From 33cf91947c845c74f78c8890c8419778ecbff488 Mon Sep 17 00:00:00 2001 From: Stanley Chu Date: Sun, 9 Aug 2020 13:07:34 +0800 Subject: [PATCH] scsi: ufs: Fix possible infinite loop in ufshcd_hold [ Upstream commit 93b6c5db06028a3b55122bbb74d0715dd8ca4ae0 ] In ufshcd_suspend(), after clk-gating is suspended and link is set as Hibern8 state, ufshcd_hold() is still possibly invoked before ufshcd_suspend() returns. For example, MediaTek's suspend vops may issue UIC commands which would call ufshcd_hold() during the command issuing flow. Now if UFSHCD_CAP_HIBERN8_WITH_CLK_GATING capability is enabled, then ufshcd_hold() may enter infinite loops because there is no clk-ungating work scheduled or pending. In this case, ufshcd_hold() shall just bypass, and keep the link as Hibern8 state. Link: https://lore.kernel.org/r/20200809050734.18740-1-stanley.chu@mediatek.com Reviewed-by: Avri Altman Co-developed-by: Andy Teng Signed-off-by: Andy Teng Signed-off-by: Stanley Chu Signed-off-by: Martin K. Petersen Signed-off-by: Sasha Levin --- drivers/scsi/ufs/ufshcd.c | 5 ++++- 1 file changed, 4 insertions(+), 1 deletion(-) diff --git a/drivers/scsi/ufs/ufshcd.c b/drivers/scsi/ufs/ufshcd.c index 136b863bc1d45..2b5acdfb17820 100644 --- a/drivers/scsi/ufs/ufshcd.c +++ b/drivers/scsi/ufs/ufshcd.c @@ -1566,6 +1566,7 @@ static void ufshcd_ungate_work(struct work_struct *work) int ufshcd_hold(struct ufs_hba *hba, bool async) { int rc = 0; + bool flush_result; unsigned long flags; if (!ufshcd_is_clkgating_allowed(hba)) @@ -1597,7 +1598,9 @@ int ufshcd_hold(struct ufs_hba *hba, bool async) break; } spin_unlock_irqrestore(hba->host->host_lock, flags); - flush_work(&hba->clk_gating.ungate_work); + flush_result = flush_work(&hba->clk_gating.ungate_work); + if (hba->clk_gating.is_suspended && !flush_result) + goto out; spin_lock_irqsave(hba->host->host_lock, flags); goto start; }