On 2/4/25 16:37, Alexander Atanasov wrote:
On 4.02.25 10:19, Pavel Tikhomirov wrote:
On 2/4/25 15:57, Alexander Atanasov wrote:
On 4.02.25 9:44, Pavel Tikhomirov wrote:
On 2/3/25 15:57, Alexander Atanasov wrote:
On 3.02.25 9:49, Pavel Tikhomirov wrote:
On 2/3/25 15:42, Alexander Atan
Ack
On Tue, Feb 4, 2025 at 4:41 PM Liu Kui wrote:
>
> It is a normal scenario that the request may have been killed when a
> response is returned. So we shouldn't return -EINVAL to the response
> writer, just need to complete the request with error. Otherwise it
> breaks the userspace thread.
>
>
On 04.02.2025 07:39, Alexander Atanasov wrote:
On 4.02.25 8:30, Pavel Tikhomirov wrote:
On 2/3/25 14:50, Alexander Atanasov wrote:
On 31.01.25 10:32, Pavel Tikhomirov wrote:
On 1/24/25 23:35, Alexander Atanasov wrote:
@@ -353,23 +352,20 @@ static void ploop_dispatch_pio(struct ploop
*ploo
On 2/4/25 15:57, Alexander Atanasov wrote:
On 4.02.25 9:44, Pavel Tikhomirov wrote:
On 2/3/25 15:57, Alexander Atanasov wrote:
On 3.02.25 9:49, Pavel Tikhomirov wrote:
On 2/3/25 15:42, Alexander Atanasov wrote:
On 3.02.25 9:27, Pavel Tikhomirov wrote:
On 2/3/25 14:45, Alexander Atana
On 4.02.25 10:19, Pavel Tikhomirov wrote:
On 2/4/25 15:57, Alexander Atanasov wrote:
On 4.02.25 9:44, Pavel Tikhomirov wrote:
On 2/3/25 15:57, Alexander Atanasov wrote:
On 3.02.25 9:49, Pavel Tikhomirov wrote:
On 2/3/25 15:42, Alexander Atanasov wrote:
On 3.02.25 9:27, Pavel Tikhomirov
On 1/24/25 23:35, Alexander Atanasov wrote:
@@ -1124,8 +1124,13 @@ static int ploop_alloc_cluster(struct ploop *ploop,
struct ploop_index_wb *piwb,
goto out;
to = kmap_local_page(piwb->bat_page);
+ spin_lock_irqsave(&piwb->md->md_lock, flags);
WRITE_ONCE(to[c
It is a normal scenario that the request may have been killed when a
response is returned. So we shouldn't return -EINVAL to the response
writer, just need to complete the request with error. Otherwise it
breaks the userspace thread.
Related to #VSTOR-98704
Signed-off-by: Liu Kui
---
fs/fuse/dev
