[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <CAMGffEmgcKh1xGOrRNsBq_GrEbL0a1UdG3dgg6Mg0vrsKzW_Vg@mail.gmail.com>
Date: Fri, 14 Jun 2024 07:32:51 +0200
From: Jinpu Wang <jinpu.wang@...os.com>
To: TJ Adams <tadamsjr@...gle.com>
Cc: Jack Wang <jinpu.wang@...ud.ionos.com>,
"James E . J . Bottomley" <James.Bottomley@...senpartnership.com>,
"Martin K . Petersen" <martin.petersen@...cle.com>, linux-scsi@...r.kernel.org,
linux-kernel@...r.kernel.org, Igor Pylypiv <ipylypiv@...gle.com>
Subject: Re: [PATCH 1/3] scsi: pm80xx: Set phy->enable_completion only when we
wait for it
On Fri, Jun 7, 2024 at 7:57 PM TJ Adams <tadamsjr@...gle.com> wrote:
>
> From: Igor Pylypiv <ipylypiv@...gle.com>
>
> pm8001_phy_control() populates the enable_completion pointer with a
> stack address, sends a PHY_LINK_RESET / PHY_HARD_RESET, waits 300 ms,
> and returns. The problem arises when a phy control response comes late.
> After 300 ms the pm8001_phy_control() function returns and the passed
> enable_completion stack address is no longer valid. Late phy control
> response invokes complete() on a dangling enable_completion pointer
> which leads to a kernel crash.
>
> Signed-off-by: Igor Pylypiv <ipylypiv@...gle.com>
> Signed-off-by: Terrence Adams <tadamsjr@...gle.com>
Acked-by: Jack Wang <jinpu.wang@...os.com>
> ---
> drivers/scsi/pm8001/pm8001_sas.c | 4 +++-
> 1 file changed, 3 insertions(+), 1 deletion(-)
>
> diff --git a/drivers/scsi/pm8001/pm8001_sas.c b/drivers/scsi/pm8001/pm8001_sas.c
> index a5a31dfa4512..ee2da8e49d4c 100644
> --- a/drivers/scsi/pm8001/pm8001_sas.c
> +++ b/drivers/scsi/pm8001/pm8001_sas.c
> @@ -166,7 +166,6 @@ int pm8001_phy_control(struct asd_sas_phy *sas_phy, enum phy_func func,
> unsigned long flags;
> pm8001_ha = sas_phy->ha->lldd_ha;
> phy = &pm8001_ha->phy[phy_id];
> - pm8001_ha->phy[phy_id].enable_completion = &completion;
>
> if (PM8001_CHIP_DISP->fatal_errors(pm8001_ha)) {
> /*
> @@ -190,6 +189,7 @@ int pm8001_phy_control(struct asd_sas_phy *sas_phy, enum phy_func func,
> rates->maximum_linkrate;
> }
> if (pm8001_ha->phy[phy_id].phy_state == PHY_LINK_DISABLE) {
> + pm8001_ha->phy[phy_id].enable_completion = &completion;
> PM8001_CHIP_DISP->phy_start_req(pm8001_ha, phy_id);
> wait_for_completion(&completion);
> }
> @@ -198,6 +198,7 @@ int pm8001_phy_control(struct asd_sas_phy *sas_phy, enum phy_func func,
> break;
> case PHY_FUNC_HARD_RESET:
> if (pm8001_ha->phy[phy_id].phy_state == PHY_LINK_DISABLE) {
> + pm8001_ha->phy[phy_id].enable_completion = &completion;
> PM8001_CHIP_DISP->phy_start_req(pm8001_ha, phy_id);
> wait_for_completion(&completion);
> }
> @@ -206,6 +207,7 @@ int pm8001_phy_control(struct asd_sas_phy *sas_phy, enum phy_func func,
> break;
> case PHY_FUNC_LINK_RESET:
> if (pm8001_ha->phy[phy_id].phy_state == PHY_LINK_DISABLE) {
> + pm8001_ha->phy[phy_id].enable_completion = &completion;
> PM8001_CHIP_DISP->phy_start_req(pm8001_ha, phy_id);
> wait_for_completion(&completion);
> }
> --
> 2.45.2.505.gda0bf45e8d-goog
>
Powered by blists - more mailing lists