > Running tests with a debug kernel shows that bnx2fc_recv_frame is > modifying the per_cpu lport stats counters in a non-mpsafe way. > Just boot a debug kernel and run the bnx2fc driver with the hardware > enabled. ^^^^^
John,
Erm. Hit the wrong key. Applied to 5.17/scsi-fixes, of course.
> Martin, is it too late to get this into staging for v5.17-rc2?
Patches sit in linux-next for about a week before they get sent to
Linus.
On Mon, 24 Jan 2022 09:51:10 -0500, John Meneghini wrote: > Running tests with a debug kernel shows that bnx2fc_recv_frame is > modifying the per_cpu lport stats counters in a non-mpsafe way. > Just boot a debug kernel and run the bnx2fc driver with the hardware > enabled. > > [ 1391.699147] BUG: using smp_processor_id() in preemptible [00000000] code: bnx2fc_ > [ 1391.699160] caller is bnx2fc_recv_frame+0xbf9/0x1760 [bnx2fc] > [ 1391.699174] CPU: 2 PID: 4355 Comm: bnx2fc_l2_threa Kdump: loaded Tainted: G B > [ 1391.699180] Hardware name: HP ProLiant DL120 G7, BIOS J01 07/01/2013 > [ 1391.699183] Call Trace: > [ 1391.699188] dump_stack_lvl+0x57/0x7d > [ 1391.699198] check_preemption_disabled+0xc8/0xd0 > [ 1391.699205] bnx2fc_recv_frame+0xbf9/0x1760 [bnx2fc] > [ 1391.699215] ? do_raw_spin_trylock+0xb5/0x180 > [ 1391.699221] ? bnx2fc_npiv_create_vports.isra.0+0x4e0/0x4e0 [bnx2fc] > [ 1391.699229] ? bnx2fc_l2_rcv_thread+0xb7/0x3a0 [bnx2fc] > [ 1391.699240] bnx2fc_l2_rcv_thread+0x1af/0x3a0 [bnx2fc] > [ 1391.699250] ? bnx2fc_ulp_init+0xc0/0xc0 [bnx2fc] > [ 1391.699258] kthread+0x364/0x420 > [ 1391.699263] ? _raw_spin_unlock_irq+0x24/0x50 > [ 1391.699268] ? set_kthread_struct+0x100/0x100 > [ 1391.699273] ret_from_fork+0x22/0x30 > > [...] Applied to 5.17/scsi-fixes, thanks! [1/1] scsi: bnx2fc: make bnx2fc_recv_frame mp safe https://git.kernel.org/mkp/scsi/c/936bd03405fc
diff --git a/drivers/scsi/bnx2fc/bnx2fc_fcoe.c b/drivers/scsi/bnx2fc/bnx2fc_fcoe.c index 71fa62bd3083..e41a94dc2d1f 100644 --- a/drivers/scsi/bnx2fc/bnx2fc_fcoe.c +++ b/drivers/scsi/bnx2fc/bnx2fc_fcoe.c @@ -508,7 +508,8 @@ static int bnx2fc_l2_rcv_thread(void *arg) static void bnx2fc_recv_frame(struct sk_buff *skb) { - u32 fr_len; + u64 crc_err; + u32 fr_len, fr_crc; struct fc_lport *lport; struct fcoe_rcv_info *fr; struct fc_stats *stats; @@ -542,6 +543,11 @@ static void bnx2fc_recv_frame(struct sk_buff *skb) skb_pull(skb, sizeof(struct fcoe_hdr)); fr_len = skb->len - sizeof(struct fcoe_crc_eof); + stats = per_cpu_ptr(lport->stats, get_cpu()); + stats->RxFrames++; + stats->RxWords += fr_len / FCOE_WORD_TO_BYTE; + put_cpu(); + fp = (struct fc_frame *)skb; fc_frame_init(fp); fr_dev(fp) = lport; @@ -624,16 +630,15 @@ static void bnx2fc_recv_frame(struct sk_buff *skb) return; } - stats = per_cpu_ptr(lport->stats, smp_processor_id()); - stats->RxFrames++; - stats->RxWords += fr_len / FCOE_WORD_TO_BYTE; + fr_crc = le32_to_cpu(fr_crc(fp)); - if (le32_to_cpu(fr_crc(fp)) != - ~crc32(~0, skb->data, fr_len)) { - if (stats->InvalidCRCCount < 5) + if (unlikely(fr_crc != ~crc32(~0, skb->data, fr_len))) { + stats = per_cpu_ptr(lport->stats, get_cpu()); + crc_err = (stats->InvalidCRCCount++); + put_cpu(); + if (crc_err < 5) printk(KERN_WARNING PFX "dropping frame with " "CRC error\n"); - stats->InvalidCRCCount++; kfree_skb(skb); return; }
Running tests with a debug kernel shows that bnx2fc_recv_frame is modifying the per_cpu lport stats counters in a non-mpsafe way. Just boot a debug kernel and run the bnx2fc driver with the hardware enabled. [ 1391.699147] BUG: using smp_processor_id() in preemptible [00000000] code: bnx2fc_ [ 1391.699160] caller is bnx2fc_recv_frame+0xbf9/0x1760 [bnx2fc] [ 1391.699174] CPU: 2 PID: 4355 Comm: bnx2fc_l2_threa Kdump: loaded Tainted: G B [ 1391.699180] Hardware name: HP ProLiant DL120 G7, BIOS J01 07/01/2013 [ 1391.699183] Call Trace: [ 1391.699188] dump_stack_lvl+0x57/0x7d [ 1391.699198] check_preemption_disabled+0xc8/0xd0 [ 1391.699205] bnx2fc_recv_frame+0xbf9/0x1760 [bnx2fc] [ 1391.699215] ? do_raw_spin_trylock+0xb5/0x180 [ 1391.699221] ? bnx2fc_npiv_create_vports.isra.0+0x4e0/0x4e0 [bnx2fc] [ 1391.699229] ? bnx2fc_l2_rcv_thread+0xb7/0x3a0 [bnx2fc] [ 1391.699240] bnx2fc_l2_rcv_thread+0x1af/0x3a0 [bnx2fc] [ 1391.699250] ? bnx2fc_ulp_init+0xc0/0xc0 [bnx2fc] [ 1391.699258] kthread+0x364/0x420 [ 1391.699263] ? _raw_spin_unlock_irq+0x24/0x50 [ 1391.699268] ? set_kthread_struct+0x100/0x100 [ 1391.699273] ret_from_fork+0x22/0x30 To fix the problem: restore the old get_cpu/put_cpu code with some modifications to reduce the size of the critical section. Fixes: d576a5e80cd0 ("bnx2fc: Improve stats update mechanism") Tested-by: Guangwu Zhang <guazhang@redhat.com> Signed-off-by: John Meneghini <jmeneghi@redhat.com> --- drivers/scsi/bnx2fc/bnx2fc_fcoe.c | 21 +++++++++++++-------- 1 file changed, 13 insertions(+), 8 deletions(-)